The business planning and evaluation process

Abstract#

This chapter examines the business planning and evaluation process — from its intellectual origins to contemporary analytical tools. The theoretical part traces the development of planning thought through four paradigms: administrative planning (Taylor, Fayol, Weber), strategic planning (Ansoff, Andrews, Porter), strategic thinking (Mintzberg, Barney, Prahalad and Hamel) and adaptive planning (Ries, Sarasvathy, McGrath, Teece). It presents the three-level hierarchy of planning (strategic, tactical, operational), a five-stage model of the planning process (assumptions → forecasting → modelling → evaluation → decision) and the integration of evaluation at every stage (ex ante, in itinere, ex post). The critical part analyses cognitive biases, the epistemological limits of forecasting, information asymmetry and the shortcomings of integration models.

The central thesis of the chapter is that business planning is not a one-off document but a continuous process of modelling decisions and evaluating them analytically, in which every business decision has alternatives that can be modelled quantitatively and assessed analytically — while remaining aware of the sensitivity of assumptions, the uncertainty of forecasts and the limits of models. The analytical part presents the basic tools of forecasting and evaluation (CAGR, the simple moving average, exponential smoothing, linear regression, scenario analysis, sensitivity analysis), illustrated by two case analyses — a business plan for a Lithuanian farm applying for support and the planning of a technology start-up — together with instructions for building a spreadsheet model.

The chapter is directly linked to the whole later structure of the book: the foundations of value in Chapter 1 provide the ontological basis, Chapters 3–4 examine risk in more detail, Chapters 5–8 apply the planning process to particular functional areas, Chapters 9–15 to financial analysis and valuation, Chapter 16 to an integrated financial model of a business plan, and Chapters 17–18 to plan implementation and exit. The planning cycle presented here is the system within which all these topics acquire meaning.

Learning outcomes#

After reading this chapter and completing the exercises, the reader will be able to:

1. Recognise the levels of planning (strategic, tactical, operational) and the five stages of the planning process in a specific organisational situation (recognition).
2. Explain the development of the four planning paradigms and state the conditions under which each of them remains the most appropriate (comprehension).
3. Calculate the growth rate, simple moving average and exponential smoothing forecasts, the weighted expected value and the elasticity coefficient, and interpret the result for a decision (application).
4. Analyse the assumptions and forecasts of a plan: identify the load-bearing assumptions, the signs of a structural break and the traces of cognitive biases (analysis).
5. Construct a three-scenario financial model with a separate assumptions area and a sensitivity table (synthesis).
6. Assess whether a particular plan can be relied on when making a decision: whether evaluation is carried out ex ante, in itinere and ex post, and under what conditions the conclusion would change (evaluation).

2.1 Theoretical part#

In Chapter 1, value was defined as the present expression of future cash flows, dependent on the assumptions made about those flows. This leaves open the question of where the assumptions come from and who tests them. The present chapter answers it: planning is the process in which assumptions are formulated, turned into forecasts and models, and evaluation tests them at every stage. Without this link, the concept of value in Chapter 1 remains theoretical — only what has first been planned and justified can be discounted.

2.1.1 Historical and conceptual context#

Why the history of planning matters to the practitioner#

Business planning today looks like a self-evident management function — it is hard to imagine an organisation without at least a budgeting process. Yet what we now call business planning is the result of several decades of intellectual evolution, and understanding that evolution is not mere academic curiosity. The different traditions of planning — rational strategic planning, emergent strategy, the adaptive model — did not replace one another in sequence; they coexist in contemporary practice and often cause conceptual confusion. A manager who applies a formal strategic planning process in a start-up environment, and an entrepreneur who abandons any plan at all “because Mintzberg proved that plans do not work”, both make the same mistake: they apply one paradigm where another is required.

The purpose of this section is to trace how the concept of business planning was formed and changed over the past century, and to show that effective contemporary planning is not the triumph of one school but a synthesis of several traditions, in which the planning method must match the level of uncertainty and the context of the organisation.

The era of administrative management: planning as a function (1910–1960)#

The origins of business planning are linked to the management needs created by the industrial revolution. Until the end of the nineteenth century the activity of most firms was simple enough for the intuition and experience of an owner-manager to suffice. As the scale and complexity of organisations grew, however, a need arose to formalise decision-making.

Frederick Winslow Taylor (1856–1915) published The Principles of Scientific Management in 1911, formulating scientific management — the view that work processes can be analysed, measured and optimised by systematic methods. Taylor’s contribution to the theory of planning is indirect but essential: he established the idea that management is not an art or a matter of intuition but a systematic, analytical process in which data and measurement are the principal instruments of decision-making (Taylor, 1911). Taylorism created the intellectual environment in which planning became a legitimate managerial function. Taylor’s vision of planning, however, was essentially operational: it covered the breaking down, standardisation and optimisation of work operations, but it did not question strategic goals and took no account of environmental uncertainty. The future was treated as stable and predictable — it was enough to find the “best method” and adhere to it.

Henri Fayol (1841–1925) is regarded as the originator of the classical concept of planning. In his 1916 work Administration Industrielle et Générale, planning (Fr. prévoyance) is identified as the first of five management functions: planning, organising, commanding, co-ordinating and controlling. Fayol defined planning as “examining the future and drawing up the plan of action” (Fayol, 1916/1949, pp. 5–6), and he called the preparation of the plan itself one of the most difficult and most important matters of every business, one that brings into play all departments and all functions (p. 43).

Fayol’s approach, however, was more subtle than textbooks usually present it. He was the first to view the organisation “top-down” as a single system of management, and he formulated fourteen principles of management — among them the division of work, unity of command, unity of direction and the scalar chain (the formal hierarchy of authority). Yet Fayol clearly rejected the logic of a recipe that fits everyone: in management affairs, in his words, there is nothing rigid or absolute — everything is a question of proportion, and therefore the principles are flexible guidelines, adaptable to every need, while knowing how to make use of them is a difficult art (Fayol, 1916/1949, p. 19). Fayol’s most important insight is that planning is not a one-off act but a continuous process encompassing the setting of objectives, the design of a sequence of actions and the allocation of resources. A plan, according to Fayol, must be flexible enough to accept the adjustments it is considered well to introduce, whether from the pressure of circumstances or for other reasons (Fayol, 1916/1949, p. 45) — an idea that anticipates later discussions of adaptiveness.

In this context the contribution of Max Weber (1864–1920) and his model of bureaucracy should also be mentioned. Formalised procedures, hierarchy and a rational system of decision-making formed the organisational infrastructure within which planning could function as an institutionalised practice (Weber, 1922/1978). Weber’s bureaucratic model was not a theory of planning per se, but it created the structural basis — a clear chain of responsibility, documented processes, rationality as an organisational principle — without which systematic planning would not have been possible.

The era of administrative management thus established a fundamental assumption: management is a rational activity that can be analysed, and planning is its point of departure. That assumption has survived to this day, although its interpretation has changed substantially.

The formation of the strategic planning paradigm (1960–1990)#

The real breakthrough in the theory of planning came in the 1960s, when strategic planning developed out of corporate planning, which had largely consisted of extrapolating financial forecasts. This was not merely a change of terminology: strategic planning shifted attention from internal operations to the external environment, and from short-term budgeting to long-term competitive positioning.

H. Igor Ansoff (1918–2002) is regarded as the founder of the discipline of strategic planning. His 1965 work Corporate Strategy: An Analytic Approach to Business Policy for Growth and Expansion formulated an analytical system for taking corporate decisions on growth and diversification. Ansoff proposed a product–market matrix (known as the Ansoff matrix), classifying growth strategies along two dimensions — an existing or a new product and an existing or a new market: market penetration, product development, market development and diversification (Ansoff, 1965).

More important than the particular matrices, however, is Ansoff’s vision of strategic planning as a rational, staged process: analysis of the environment → setting objectives → generating strategic alternatives → evaluation and choice → implementation planning. This model rested on several essential assumptions: (1) the future is at least partly predictable and can be forecast; (2) the management of a firm can and should formulate clear long-term objectives; (3) strategic decisions can be taken in a systematic, analytical manner; (4) the environment is essentially manageable. Mintzberg and his colleagues later called this approach the “planning school” — one of ten schools of strategy formation and one of the three prescriptive schools, alongside the design and positioning schools (Mintzberg et al., 1998, pp. 5, 48).

Ansoff subsequently supplemented his concept with two important ideas. The first is weak signals analysis (Ansoff, 1975): organisations must monitor the environment continuously for early indicators of strategic discontinuities, instead of waiting until the changes become obvious. The second is the Contingent Strategic Success Formula (CSSF): the profitability of an organisation depends on strategic aggressiveness, managerial responsiveness and their match with the level of environmental turbulence (Ansoff, 1979). This idea — that the planning method must match the level of environmental uncertainty — became one of the most important axioms of contemporary planning theory.

In parallel with Ansoff, Kenneth Andrews (1916–2005) developed at Harvard Business School a less formalised but widely adopted conception of strategy formation. His work The Concept of Corporate Strategy (1971), together with the earlier work of Learned, Christensen, Andrews and Guth (1965), formulated the method of SWOT analysis (strengths, weaknesses, opportunities, threats), which became one of the most widely used instruments of strategic planning in practice.

Consultants at the Boston Consulting Group (BCG) and McKinsey created portfolio management matrices in the 1970s (the BCG matrix, the GE/McKinsey matrix) that allowed corporations to apply a systematic planning approach to managing a diversified portfolio of business units. This was the high point of formal strategic planning — large corporations set up entire strategic planning departments with dozens of specialists (Kiechel, 2010). Planning became an industry — and for that very reason it became vulnerable to criticism.

Michael E. Porter (b. 1947) brought an essential analytical dimension into the theory of planning: the structural analysis of the competitive environment. His Competitive Strategy (1980) presented the Five Forces Framework: the threat of new entrants, the bargaining power of buyers, the bargaining power of suppliers, the threat of substitute products and the intensity of rivalry within an industry. Competitive Advantage (1985) supplemented this analysis with the concept of the value chain.

Porter’s contribution to the planning context is twofold. First, he gave planning a dimension of content: before actions are planned, the structure of the industry, the competitive forces and the logic of value creation must be understood. This supplemented Ansoff’s procedural approach (how to plan) with content (what to plan) (Porter, 1980). Second, the three generic strategies — cost leadership, differentiation and focus — gave planning a clear structure of choices. Porter warned that being “stuck in the middle” — having no clear strategic position — is risky; such a firm is usually far less profitable than competitors that have chosen one of the generic strategies (Porter, 1980, p. 41; 1985, pp. 16–17).

Mintzberg’s critique and emergent strategy#

The most serious blow to formal strategic planning was delivered by Henry Mintzberg (b. 1939). His long cycle of research, culminating in The Rise and Fall of Strategic Planning (1994), called the entire paradigm of the planning school into question.

Mintzberg’s principal arguments are as follows.

1. The false promise of formalisation. Strategy formation is a creative process requiring synthesis, and it cannot be broken down into formalised stages. Planning can analyse and codify an existing strategy, but it cannot create a new one — Mintzberg called this the “programming” of strategy rather than the “formation” of strategy (Mintzberg, 1994).

2. Intended and emergent strategy. Mintzberg distinguished intended strategy — the one deliberately planned by management — from emergent strategy, which takes shape gradually, without an explicit plan, through the interaction between the actions of the organisation and its environment. Between them lie deliberate strategy, which was planned and implemented, and unrealised strategy, which remained on paper. In real organisations most strategies are mixed: partly intended, partly emergent (Mintzberg & Waters, 1985). This insight is fundamental: if a real strategy is always partly the result of a plan and partly of chance, then the role of formal planning is limited, and an organisation that does not acknowledge emergent strategies loses the ability to learn from its own experience.

3. The limits of forecasting. Strategic planning rests on the assumption that the future can be forecast at least in part. Environmental uncertainty, however — technological, geopolitical and market discontinuities — makes long-term forecasts fundamentally unreliable. Planning that does not appreciate the limits of its own forecasts is not merely ineffective but dangerous, because it creates a false illusion of certainty.

4. Detachment from strategic thinking. A formal planning process often becomes detached from genuine strategic thinking at the top of the organisation — planning departments produce documents that do not become real decisions.

Mintzberg’s critique should not, however, be interpreted as a rejection of planning per se. Mintzberg himself accepted that organisations need planning — but that it should be understood as strategic thinking rather than strategic planning in the narrow sense (Mintzberg, 1994, pp. 256, 273, 382). Planning is useful as the programming of strategy — the elaboration, co-ordination and communication of strategic decisions already taken — but not as a means of creating strategy. This distinction has important implications for the practice of business planning: effective planning requires not only formal procedures but also the creative thinking that “shapes” the future rather than merely “forecasting” it.

The resource-based view: criticism from the other side#

Porter’s models were criticised not only by Mintzberg but also from another perspective — that of the Resource-Based View (RBV). Jay Barney (1991) showed that competitive advantage arises not only from the structure of the industry (the external environment) but also from the unique, hard-to-imitate resources and capabilities of the organisation (the internal environment). A resource confers a sustained competitive advantage if it is valuable, rare, hard to imitate and has no equivalent substitutes — the so-called VRIN criterion (Barney, 1991).

C. K. Prahalad and Gary Hamel (1990) formulated the related concept of core competencies: a core competence is the collective learning of the organisation — above all the ability to co-ordinate diverse production skills and integrate several streams of technology — and three tests help to identify it: it provides potential access to a variety of markets, it contributes significantly to the customer’s perceived benefit from the end product, and it is difficult to imitate (Prahalad & Hamel, 1990, pp. 81, 83). The authors proposed seeing the firm not only as a portfolio of businesses but also as a portfolio of competencies, so that planning begins with the question of what the organisation can do better than others, rather than with the analysis of market structure alone. This perspective significantly enriched the content of the planning discourse: planning should assess not only the external environment but also internal capabilities, knowledge and organisational culture as a source of value.

The debate between Porter (positioning in the market) and Barney–Prahalad–Hamel (internal resources) reveals a fundamental dilemma of planning theory: where does value creation begin — in the market or inside the organisation? The contemporary answer — in both places — means that a business plan must integrate both an analysis of the external environment and an assessment of internal resources.

The contemporary paradigm of adaptive planning#

In the first decades of the twenty-first century the theory of planning underwent another fundamental transformation, driven by several factors: the pace of change created by digital technologies, the uncertainty increased by globalisation, the growth of the start-up economy and the influence of the Agile approach to software development on management thinking.

Eric Ries (2011), in The Lean Startup, formulated the principles of the lean startup methodology, at the centre of which lies the build–measure–learn cycle: a business assumption is turned into a Minimum Viable Product (MVP), this is tested in the market, and the results are used for learning — by adjusting the assumption or by changing direction (a pivot). In essence, Ries proposed a planning model in which the plan is not a long-term document but a set of continuously tested hypotheses. Every planning assumption must be turned into a testable hypothesis and checked in the market as quickly and as cheaply as possible (Ries, 2011). This approach inverted the logic of planning: instead of “plan everything and then implement it”, the rule becomes “plan a hypothesis and test it at once”.

Saras Sarasvathy (2001) formulated effectuation theory — a conceptual model that she develops through two thought experiments and concludes with a list of propositions to be tested empirically (Sarasvathy, 2001, pp. 243–244). In contrast to traditional planning, which starts from goals and looks for the means of achieving them (causal logic), effectuation starts from the means at hand — “who I am, what I know, whom I know” (p. 250) — and asks what goals can be achieved with those resources. Effectuation does not reject planning, but it fundamentally changes its direction: from goals towards means, from prediction towards control, from competitive analysis towards the building of partnerships (Sarasvathy, 2001).

Rita Gunther McGrath (2013) formulated the concept of transient competitive advantage: a sustainable competitive advantage is an idea of the last century; in the contemporary environment advantages are transient, and organisations must plan not for static positions but for a dynamic process of creating advantages and of healthy disengagement from those that are fading (McGrath, 2013, pp. xi, 51–74). This means that a business plan cannot be a static document intended to preserve the current position — it must be a dynamic instrument for continuous adaptation.

Jeremy Hope and Robin Fraser (2003), in Beyond Budgeting, formulated an alternative to traditional budgeting: abandon fixed annual budgets, use rolling forecasts, link the allocation of resources to value creation rather than to historical budgets, and delegate decision-making to lower levels of the organisation. This conception is directly related to the approach advocated in this book: planning as a continuous process of modelling decisions rather than a one-off document.

The Agile approach, which originated in software development (Beck et al., 2001), has had a wider influence on business planning: iterative development, rapid feedback, adjustment of the plan in real time. The Agile approach shows that planning can be effective even without long-term detailed forecasts — it is enough to have a clear vision, short-term sprints and continuous learning from results.

David Teece’s (2007) theory of dynamic capabilities supplemented the resource-based view with a time dimension: what matters is not only holding valuable resources but also being able to reconfigure them continuously as environmental conditions change. In the context of planning this means that a plan must cover not only the allocation of resources but also the organisation’s ability to change — adaptiveness as a strategic competence.

The development of planning paradigms: a synthesis#

The development of planning thought reviewed above can be structured into four paradigms, each of which supplemented rather than replaced those before it.

Table 2.1. The development of business planning paradigms

This evolution reveals a clear regularity: the greater the uncertainty of the environment, the more planning moves from forecasting (attempting to predict the future) to adaptation (the ability to respond to it). That does not mean, however, that the earlier paradigms have become useless. Stable industries in which change is slow and predictable can still apply the rational Ansoff–Porter model effectively. The essential contemporary principle is that the planning method must match the level of uncertainty (Courtney et al., 1997): low uncertainty permits analytical planning, medium uncertainty calls for scenario analysis, and high uncertainty for an adaptive, experimental approach.

The paradox of planning and the position taken in this book#

All the paradigms discussed converge on one essential paradox: plans are almost always wrong, yet planning is indispensable. No forecast coincides exactly with reality, but an organisation that does not plan is in an even worse position than one whose plans were only partly borne out. As Eisenhower put it: “Plans are worthless, but planning is everything” (Eisenhower, 1957) — he added that he had heard the saying during his time in the Army. Russell Ackoff (1970, p. 15) expressed a similar thought: the principal value of planning lies not in the plans produced but in the process of producing them.

The essence of the paradox is that the value of planning arises not from the accuracy of forecasts but from several other functions.

1. Identifying and testing assumptions. Planning forces assumptions to be stated explicitly that would otherwise remain unexamined — and, at the same time, vulnerable. Is the growth of sales consistent with the size of the market? Are the margins consistent with the competitive environment?
2. Co-ordination and communication. A plan ensures that the members of the organisation work towards shared objectives and understand the common logic.
3. A structure for decisions. Planning creates an analytical system within which alternatives can be compared and well-founded decisions taken.
4. Preparedness for change. Scenario analysis and “what if” thinking prepare the organisation for adaptation.
5. Learning. Arie de Geus (1988) defined planning as learning and corporate planning as institutional learning: the real purpose of planning, in his view, is not to produce a plan but to change the mental models that decision makers carry in their heads (de Geus, 1988, p. 71, as cited in Mintzberg, 1994, p. 377). This idea links the disciplines of planning and evaluation: evaluation becomes not a “final answer” but an instrument of structured learning about the drivers of business value.

This book takes a position that integrates the best of all the paradigms discussed: business planning is a continuous process of modelling decisions and evaluating them analytically, in which every business decision has alternatives that can be modelled quantitatively and assessed analytically — while remaining aware of the sensitivity of assumptions, the uncertainty of forecasts and the limits of models, yet still grounding decisions in figures rather than in intuition alone. This is a rational core (Ansoff, Porter) combined with the recognition of emergent strategy (Mintzberg) and the form of an adaptive, iterative process (Ries, Sarasvathy, McGrath).

The following sections develop this position in detail: section 2.1.2 presents the specific models and levels of the planning process, section 2.1.3 offers a critical discussion of the limits and paradoxes of planning, and section 2.2 sets out the mathematical and analytical tools by which this philosophy is turned into practice.

2.1.2 Principal theories and models#

In the previous section we saw how planning thought evolved from the administrative approach of Taylor and Fayol to the contemporary adaptive model. This section moves from history to structure: how the planning process is organised, what stages it consists of, at what levels it operates, and how evaluation is integrated into every stage of the process. This is the analytical model on which all the later chapters of the book rest.

Levels of planning: strategic, tactical, operational#

One of the foundational structures of planning theory is the three-level hierarchy formulated by Robert Anthony (1965) in his model of management control systems, which became the standard approach both in the academic literature and in business practice. Herbert Simon (1960) supplemented this structure with a dimension of decision typology: the higher levels of planning are dominated by non-programmed decisions — novel, ambiguous, strategically significant — and the lower levels by programmed decisions, which are routine, repetitive and clearly defined.

The strategic level covers decisions with the longest horizon (3–10 years, and in fields such as energy or infrastructure up to 20 years or more): the direction of the organisation, competitive positioning, the structure of the business portfolio, major strategic investments and the transformation of the business model. The level of detail is the lowest — the categories are aggregate ones (markets, product groups, strategic business units). The principal questions are: in which markets should the firm compete? What competitive position should it hold? What are the main priorities for allocating capital? How is a sustainable competitive advantage to be built? The instruments of evaluation at this level are scenario analysis, the discounted cash flow (DCF) model with alternative scenarios, real options, the analysis of strategic investments and economic value added (EVA). Strategic decisions are usually hard to reverse and have a large effect on the whole organisation (Anthony, 1965; Johnson et al., 2017).

The tactical level covers a horizon of one to three years and is the bridge between strategy and day-to-day activity. Here strategic objectives are transformed into specific programmes of action, functional plans and the allocation of resources: the marketing plan, the production plan, the personnel plan, the investment portfolio, departmental budgets. The tactical level is the critical point of integration — it is here that marketing forecasts, supply chain constraints, production capacity and financial targets intersect. Where tactical planning is not formalised, an organisation naturally drifts towards functionally separate plans that may be locally rational but globally uncoordinated (Dreyer et al., 2018). The instruments of evaluation are budgets, return on investment (ROI) analysis, project appraisal, comparative analysis and functional key performance indicators (KPIs).

The operational level covers the shortest horizon — from days to one year — and the highest level of detail: production schedules, inventory management, cash flow management, the allocation of labour, order fulfilment. Decisions at this level are structured and repetitive, yet their effectiveness directly determines whether the strategic objectives will be achieved. The instruments of evaluation are variance analysis, efficiency indicators, service level indicators and capacity utilisation rates.

It should be stressed that the hierarchy is effective only where vertical consistency is ensured: strategic objectives are cascaded consistently into tactical and operational plans, and operational results — through feedback mechanisms — inform tactical and strategic decisions (Meal et al., 1987). Without this two-way link the hierarchy remains a formal structure with no real content.

Figure 2.1. The three-level planning hierarchy with feedback loops

graph TD
    S["Strategic level<br/><i>3–10 years · Direction, positioning</i>"]
    T["Tactical level<br/><i>1–3 years · Functional plans, budgets</i>"]
    O["Operational level<br/><i>Days–1 year · Schedules, execution</i>"]

    S -->|"Objectives and priorities ↓"| T
    T -->|"Programmes of action ↓"| O
    O -->|"Results and variances ↑"| T
    T -->|"Feedback ↑"| S

    style S fill:#2c3e7a,color:#fff
    style T fill:#2980b9,color:#fff
    style O fill:#27ae60,color:#fff

A plan and a forecast are two different things#

In everyday language a plan and a forecast are often used as synonyms, yet from a management point of view they are two different objects, and confusing them is one of the commonest errors in planning.

A forecast is an estimate of a likely state: what will most probably happen if nothing changes. It is descriptive — it describes the world and can be right or wrong. A plan is a normative decision: what the organisation will pursue and what it will do to that end. It is neither right nor wrong — it is either met or not met, ambitious or cautious.

The practical significance of the distinction becomes clear in the question of responsibility. The accuracy of a forecast is the responsibility of analysis and data; the fulfilment of a plan is the responsibility of people. When the two are merged into a single figure, one of two things happens: either the forecast is raised to the status of a target (“we shall grow by 15 %, because that is what the plan says”), or the target is lowered to the level of the forecast (“we plan for what will happen anyway”). In the first case the plan becomes a statement of expectations; in the second it loses its managerial meaning.

The correct sequence is this: a forecast updates a plan but does not replace it. On receiving a new forecast, the organisation decides whether the target of the plan is still attainable, whether it should be pursued by other means, or whether the target should be revised. It is this decision, not the figure itself, that is the output of the planning process.

A model of the planning process: from assumptions to decision#

The planning process — irrespective of level — comprises five integrated stages that form a cyclical rather than a linear sequence.

1. Identifying and formulating assumptions (assumption setting). Every plan begins with assumptions — explicit or implicit statements about the future that are taken to be true for the purposes of planning, although their actual accuracy is unknown. Assumptions may be macroeconomic (GDP growth, inflation, interest rates), sectoral (competitive dynamics, regulatory change), firm-level (market share, pricing, efficiency improvements) or financial (the discount rate, tax rates, working capital requirements). Planning without assumptions is impossible — planning is the modelling of the future, and the future is unknown. The quality of the assumptions — their clarity, their justification and their regular review — is therefore the single most important factor in the quality of planning.

2. Forecasting. Forecasting is the systematic estimation of future values on the basis of the information available. It is the technical infrastructure that supports financial projections and valuation models. Methods of forecasting are divided into quantitative ones (time series methods, regression, Monte Carlo simulation, machine learning) and qualitative ones (expert judgement, the Delphi method, scenario planning). Quantitative methods are suitable where numerical data about the past are available and it is reasonable to assume that some of the past patterns will continue — usually shorter horizons; over longer horizons qualitative methods and scenario analysis become more important, because structural breaks — technological, regulatory, demographic — turn into the principal source of uncertainty (Hyndman & Athanasopoulos, 2021, §1.2, §1.4). An important principle: forecasts should be given as intervals (a range) rather than as point estimates (a single figure) — this reflects uncertainty more adequately and helps in taking better decisions.

3. Modelling. Modelling is the structuring of assumptions and forecasts into analytical models that make it possible to test the relationships between variables and to compare decision alternatives. Models may be deterministic (one value of the result for each combination of assumptions) or stochastic (the variables are treated as probability distributions and the result is a distribution). Deterministic models are simpler and more widely used in practice, but they create an illusion of spurious precision; stochastic models reflect uncertainty more realistically but require more data and more expertise. A practical compromise — a deterministic model with sensitivity and scenario analysis — makes it possible to take some account of uncertainty without the complexity of the full stochastic apparatus. The modelling stage is essential in the context of this book: it is here that every business decision — an investment, a financing structure, a marketing channel, a personnel plan — is turned into alternatives that can be compared quantitatively.

4. Evaluation. Evaluation is the assessment of the value and the prospects of the alternative courses of action. This function is the central thesis of this book: evaluation is not a procedure carried out “after” planning but an integral part of planning — an instrument for modelling decisions. Evaluation is carried out in three time modes:

• Ex ante (before the decision): comparing alternatives and justifying the choice. Which investment creates the most value? Which source of financing is optimal? Which marketing channel is the most effective?
• In itinere (during implementation): monitoring assumptions and adjusting the plan. Is the growth of sales consistent with the assumptions? Is the cost structure consistent with the model? Does the plan need to be adjusted?
• Ex post (after implementation): learning for future planning cycles. Where were the assumptions wrong? Why? How can the process be improved?

This threefold logic of evaluation — ex ante, in itinere, ex post — transforms evaluation from a retrospective control function into a proactive instrument of decision support. The terms are taken from the practice of programme evaluation; in the value management literature the same idea is expressed as the requirement to assess assumptions before a decision, while implementing it and after it.

5. Decision. The final stage is the confirmation of the choice and the commitment of resources. Clear decision governance — who decides, on what basis and through what escalation mechanism — is a necessary condition if the planning process is not to turn into a bureaucratic exercise without real results.

Figure 2.2. The planning process cycle with evaluation integrated

graph LR
    A["1. Assumptions<br/><i>Identification,<br/>formulation</i>"] --> B["2. Forecasting<br/><i>Quantitative,<br/>qualitative</i>"]
    B --> C["3. Modelling<br/><i>Structuring the<br/>alternatives</i>"]
    C --> D["4. Evaluation<br/><i>Ex ante ·<br/>In itinere ·<br/>Ex post</i>"]
    D --> E["5. Decision<br/><i>Authorisation,<br/>resources</i>"]
    E -.->|"Feedback"| A

    style A fill:#8e44ad,color:#fff
    style B fill:#2980b9,color:#fff
    style C fill:#27ae60,color:#fff
    style D fill:#e67e22,color:#fff
    style E fill:#c0392b,color:#fff

It should be stressed that this process is cyclical and iterative rather than a one-off linear sequence. Three reasons make it iterative.

1. Incomplete information — at the start of planning part of the information is unknown; assumptions are set and tested later.
2. The interdependence of plans — sales forecasts affect production plans, production plans affect financial requirements, and financial constraints may force a revision of the sales targets.
3. Change in the environment — the external environment changes both during and after planning, so plans must be updated continuously.

The business plan as an integrated document#

A business plan is not merely a document of financial projections but a structured system for modelling decisions, integrating strategic direction, tactical and operational co-ordination, financial forecasts, an implementation plan and risk monitoring. Although the form and the level of detail of a business plan vary with the context — a plan addressed to the investors of a start-up differs from an application to the National Paying Agency (NMA), the Lithuanian body administering EU agricultural and rural support, under the rural development support measures — its conceptual structure comprises three hierarchical layers that correspond to the planning hierarchy.

The strategic layer (a 3–10 year perspective): mission and vision, market positioning and competitive strategy, business portfolio decisions, major investment programmes, high-level financial objectives (revenue growth, profitability ranges, return on invested capital (ROIC) targets).

The tactical layer (a 1–3 year perspective): functional strategies — the marketing plan (target segments, pricing, distribution channels), the operations plan (capacity decisions, the supply chain, quality targets), the personnel plan (headcount, the development of competencies), the finance plan (the allocation of investment, working capital management, capital structure). Also departmental budgets, the project portfolio, the KPI system and an assessment of assumptions and risks.

The operational layer (months to a year): detailed implementation schedules, the allocation of resources, process documentation, the measurement of performance, contingency plans.

The structure of this book reflects that logic directly: Chapters 5–8 examine functional plans at the tactical level (structure, operating processes, marketing, personnel), and Chapters 9–15 financial analysis and valuation, which is a cross-cutting function permeating all the levels.

Mechanisms of co-ordination: S&OP and IBP#

The planning hierarchy works effectively only where mechanisms of co-ordination exist that ensure horizontal (between functions) and vertical (between levels) integration. The two most important such mechanisms in contemporary practice are sales and operations planning (S&OP) and integrated business planning (IBP).

S&OP is a tactical process designed to balance demand and supply over a medium-term horizon. It comprises five elements.

1. The demand review — checking the sales forecast against the latest information.
2. The supply and capacity review — assessing production and supply capacity.
3. Financial reconciliation — whether the financial consequences are consistent with the objectives of the firm.
4. The management review — cross-functional dialogue and problem solving.
5. Confirmation of decisions and the assignment of responsibility.

S&OP seeks to give the organisation a “single set of numbers”: all the functions — sales, production, finance, supply — should work with the same assumptions and projections rather than with different versions (Dreyer et al., 2018, p. 114; Markin et al., 2021).

Without the co-ordination provided by S&OP, each function naturally optimises its own area: sales pursues growth, production stability, finance profitability and cash flow, and supply costs. These objectives conflict with one another, and without a mechanism of integration the organisation faces internal inconsistency — functional fragmentation (Dreyer et al., 2018).

IBP is a more advanced form of S&OP that systematically integrates financial management, strategic priorities and scenario thinking. Whereas S&OP is mainly concerned with co-ordinating demand and supply, IBP extends the logic to the operationalisation of strategic objectives, financial and operational alignment, organisational learning and the reallocation of resources between priorities. IBP is treated not merely as a tactical balancing process but as a mechanism for executing strategy — a critical bridge between the formulation of strategy and everyday activity (Nyman et al., 2025).

The effectiveness of a planning system depends not only on good forecasts or advanced IT tools but on organisational alignment: shared objectives, shared assumptions, clear decision channels, agreed criteria for assessing performance, and disciplined iteration with explicit change management (Markin et al., 2021; Dreyer et al., 2018). The valuation literature suggests assessing a planning process not by the accuracy of its forecasts but by the extent to which it reduces bias in decisions — a separate discussion of debiasing is devoted to this (Koller et al., 2020, pp. 572, 576–580).

The balanced scorecard and objectives and key results as instruments of integration#

The Balanced Scorecard (BSC), created by Robert Kaplan and David Norton (1992, 1996), is one of the best-known mechanisms for formalising the feedback loop between strategy and operations. The BSC transforms strategic objectives into KPIs across four perspectives.

1. The financial perspective — profitability, growth, the creation of shareholder value.
2. The customer perspective — customer satisfaction, loyalty, market share.
3. The internal process perspective — operational efficiency, quality, innovation.
4. The learning and growth perspective — employee competencies, technological infrastructure, organisational culture.

The value of the BSC in the context of planning lies not only in measurement but also in identifying causal links: investment in learning should improve processes, better processes should satisfy customers, and satisfied customers should generate financial results (Kaplan & Norton, 2001). As will be discussed in section 2.1.3, however, this chain is a hypothesis rather than an established fact: Nørreklit argues that the relationship between the perspectives is logical rather than causal (Nørreklit, 2003, pp. 615–616).

Objectives and Key Results (OKR) is a contemporary tactical and operational instrument of execution whose origins go back to the period when Andy Grove (1936–2016) led the Intel corporation. Drawing on Peter Drucker’s (1954) concept of Management by Objectives (MBO), Grove simplified it and turned it into a more operational mechanism: every Objective has three to five quantitative Key Results by which progress is measured objectively. John Doerr, a former Intel employee, brought the OKR system to Google in 1999, where it became a standard instrument of execution (Doerr, 2018).

OKRs work in short cycles (quarters, sometimes months), set ambitious stretch goals, are transparent (all the OKRs of the organisation are visible to all employees) and are kept separate from the remuneration system so as to encourage ambition rather than caution. It should be noted that OKR is an instrument of practitioners rather than of academics — its empirical grounding is still limited, and its effectiveness is documented in case studies rather than in controlled studies. Even so, OKRs are spreading rapidly beyond the technology sector — they are used by organisations of all sizes around the world (Niven & Lamorte, 2016, pp. 6, 151).

OKR does not compete with the BSC: the BSC is a system for “translating” strategy (long-term), while OKR is a system of tactical and operational execution (short-term). An effective integration is this: the BSC defines the strategic objectives and metrics, and OKR carries them into quarterly execution targets. Both instruments, however, share a common root — Drucker’s MBO — and address the same problem: how to turn a strategic vision into everyday action.

The five dimensions of a planning system: a synthesis#

To summarise the models presented in this section, a planning system can be structured along five dimensions.

Table 2.2. The dimensions of a planning system

These five dimensions constitute an analytical model with which the maturity of any organisation’s planning system can be assessed. In the chapters that follow, each dimension is made specific for particular business decisions: operating processes (Chapter 6), marketing (Chapter 7), personnel (Chapter 8), financial analysis (Chapter 9) and valuation (Chapter 13).

2.1.3 Critical analysis and discussion#

The previous two sections presented the history of planning thought and contemporary models. A university textbook, however, should not confine itself to presenting models — it must assess critically their limits, contradictions and unresolved problems. This section examines five fundamental problems of planning: the tension between rationality and emergence, cognitive biases, the epistemological limits of forecasting, information asymmetry, and criticism of the models of integration.

Rational and emergent planning: not a choice but a spectrum#

Mintzberg’s critique and the concept of emergent strategy, presented in section 2.1.1, are sometimes simplified into a dichotomy: either you plan rationally (Ansoff) or you acknowledge emergence (Mintzberg). Empirical research, however, shows that this dichotomy is misleading.

Brews and Hunt (1999) examined the relationship between formal planning and organisational performance and found that the organisations achieving better results combine a formal planning process with a readiness to adjust strategies in the light of feedback, while the association with performance emerges only where the experience of formal planning extends to four years or more. This means that rationality and emergence are not alternatives — they are combined within the same process, in different degrees over different time horizons.

Mintzberg, Ahlstrand and Lampel (1998), in Strategy Safari, identified ten schools of strategy formation — three prescriptive (design, planning, positioning) and seven descriptive (entrepreneurial, cognitive, learning, power, cultural, environmental, configuration). Their main conclusion is that no single school explains the whole process of strategy formation; real organisations use elements of several schools depending on the context. This insight has a direct practical implication: whoever appraises a business plan (an investor, a bank, an expert at the National Paying Agency) should assess not only the quality of the plan’s content but also the adequacy of the planning process — whether the planning method used matches the level of uncertainty in the business environment.

The contemporary answer to the rationality–emergence dilemma is strategic ambidexterity — the ability of an organisation to exploit existing advantages (rational planning) and explore new opportunities (emergent learning) at the same time. This requires not one mode of operation but two: a structured mode for stable activity and an experimental mode for innovation and adaptation.

Cognitive biases: planning as a psychological problem#

Even where a planning process is well designed structurally, it remains vulnerable to systematic errors of thinking — cognitive biases, revealed by the research programme of Daniel Kahneman and Amos Tversky, from the regularities of decision-making under risk in prospect theory (Kahneman & Tversky, 1979) to later work on anchoring and intuitive prediction.

Optimism bias (data on Lithuanian projects are given in section 2.1.5). Planners systematically overestimate favourable outcomes and underestimate unfavourable ones — costs, time and risks. This is not deliberate manipulation but a deeply rooted cognitive tendency. Kahneman and Lovallo (1993) found that optimism bias is caused by the dominance of the “inside view”: the planner concentrates on the specifics of the particular project and pays too little attention to the statistical results of similar projects. The corrective mechanism is reference class forecasting (also called comparison class forecasting, RCF): testing a forecast against the statistical distribution of similar projects rather than against the circumstances of the project alone.

The planning fallacy. Related to optimism bias, but specifically covering the systematic underestimation of time and costs and the overestimation of benefits. Bent Flyvbjerg’s research has provided empirical evidence of the “iron law of megaprojects”: cost overrun is the rule rather than the exception — nine out of ten megaprojects exceed their budget, and the average overrun in real terms reaches 44.7 % in rail projects and 20.4 % in road projects (Flyvbjerg, 2014, p. 9). In information technology projects the average overrun is smaller — about 27 % (Flyvbjerg & Budzier, 2011) — yet one project in six deviates so far that its costs exceed the estimate by an average of 200 % (Flyvbjerg, 2014, p. 10). These figures show that the planning fallacy is not a random inaccuracy: it is systematic and directed one way (the optimistic one).

Anchoring. The information received first has a disproportionate influence on later estimates. If a specific sales target or budget figure was set at the start of planning, subsequent adjustments are usually too small and remain too close to the initial anchor. This is particularly dangerous in iterative planning, where the figures of the previous cycle become the “anchors” of the new one.

The sunk cost fallacy. The tendency to continue projects because “so much has already been invested”, although a rational decision should depend only on future cash flows, irrespective of past investment (Arkes & Blumer, 1985). In business planning this error appears when an organisation refuses to abandon an unpromising project because the resources already committed create an emotional commitment.

Groupthink. Collective planning can reinforce biases where the members of a group seek consensus rather than the truth — critical voices are silenced, alternative scenarios are not discussed, and the optimistic scenario becomes the “shared understanding” (Janis, 1972). The corrective mechanism is pre-mortem analysis: the team imagines that the project has failed and works backwards, identifying the logical causes of failure. This legitimises scepticism and helps to identify risks that are usually left unspoken because of social pressure.

These biases have a fundamental implication for this book: even the most perfect planning model cannot eliminate human errors of thinking. A planning process must therefore include not only analytical tools (formulas, models) but also behavioural safeguards — independent testing of assumptions, comparison with reference classes, pre-mortem analysis and a “devil’s advocate” within the planning team.

The epistemological limits of forecasting#

Forecasting is a necessary precondition of planning — without forecasts it is impossible to model future scenarios. The accuracy of forecasting, however, has fundamental epistemological limits, and these must be understood if the results of planning are to be interpreted correctly.

A structural break is the moment at which the regularities of the past cease to hold: a sudden shock in energy prices, regulatory change, a technological upheaval, a pandemic. Contemporary econometric models attempt to identify breaks in real time by using Bayes’ rule to update parameters, yet it is in principle impossible to forecast reliably something that has never happened before — a “black swan”, in Nassim Taleb’s (2007) term.

Lessons from the forecasting competitions. Makridakis’s series of forecasting competitions (M1–M5, 1982–2020) produced four significant empirical insights.

1. Simpler methods often outperform more complex ones.
2. Combinations of statistical and judgemental methods outperform individual methods.
3. The complexity of a model is not directly related to its accuracy.
4. Machine learning methods first showed a significant advantage only in the M4 and M5 competitions: M4 was won by hybrid solutions combining neural networks with statistical models (Makridakis et al., 2020), and M5 by purely machine learning methods (Makridakis et al., 2022).

For business planning this means that accuracy should be expected not from a more complex model but from a better process — combining several methods, reviewing regularly, and producing interval rather than point forecasts.

Philip Tetlock examined the accuracy of the forecasts made by political and economic experts and found that the average expert forecasts about as accurately as a dart-throwing chimpanzee — particularly when the question concerns a horizon of three to five years; the average, as the author himself stresses, conceals differences: some experts had modest but real foresight (Tetlock & Gardner, 2015, pp. 5, 19, 68). A rare category of “superforecasters”, however, achieves consistently higher accuracy. Their common features are intellectual humility, the continuous updating of probabilities (Bayesian logic), the search for information from a variety of sources and attention to base rates. For business planning this implies that the quality of a forecast depends more on the forecaster’s style of thinking than on the model chosen.

Extrapolation without structural analysis is the commonest and potentially the most dangerous error in forecasting. Projecting current trends directly into the future without regard to structural change offers great advantages for operational purposes, but over the long term it can become the greatest obstacle to success (Makridakis et al., 1998, pp. 456, 568). In business plans this error appears when the growth of sales is extrapolated straight from past data without regard to market saturation, competitive response or macroeconomic cycles.

These limits do not mean that forecasting is pointless — they mean that the results of forecasting must be interpreted as ranges with probabilities rather than as exact figures. A business plan that presents a single scenario with exact figures (“revenue after five years will be EUR 2 347 521”) is epistemologically incorrect — not because the figure is necessarily wrong, but because it conceals the uncertainty that is an inseparable property of the future. This is the problem of spurious precision, which this book criticises consistently.

Information asymmetry and strategic misrepresentation#

The problems discussed so far — cognitive biases and the limits of forecasting — are unintentional errors of planning. In practice, however, deliberate distortion of planning also exists, and it arises from information asymmetry.

George Akerlof (1970), in his famous article on “lemons”, showed how information asymmetry between the parties to a transaction creates adverse selection; he mentions the related notion of moral hazard with reference to Arrow’s (1963) work on markets for medical care. In business planning this appears in several forms.

• The plan as a “sales document”. Where a business plan is prepared for investors, for a bank or for an application (for example for support administered by the National Paying Agency), its author has an incentive to present an optimistic picture — not because of deliberate deception, but because the planning process is naturally optimised for its addressee. The investor knows this and applies a discount — but the size of the discount is a guess, not an analysis.
• Internal information asymmetry. Managers at lower levels may withhold information from higher levels in order to obtain easier targets or larger budgets. A sales department may understate its forecast in order to “beat” the target later; a production department may demand larger inventories “for safety”. This internal asymmetry distorts the planning process systematically.
• Strategic misrepresentation. Flyvbjerg (2014) distinguished cognitive errors (unconscious optimism bias) from the deliberate manipulation of forecasts; he had earlier named the latter the “Machiavelli factor” (Flyvbjerg, 2003, p. 121). Cost overruns in megaprojects do not always arise from cognitive errors — often the initiators of a project deliberately understate costs so that the project will be approved, in the expectation that a project once started will be hard to stop.

The problem of information asymmetry cannot be solved by better models alone — it requires institutional safeguards: independent appraisal, audit, the testing of assumptions against reference classes, and clear mechanisms of accountability for the quality of assumptions. These aspects are examined in more detail in Chapter 3 (business risk) and Chapter 10 (the appraisal of investment decisions).

Criticism of the models of integration: the BSC and the limits of the planning process#

The Balanced Scorecard, presented in section 2.1.2, has attracted significant academic criticism. Hanne Nørreklit (2000, 2003) showed that the relationships the BSC postulates between the four perspectives — learning → processes → customers → finance — are not causal: they rest on financial calculus and are therefore logical rather than causal (Nørreklit, 2003, pp. 615–616). Empirical research shows, in addition, that the relationship between customer satisfaction and future performance is non-linear, and that firm-level results vary by industry — in some industries the relationship is positive, in others negative or insignificant (Ittner & Larcker, 2001, p. 370). From a planning point of view it also matters that the effect is lagged: an investment in learning may produce a financial result only after several years. The BSC risks becoming a “pseudo-scientific” instrument if its users accept the postulated causal links as empirical facts rather than as hypotheses to be tested.

Similarly, S&OP and IBP processes — although conceptually well founded — are implemented with very unequal degrees of maturity. Case studies show that cross-functional integration often remains limited: formally a “single set of numbers” is pursued, but sales managers go on telephoning production directly and some production managers prepare their own forecasts; the executive meeting turns into a forum in which everything is discussed, so that the most important questions are left aside (Dreyer et al., 2018; Ivert & Jonsson, 2014, p. 11). This shows that the effectiveness of a planning system depends not on the choice of model but on the organisation’s ability to implement it — it is a problem of process rather than of content.

The problem of planning nervousness should also be mentioned: rolling horizon procedures, which in theory ensure adaptiveness, may in practice increase the variability of plans. Studies that model rolling horizon planning with chance constraints show that the stability of a plan has to be secured by separate means (Lin & Uzsoy, 2016); in practice this is done with “frozen zones”, tolerance bands and decision priorities — without them, continuous change creates chaos rather than flexibility. Iteration is a virtue only where the frequency and the consequences of change are under control.

Learning from mistakes: single-loop and double-loop learning#

The planning process does not only generate decisions — it should also generate learning. Chris Argyris and Donald Schön (1978), and later Peter Senge (1990) in The Fifth Discipline, distinguished two types of learning.

• Single-loop learning: correcting errors without changing fundamental assumptions. An example: the sales forecast was too optimistic → next time we forecast more conservatively. This is operational but limited learning.
• Double-loop learning: questioning the fundamental assumptions, the objectives and the normative foundations. An example: the sales forecast was too optimistic → is our forecasting process adequate? Is our model of the market correct? Does our planning culture tolerate the communication of bad news? This is deep learning, which changes not only actions but also mental models.

Contemporary evaluation should aim at double-loop learning — not merely correcting errors on the basis of existing assumptions, but questioning the assumptions themselves. This is consistent with the position of this book: the value of planning arises not from the document but from the process in which an organisation learns continuously about the logic of its business and about its environment.

A concluding discussion: what does a “good” business plan mean?#

The problems discussed in this section — cognitive biases, the limits of forecasting, information asymmetry, the shortcomings of the models of integration — may give the impression that planning is a hopeless undertaking. Such a conclusion would be wrong. Planning is necessary even when it is imperfect; the question is not “whether to plan?” but “how to plan honestly?”.

A “good” business plan, in the view taken by this book, has the following properties.

1. The assumptions are stated explicitly rather than hidden in the depths of the model.
2. The forecasts are given as intervals — a range is presented rather than a single figure.
3. Sensitivity has been examined — the factors on which the result depends most have been identified.
4. The alternatives have been compared — not only a “base scenario” but also alternative courses of decision.
5. The risks have been identified and assessed rather than ignored or concealed beneath optimistic assumptions.
6. The process is iterative — the plan is reviewed and adjusted regularly in the light of new information.
7. Learning is institutionalised — the organisation has a mechanism by which errors in assumptions become a source of better understanding.

These seven properties are not a portrait of the ideal plan but the minimum threshold of quality below which planning becomes dangerous — not because it provides no information, but because it provides misleading information with the appearance of certainty.

How to determine the level of uncertainty#

The proposition that the planning method must match the level of uncertainty remains empty until it is said how that level is to be determined. In practice a sequence of two questions helps.

First — what kind of uncertainty predominates? It is useful to distinguish three layers.

• Parameter uncertainty. How the relationship works is known, but the exact values are not (what the price will be, what the yield will be). This is the weakest layer: it is managed with scenarios, ranges and sensitivity analysis.
• Model uncertainty. The very form of the relationship is in doubt: do sales depend on the marketing budget linearly at all, or does saturation set in? Here the comparison of several models and the measurement of forecast accuracy help.
• Structural uncertainty. Not only the values are unknown but also the possible outcomes: the market may disappear, the technology may change fundamentally. Neither scenarios nor sensitivity tables help here, because what is calculated is only what can already be imagined.

Second — what do the data show, and what is the cost of error? Observed variability can be measured (the dispersion of an indicator, forecast error, the clarity of a trend), but historical calm does not remove structural uncertainty — it merely shows that no break has occurred yet. Two qualitative questions are therefore always added to the quantitative estimate: is the decision reversible, and how much does an experiment cost? Where a decision is irreversible and an experiment is expensive, analytical planning remains superior even under high variability; where the opposite holds, it is cheaper to test than to forecast.

Figure 2.3. Choosing a planning approach according to the level of uncertainty and the reversibility of the decision

graph TD
    K["What kind of uncertainty?"]
    K -->|"Parameter:<br/>the relationship is known,<br/>the values are not"| P["Analytical planning<br/><i>Forecast + scenarios<br/>+ sensitivity analysis</i>"]
    K -->|"Model:<br/>the form of the relationship is in doubt"| M["Scenario planning<br/><i>Several models,<br/>measurement of forecast accuracy</i>"]
    K -->|"Structural:<br/>the outcomes are unknown"| S["Adaptive planning<br/><i>Experiment,<br/>short cycles</i>"]

    P --> G{"Is the decision<br/>reversible?"}
    M --> G
    S --> G
    G -->|"No: construction, a licence,<br/>a long-term contract"| A["Analysis before action<br/><i>Staging, options</i>"]
    G -->|"Yes: a cheap test"| B["Action before analysis<br/><i>Learning from the result</i>"]

    style K fill:#2c3e7a,color:#fff
    style P fill:#2980b9,color:#fff
    style M fill:#27ae60,color:#fff
    style S fill:#e67e22,color:#fff
    style G fill:#8e44ad,color:#fff
    style A fill:#7f8c8d,color:#fff
    style B fill:#7f8c8d,color:#fff

The diagram does not give an automatic answer, and it should not: it shows which two questions have to be answered before a planning approach is chosen. It is this logic that underlies the extended exercise in section 2.6, in which the student has to justify why the planning conception assigned to them suits a particular business.

2.1.4 The technology perspective#

The practice of business planning is one of the areas of management most transformed by technology. Each technological wave has not only provided new tools but also changed the very paradigm of planning — who plans, how, at what speed and with what precision.

The era of computerisation (1970–1995): the spreadsheet as a revolution#

Before electronic spreadsheets, the “financial part” of a business plan was a table drawn by hand on paper. Every change in an assumption required the whole model to be recalculated — a process taking hours or days. In practice this meant that one scenario was built — the “base” one — while sensitivity analysis was a theoretical notion rather than a working tool.

VisiCalc (1979), Lotus 1-2-3 (1983) and Microsoft Excel (1985 for the Macintosh, 1987 for Windows) changed that situation fundamentally. The spreadsheet made interactive modelling possible for the first time: change an assumption and see the effect on the whole model instantly. This was a change not only of speed but of the paradigm of thinking. In the words of VisiCalc’s creator Daniel Bricklin, “what if?” analysis became available to all, not just to big corporations with staffs of clerks and MBAs (Bricklin, 2009, p. 21); planning turned into the continuous testing of assumptions rather than the one-off drawing up of a table.

At the same time, ERP (Enterprise Resource Planning) systems — beginning with MRP (Material Requirements Planning), which spread in the 1970s, and culminating in SAP R/3, released in 1992 — brought organisational information together. Planning became an integrated computerised process linking production, finance and sales. In Lithuania and the Baltic states the prominence of regional and localised ERP solutions such as Rivilė (a Lithuanian accounting and ERP vendor) and Microsoft Dynamics 365 Business Central (formerly Navision) became particularly marked in shaping the financial reporting of small and medium-sized enterprises. Such integration demonstrated that a realistic growth forecast requires the operational interconnection of all processes in a database, rather than a separate spreadsheet file on the computers of individual managers.

Over the past decade, the same simplicity has been carried further by no-code and low-code solutions in the cloud. Budgeting platforms such as Anaplan and Adaptive Insights have allowed managers without a technical background to model business scenarios without programming knowledge, turning planning into a continuous process rather than a periodic reporting session. Lithuania’s electronic platforms have also been renewed — in particular the systems of the NMA and of the EU structural funds — so that business plans are now appraised by formal error-blocking rules (integrity algorithms). The requirements of the state left no room for narrative elegance: support was awarded to plans whose argument was structured as a model and in mathematical terms, verifiable by automated electronic checks.

The era of the internet and data (1995–2020): from the model to the data#

The internet transformed planning in two directions. First, the availability of information: market data, the financial statements of competitors and macroeconomic forecasts became accessible not only to large corporations but also to small firms and start-ups. This democratised the quality of planning, but it also created the problem of information overload.

Second, speed and iteration. Cloud computing allowed teams to work on the same plan in real time. Google Sheets and similar tools fundamentally changed the sociology of planning — from a single analyst producing a “planning book” to a team process in which several participants adjust assumptions, scenarios and models at the same time.

This era also gave rise to the practice of agile planning. The Lean Startup methodology (Ries, 2011), discussed in section 2.1.1, is not only an intellectual but also a technological innovation — it became possible once data on customer behaviour began to be collected in real time through digital platforms. A/B testing, cohort analysis and growth hacking are planning methods born of technological infrastructure rather than of academic theory. At the same time, business intelligence (BI) tools such as Tableau and Power BI moved planning from static tables to continuously updated dashboards, while cloud planning platforms such as Anaplan and Adaptive Insights made it possible to turn budgeting and forecasting into a continuous process rather than an annual ceremony. The most recent tendency in this field is the concept of Extended Planning and Analysis (xP&A), formulated by the analyst firm Gartner (2020), which brings financial and operational planning together into a single data-driven process.

The era of data, however, also intensified certain risks. The problem of spurious precision (see section 2.1.3) became sharper: when a model is fed large volumes of data, its results look more “scientific”, although the fundamental assumptions may be just as inaccurate as in a manual analysis. Large volumes of data do not compensate for poor conceptual logic.

The era of AI and AI agents (2020→): planning as a dialogue between human and machine#

The integration of artificial intelligence (AI) into business planning opens up a new paradigm — planning as a dialogue between human and machine rather than a human act alone.

The transformation of forecasting. Although traditional statistical methods were long thought sufficient for time series forecasting, the M4 and M5 forecasting competitions showed otherwise. In the M5 experiment, machine learning (ML) models significantly outperformed the best classical methods: the advantage of the top five teams over the benchmark method exceeded 20 % (Makridakis, Spiliotis, & Assimakopoulos, 2022). This progress confirms that automated algorithms can achieve greater forecast accuracy in time series tasks — demand forecasting, cash flow planning and similar areas — than traditional linear models. Nevertheless, where no validated statistical model exists — structural breaks, geopolitical wars, unexpected amendments to legislation — the most robust arrangement proves to be a combination of human and algorithm rather than full autonomy for either side; beyond the forecasting horizon all that remains is to plan for adaptation and resilience (Tetlock & Gardner, 2015, pp. 21, 23, 244).

AI agents as planning partners. Generative AI models are able to produce initial drafts of a business plan, generate scenarios, check the internal consistency of assumptions and carry out sensitivity analysis. This means that the role of the analyst is changing: from “model builder” to “model curator” — the human formulates the questions, assesses the assumptions and takes the decisions, while the AI performs the calculations and generates alternatives.

The democratisation of planning. AI lowers the barrier to professional business planning — small firms that previously could not afford a team of analysts can obtain better forecasts and scenario analyses. This is particularly relevant in the Lithuanian and Baltic context, which is dominated by small and medium-sized enterprises. An analogous logic can be seen in public sector planning processes, in the systems for electronic applications, reporting and appraisal: the digitalisation of the National Paying Agency (NMA) and of other bodies administering support, for example, changed not only the speed of submitting documents but also the architecture of expectations itself — a plan must increasingly be founded on structured data rather than on a textual narrative alone.

The era of AI, however, also intensifies the epistemological problems already discussed. Optimism bias (see section 2.1.3) may be reinforced if AI models are trained on historical data dominated by successful cases (survivorship bias). The risk of spurious precision increases further: a complex AI model presenting a result to several decimal places creates a stronger illusion of being “scientific” than a simple spreadsheet model, although both rest on the same uncertain assumptions. Finally, there is the problem of accountability: when a business plan is produced with the help of an AI assistant, who is responsible for the quality of the assumptions — the human or the algorithm?

The position of this book is that AI is a powerful instrument of planning but not a subject of planning. Ultimate responsibility for the assumptions, for the choice of scenarios and for taking decisions remains with the human. A deeper analysis of the relationship between technology and business planning is given in the final chapter of the book.

2.1.5 The Lithuanian context#

The paradigms and models discussed so far were formulated by observing large organisations in large economies. The structure of business in Lithuania is different: micro-enterprises predominate, and in them planning rarely has a separate function — often it does not even have a separate document. This does not mean that no planning takes place; it means that planning frequently happens in the manager’s head rather than on paper, and that a formal plan appears when somebody outside the firm demands one. No representative Lithuanian study measuring this relationship exists, so this is a working assumption rather than an established fact. In the Lithuanian context, therefore, the key question is not “whether to plan” but who makes the plan be written down, and when.

The plan as an external requirement. The three typical cases in which a small Lithuanian business prepares a formal plan are an application for support, a loan application and the raising of investment. The data below come from three specific sets of publicly available requirements — this is not a survey of the whole market but three verified examples.

One thing worth noticing before a first plan is drawn up emerges from this table: none of these three sets of requirements examined calls for scenarios or for sensitivity analysis. What is checked is viability at a single point — whether the ratios fall within the limits if everything happens as written. The robustness of the decision, that is, what would happen if the assumptions of the plan were not borne out, is not formally examined in these requirements. This does not mean that no financier asks for it; it means that no such requirement appears in the three sets of documents examined.

That observation supports a proposition related to section 2.1.3. If all that is assessed is whether the ratios fall within the required limits, it is rational for an applicant to submit a plan whose ratios do fall within them — not necessarily the plan that best represents reality. In such a system the sixth proposition of section 2.3 — that a good plan is not one whose forecasts are accurate but one in which the assumptions are stated and the sensitivity examined — ceases to be academic: the plan that passes the appraisal and the plan that helps the decision are not always the same document. At the same time this is an opportunity: an applicant who adds scenarios and a sensitivity analysis on their own initiative shows the appraiser not better figures but a better understanding of the assumptions.

What Lithuanian forecasting experience teaches. In December 2021 the Bank of Lithuania forecast average annual inflation for 2022, measured by the harmonised index of consumer prices (HICP), at 5.1 %, and the Ministry of Finance at 4.0 %. Actual inflation in 2022 reached 18.9 % (Bank of Lithuania, 2021, 2023; Ministry of Finance of the Republic of Lithuania, 2021). The forecast deviations were therefore about 13.8 and 14.9 percentage points respectively.

This figure should not be read as evidence that the institutions forecast badly. The energy price shock of 2022 occurred after both forecasts had been published, so this is a single shock year rather than a typical deviation; calculating a typical error would require a consistently defined series covering a decade, and no such published analysis could be found in the publications of the Bank of Lithuania. The lesson is a different one, and it corresponds exactly to the structural break argument of section 2.1.3: a forecast holds only as long as its conditions hold. In a business plan this means that an assumption is worth writing down together with the condition whose change would invalidate it — as is done in a review condition (Appendix A, A.4).

What Lithuanian project experience teaches. An audit by the National Audit Office of Lithuania (Valstybės kontrolė) in 2025 reviewed 16 real estate renovation projects completed by the state enterprise Turto bankas in 2021–2024. In all sixteen the final cost exceeded the amount originally planned: the total value of the projects rose from EUR 29.11 million to EUR 64.93 million, and nine of the sixteen also exceeded the permitted deviation from the completion deadline (National Audit Office of Lithuania, 2025).

In interpreting these data it is important not to skip a step. The audit records a deviation from the plan but does not measure its cause. Optimism bias (section 2.1.3) is one possible cause; others are the strategic understating of price at the application stage, the duration of procurement procedures, or genuinely unforeseen work. From the point of view of this book what matters is not the attribution of blame but the observation that in this sample all the deviations went the same way — not one project cost less than planned. Whether that also holds beyond this sample the audit does not answer. In international research the one-directional deviation is recorded consistently (Flyvbjerg, 2014), and the Lithuanian audit data are compatible with that conclusion, although they do not confirm it. That is precisely why the safeguards described in section 2.1.3 — reference class forecasting, the independent testing of assumptions, the pre-mortem — are not an academic exercise.

A research gap worth knowing about. The proposition in this chapter that the most effective organisations combine formal planning with adaptation rests on Brews and Hunt’s (1999) study of 656 firms, in which the association with performance appeared only after four or more years of formal planning. No replication of that study in Lithuanian or other Baltic firms could be found. The international results are not unambiguous either: Schwenk and Shrader’s (1993) meta-analysis finds a positive, though small, relationship in small firms; in Kraus and colleagues’ (2006) study of 290 small Austrian enterprises, more formal planning was associated with a higher probability of growth; and in Gică and Negrusa’s (2011) sample of 200 Romanian SMEs the overall relationship between planning and performance was not confirmed.

A practical rule follows for the student, and it applies beyond this chapter: when an international conclusion is transferred to the Lithuanian context, it becomes a hypothesis rather than a fact. That does not mean it cannot be relied on; it means that in relying on it one must state the conditions under which it would hold and what would refute it. The planning practice of small business in Lithuania has been little covered by research; a survey of firms in the Kaunas and Šiauliai regions by Gudonavičius and colleagues (2009) shows that the attention managers pay to strategy varies widely — 14.3 % are constantly concerned with it, 49 % often, 24.5 % rarely, and 12.2 % do not implement it at all — but there is as yet no more recent representative study measuring both the formality of planning and performance. This is not a shortcoming to be concealed but an open field of research, and one of the reasons why the exercises in section 2.6 ask the student not to repeat a proposition from the literature but to test it in a specific case.

2.2 Analytical models and formulas#

This section presents the mathematical and analytical tools that turn the theoretical part into practice. The formulas cover three stages of the planning process: forecasting (how to project future values), modelling (how to structure assumptions) and evaluation (how to compare alternatives and assess uncertainty). Each formula is given with an explanation of its variables and a calculation example.

Data for assumptions: where they come from#

Every formula set out below requires data, and in planning practice the most common error is one not of calculation but of source: an assumption is taken from the first figure found on the internet and becomes the foundation of the whole model. Four layers of data are used in the exercises in this book and in a real business plan.

Three rules apply to work with any of these layers. First, every figure has a year — an “average price of EUR 200 per tonne” without a year is worthless. Second, the definition of an indicator matters more than its value: the average wage “gross” and “net” differ by a third, while inflation measured by the national index and by the harmonised index differs by a different amount again. Third, a secondary source is worth tracing back to the primary one: a figure given in a sector review often comes from official statistics, and it is the statistics that should be cited.

These rules are not an academic formality. The appraiser — a bank, a support agency or an investor — checks the soundness of the assumptions before the accuracy of the calculation (section 2.1.5), and in a student’s work the origin of the assumptions is one of the assessment criteria (Appendix A, section A.6).

Measuring growth rates: CAGR#

One of the most common indicators in business planning is the compound annual growth rate (CAGR), which makes it possible to compare growth dynamics across different periods on a single standard:

where:

• $V_n$ — the final value (for example, revenue in the last year),
• $V_0$ — the initial value (for example, revenue in the first year),
• $n$ — the number of periods (years).

Numerical example. Context. The manager of a food processing company is preparing a loan application and has to state the rate at which the company has grown: in the application this figure becomes the principal growth assumption, so the bank will ask about it first. Data (the company’s annual reports; a teaching example, with magnitudes corresponding to the scale of a typical Lithuanian processing company): revenue in 2020 — EUR 500 thousand, in 2025 — EUR 820 thousand. Calculation (CAGR over 5 years):

This means that revenue grew on average by 10.4 % a year on a compound basis. The advantage of CAGR over a simple arithmetic average is that it takes the cumulative effect (compound interest) into account. CAGR does, however, have a fundamental limitation: it shows only the starting and end points and conceals the dynamics in between — if revenue over those 5 years fluctuated between EUR 300 thousand and EUR 900 thousand, CAGR will not reflect that.

Forecasting methods#

The simple moving average#

The simplest quantitative forecasting method is the simple moving average (SMA), used for short-term forecasts when the data show no pronounced trend:

where:

• $SMA_t$ — the moving average in period $t$,
• $Y_i$ — the value observed in period $i$,
• $k$ — the number of periods included in the average (the window size).

The forecast for the next period is the latest moving average: $F_{t+1} = SMA_t$.

Numerical example. Context. The purchasing manager of a trading company has to place the May order with a supplier by the end of the month: too large an order will tie up cash in inventory, too small an order will leave the shelves empty. Data (the company’s sales records; a teaching example), in EUR thousand: January — 42, February — 38, March — 45, April — 41. Calculation — the forecast for May ($k = 3$, the average of the last 3 months):

The choice of moving average — the size of $k$ — is a trade-off: a larger $k$ smooths fluctuations more but responds to change more slowly; a smaller $k$ responds faster but is more sensitive to noise.

Exponential smoothing#

Exponential smoothing gives greater weight to the most recent data, allowing the forecast to adapt to change more quickly:

where:

• $F_{t+1}$ — the forecast for the coming period,
• $Y_t$ — the actual value in the current period,
• $F_t$ — the earlier forecast for the current period,
• $\alpha$ — the smoothing parameter ($0 < \alpha < 1$).

The larger $\alpha$ is, the more weight is given to the most recent data (a faster reaction and greater sensitivity); the smaller $\alpha$ is, the more “inert” the forecast becomes (a slower reaction and greater stability). In practice, values of $\alpha$ between 0.1 and 0.3 are often used.

Numerical example. Context. The same purchasing manager notices that the moving average is slow to respond to change and wants a forecast that gives more weight to the most recent month. Data (the same sales records): actual sales in April $Y_4 = 41$ EUR thousand, the earlier forecast $F_4 = 43$ EUR thousand, $\alpha = 0.2$. The forecast for May:

The forecast is adjusted downwards (from 43 to 42.6), because the actual outcome was lower than expected, but the adjustment is conservative (only 20 % of the deviation).

Linear regression#

When a trend is visible in the data, or when the statistical relationship between variables is to be assessed, linear regression is applied:

where:

• $\hat{Y}$ — the forecast value of the dependent variable,
• $X$ — the independent variable (the explanatory factor),
• $a$ — the intercept,
• $b$ — the regression coefficient (the slope) — it shows how much $Y$ changes when $X$ changes by one unit.

The regression coefficients are calculated by the method of ordinary least squares (OLS):

Numerical example. Context. The marketing manager is asking for additional funds in next year’s budget and has to justify what the company can expect in return. Data (the company’s management accounts; a teaching example) — annual revenue ($Y$, EUR thousand) and the marketing budget ($X$, EUR thousand) over the last 5 years:

Calculations: $\bar{X} = 30$, $\bar{Y} = 348$, $b = \frac{5 \times 55\,200 - 150 \times 1\,740}{5 \times 5\,500 - 22\,500} = \frac{276\,000 - 261\,000}{27\,500 - 22\,500} = \frac{15\,000}{5\,000} = 3.0$; $a = 348 - 3.0 \times 30 = 258$.

The regression equation is $\hat{Y} = 258 + 3.0X$. If the marketing budget planned for next year is EUR 45 thousand:

Interpretation: each additional thousand euros spent on marketing is associated with EUR 3 thousand of additional revenue. Regression, however, shows correlation and not causation — without further analysis it cannot be claimed that the marketing budget causes revenue to grow. Extrapolation beyond the range of the observed data (a budget of EUR 100 thousand, for example) is also progressively less reliable.

Measuring forecast accuracy#

The three methods above produce a figure, but none of them says whether it can be trusted. There is only one honest way of measuring accuracy: the method is applied to historical data already in hand and the extent of its deviation from them is examined. A method that reproduces the past poorly has no grounds for laying claim to the future.

The mean absolute error (MAE) shows by how many units the forecast on average misses the actual figure:

where $Y_t$ is the actual value for the period, $F_t$ is the forecast for the same period and $n$ is the number of comparisons. The absolute value is taken because deviations of opposite sign would otherwise cancel each other out and the method would look more accurate than it is.

Numerical example. Context. The same purchasing manager has to choose which method to rely on in the next quarter. Data (the same sales records): the moving average forecasts and the actual figures for four months, in EUR thousand: forecasts 41.0; 42.3; 41.7; 43.0, actual figures 38; 45; 41; 44. Calculation: $MAE = (|38-41.0| + |45-42.3| + |41-41.7| + |44-43.0|)/4 = (3.0 + 2.7 + 0.7 + 1.0)/4 = 1.85$ EUR thousand. Interpretation for the decision: the forecast misses by EUR 1.85 thousand on average, that is, by about 4 % of monthly turnover. If the cost of an error in the supplier order is lower than this error, the method can be relied on; if it is higher, either another method or an inventory buffer covering that error is needed. Comparing the MAE of two methods on the same data series shows which of them better reproduces the dynamics of this particular business.

The coefficient of determination ($R^2$) shows what share of the variation in the result is explained by the regression model:

where $\hat{Y}_t$ is the value forecast by the model and $\bar{Y}$ is the mean of the actual values. In the regression example (marketing budget and revenue) the sum of squared residuals is 30 and the sum of squared deviations from the mean is 9 030, so $R^2 = 1 - 30 / 9\,030 \approx 0.997$: the model explains almost all of the variation in revenue. This is because the teaching data are neatly arranged; in real data such a close fit is more a warning than a success. A more usual value, for example $R^2 = 0.82$, would mean that the model explains 82 % of the variation in revenue, while 18 % remains unexplained. Two things should be kept in mind when interpreting it: a high $R^2$ does not mean a causal relationship (see the warning accompanying the regression formula), and the thresholds used in interpretation — for example, that 0.7 is a “strong” relationship — are a convention rather than a mathematical boundary. In the Excel model used in this book, such thresholds are held in separate cells (CH02_REG_r2_stiprus, CH02_REG_r2_vidutinis) precisely so that they can be changed without touching the formulas.

The standard deviation in it is calculated for scenarios with probabilities. When the variation of a historical series is assessed, there are no probabilities: all observations are treated as equally important ($p_i = 1/n$), and the formula becomes the ordinary descriptive standard deviation. These are not the same measure — in the first case what is measured is the expected dispersion of outcomes, in the second the observed variability of the past.

Scenario analysis: the weighted expected value#

When a forecast is presented not as a single figure but as a set of several possible scenarios, a probability is assigned to each scenario on expert judgement and the weighted expected value is calculated:

where:

• $E(V)$ — the expected (weighted) value,
• $V_i$ — the value under the $i$-th scenario,
• $p_i$ — the probability of the $i$-th scenario ($\sum p_i = 1$),
• $s$ — the number of scenarios.

Numerical example. Context. The board is deciding whether to start production of a new product and is asking not for a single figure but for a range with probabilities. Data (estimates from the sales team and from production; a teaching example) — the annual revenue of the new product under three scenarios:

The expected value coincides with the base scenario (400), because the scenarios are symmetrical. The expected value on its own is not sufficient, however — it conceals the extent of the uncertainty. The dispersion of outcomes has to be assessed in addition. The standard deviation ($\sigma$) shows the spread of outcomes around the expected value:

In our example:

The coefficient of variation (CV) makes it possible to compare risk between projects of different size:

This means that the standard deviation amounts to 35 % of the expected value. The Excel model of this book applies indicative thresholds: up to 15 % — relatively low dispersion, 15–30 % — moderate, above 30 % — high (CH02_SCN_cv_maza, CH02_SCN_cv_vidutine), so 35 % falls into the area of high dispersion. These thresholds are a didactic convention rather than a universal norm: what counts as high dispersion in one sector may be usual in another, which is why they are held in separate, changeable cells in the model. If an alternative project has $CV = 15\;\%$, it is less risky (even though the absolute value of $\sigma$ may be larger if the project itself is bigger). Risk indicators are examined in more detail in Chapters 3–4.

Sensitivity analysis: the elasticity coefficient#

Sensitivity analysis establishes which assumptions have the greatest effect on the result. The principal measure of sensitivity is the elasticity coefficient:

where:

• $\varepsilon_{Y,X}$ — the elasticity of the result $Y$ with respect to the assumption $X$,
• $\Delta Y / Y$ — the relative change in the result,
• $\Delta X / X$ — the relative change in the assumption.

Interpretation: $\varepsilon = 2.5$ means that a change of 1 % in the assumption produces a change of 2.5 % in the result — high sensitivity. $\varepsilon = 0.3$ means low sensitivity.

Numerical example. Context. Before negotiations with a buyer, the finance director has to know how far the value of the project depends on price: if the dependence is strong, a discount will cost more than it appears to. Data (the model of the same project; a teaching example). In the base scenario the net present value (NPV) of the company is EUR 500 thousand at a selling price of EUR 10 per unit. If the price falls by 10 % (to EUR 9 per unit), NPV falls to EUR 325 thousand:

NPV is highly sensitive to the selling price: a change of 1 % in price produces a change of 3.5 % in NPV. This means that price is a load-bearing assumption, to which particular attention has to be paid in forecasting and in monitoring.

Sensitivity analysis is usually presented in the form of a sensitivity table (the example with WACC and the growth rate given in Chapter 1) or of a tornado chart, in which the variables are ranked by the size of their effect — from the largest to the smallest. This technique is applied in practice in section 2.5 (the Excel model).

Summary of the formulas#

Table 2.3. The formulas of Chapter 2

These formulas make up a basic set of analytical tools, sufficient for initial planning models. More advanced methods — Monte Carlo simulation, ARIMA time series and real options valuation — are presented in the relevant chapters (Chapters 4, 12 and 14).

2.3 Summary#

This chapter has analysed the business planning and evaluation process — from its intellectual origins to contemporary analytical instruments. The main claims of the chapter are the following.

1. Planning thought has evolved through four paradigms — administrative planning (Taylor, Fayol), strategic planning (Ansoff, Porter), strategic thinking (Mintzberg, Barney) and adaptive planning (Ries, Sarasvathy, McGrath) — and each paradigm added to the earlier ones rather than replacing them. Effective planning today requires elements from all four.

2. The planning method must match the level of uncertainty. A stable environment permits analytical planning; moderate uncertainty calls for scenario analysis; high uncertainty calls for an adaptive, experimental approach. There is no universally “best” planning method.

3. The planning process comprises five cyclical stages: identifying assumptions, forecasting, modelling, evaluation and decision. The process is iterative — not a one-off document but a continuous learning cycle.

4. Evaluation is an integral part of planning rather than a separate control function. Evaluation is carried out in three time modes: ex ante (before the decision — comparing alternatives), in itinere (during implementation — monitoring the assumptions) and ex post (after implementation — learning).

5. Planning is vulnerable to systematic errors — cognitive biases (optimism bias, the planning fallacy, anchoring), the epistemological limits of forecasting (structural breaks, spurious precision) and information asymmetry (the plan as a “sales document”, strategic misrepresentation). These errors can be mitigated but not eliminated, and the process must therefore include behavioural safeguards.

6. A “good” business plan is not one whose forecasts turn out to be accurate — it is a plan in which the assumptions are clearly stated, the forecasts are given as ranges, sensitivity has been examined, alternatives have been compared, risks have been identified, the process is iterative and learning is institutionalised.

7. The value of planning comes from the process, not from the document. Arie de Geus (1988) called planning learning, and corporate planning institutional learning. An organisation that plans honestly knows more about the logic of its own business than one that does not plan — even if the forecasts of both are equally inaccurate.

Connections with other chapters. The planning cycle presented in this chapter — assumptions → forecasting → modelling → evaluation → decision — is the framework that all the chapters of the textbook go on to fill in. Its ontological foundation comes from the concept of value in Chapter 1, the measurement of risk is deepened in Chapters 3–4, and Chapter 16 realises the whole cycle in a single financial model of a business plan. The other chapters are specifications of the same cycle for particular areas — functional plans, financial analysis, investment decisions, valuation methods, implementation and exit.

2.4 Case studies and worked examples#

Case 1: A Lithuanian farm business plan for an application for support#

Case type: [hypothetical]. The situation and all the figures have been created for teaching purposes on the basis of typical practice in the Lithuanian agricultural sector and of the requirements applying to applications for support; the orders of magnitude correspond to the scale of cereal farms in the Kėdainiai district, but these are not the data of any particular farm.

The situation#

Jonas, a farmer in the Kėdainiai district, grows cereals on 120 ha. He plans to invest in a grain dryer and storage complex (a total investment of EUR 280 thousand) in order to draw on rural development support. The National Paying Agency (NMA), the body administering EU agricultural and rural support in Lithuania, requires the application to be accompanied by a business plan with financial projections demonstrating the economic viability of the investment; this example applies the typical condition of five-year projections. The specific horizon and the set of indicators depend on the measure and on the call for applications (see Section 2.1.5), but the planning logic that the case illustrates changes little from one programming period to the next.

Applying the planning process#

1. Assumptions. Jonas has to state his assumptions clearly:

Every assumption has a source or a justification. This illustrative example gives approximate figures; in a real business plan they would be based on specific public data from the State Data Agency, the Bank of Lithuania and the NMA, with precise references. The NMA appraiser checks precisely the soundness of the assumptions — not only the final figures.

2. Forecasting. The investment does not increase revenue directly — it changes the incremental cash flows: (a) the drying service no longer has to be bought in from outside, and (b) once storage is available, the grain can be sold later rather than at the harvest peak, and so at a seasonality premium. In the base scenario:

• yield: 120 ha × 5.5 t/ha = 660 t; the value of output is 660 t × EUR 200/t = EUR 132 thousand;
• drying costs saved: (15 − 4) EUR/t × 660 t = EUR 7.26 thousand (the operating costs of the farm’s own dryer are deducted from the price of the service);
• the seasonality premium: 10 % × EUR 132 thousand = EUR 13.2 thousand (the midpoint of the assumed range of 8–12 %);
• the annual incremental flow: 7.26 + 13.2 = EUR 20.46 thousand, growing by 2 % each year together with prices.

This brings out a distinction that those who prepare applications often overlook: the application form requires five-year projections (the condition assumed in this example), but the viability of the investment is appraised over the useful life of the asset. A five-year window without a residual value systematically understates such an investment, so the calculation here is made for ten years, with the residual value conservatively set at zero.

3. Modelling. The own contribution after support of 50 % is 280 × 0.5 = EUR 140 thousand ($t_0$). The flows are discounted at 6 % for ten years; the sum of the annual growth-to-discount ratios gives the multiplier $\sum_{t=1}^{10} (1.02/1.06)^t = 8.14$. The results of the three scenarios are as follows:

A check on the base scenario: $-140 + 20.46 \times 8.14 = -140 + 166.6 = +26.6 \text{ EUR thousand}$. In the pessimistic scenario the NPV is negative — the investment would not pay back. That is not a reason to abandon the project; it is information that the success of the project depends on yield and on price — on the load-bearing assumptions.

4. Evaluation (ex ante). Consulting an agronomist and drawing on the climate statistics of the district, Jonas assigns probabilities: he puts the pessimistic scenario (a year of drought or of heavy rain) at 20 %, the base scenario (close to the historical average) at 60 % and the optimistic scenario (favourable conditions and higher prices) at 20 %. The weighted expected NPV is:

A positive expected NPV shows that the investment is economically justified. Sensitivity is tested by changing one assumption at a time: raising only the price by 10 % (from EUR 200 to EUR 220/t, leaving the yield at 5.5 t/ha) lifts the NPV to EUR 37.3 thousand, that is by 40 %; the elasticity is $\varepsilon = 40 / 10 = 4.0$. A change in the yield alone, from 5.5 to 6.0 t/ha (+9.1 %), has an even stronger effect (an elasticity of 6.2). Such sensitivity is not a calculation error: the NPV is the difference between two large quantities, so a small change in the flow changes it several times more strongly. That is precisely why the scenario table, in which all the assumptions are changed at the same time, is not suitable for measuring elasticity — it shows the overall outcome rather than the influence of an individual factor.

5. Decision. Jonas submits the application with three scenarios, clearly stating the assumptions and the sensitivity analysis. The NMA appraiser sees not only a “nice number” but also the author’s understanding of the limits of the assumptions.

Conclusions#

• A business plan for an application for support is not a “bureaucratic document” — it is a structured decision model in which every assumption must be justified and verifiable.
• A three-scenario model is the minimum quality threshold — a single-scenario plan conceals risk.
• What matters most to the NMA appraiser is not the final figure but the soundness of the assumptions and the author’s ability to identify the load-bearing assumptions.
• In the farming sector the principal risks — yield and the purchase price — are partly beyond its control, so sensitivity analysis is a necessary part of the plan rather than an optional addition.

Case 2: Planning in a start-up — the traditional and the “lean” approaches#

Case type: [hypothetical]. The situation and all the figures have been created for teaching purposes on the basis of typical practice among agricultural technology start-ups in the Baltic region; the indicators correspond to the scale of early-stage technology companies, but these are not the data of any particular company.

The situation#

DataHarvest, a technology start-up based in Vilnius, is developing an artificial intelligence (AI) platform for agricultural forecasting — on the basis of satellite data it forecasts yields, the probability of disease and the optimal time for fertilisation. The team (3 programmers and 1 agronomist) is seeking an investment of EUR 500 thousand for expansion.

The traditional planning approach#

A business plan with 5-year financial projections is prepared for the investor:

CAGR = $(2\,880/48)^{1/4} - 1 = 178\;\%$ per year. The plan looks impressive, but it raises questions:

• Is growth from 20 to 1 200 clients in 5 years realistic for a start-up at the beginning of its life?
• What is the conversion rate from free trials to paying clients?
• Has the pricing (EUR 200/month) been tested in the market?
• What are the customer acquisition costs (CAC)?
• What is the customer churn rate?

Each of these questions is an unverified assumption. The traditional plan “hid” them behind the figures.

The “Lean Startup” approach#

Instead of 5-year projections, DataHarvest identifies three critical hypotheses:

1. The value hypothesis: are farmers willing to pay EUR 200/month for AI yield forecasts? Test: a 30-day free trial for 15 farms; the success criterion is a conversion to the paid version of ≥60 %.
2. The growth hypothesis: do satisfied clients recommend the product to others? Test: the Net Promoter Score (NPS) after 3 months of use; the success criterion is an NPS ≥ 40.
3. The technological hypothesis: are the AI forecasts accurate enough to create value? Test: a comparison with actual data from the 2024 season; the success criterion is a forecast error ≤ 15 %.

Each hypothesis is tested over 2–3 months on a minimal budget. A financial model is prepared only once the hypotheses have been confirmed — and it is then based on empirical data rather than on guesses.

A comparison of the approaches#

Conclusions#

• Both approaches have their place. A traditional financial model is needed for mature companies and for formal appraisal (the NMA, a bank, an investor). The lean approach is for start-ups and for the development of new products.
• A start-up’s plan should begin with hypotheses rather than with projections. The figures come after the hypotheses have been confirmed, not before.
• Investors increasingly assess not the “beauty” of the plan but the team’s ability to learn — how quickly it identifies mistakes and changes direction (a pivot).
• From the standpoint of this textbook, one logic connects both cases: decision modelling and analytical evaluation — only the speed of iteration and the way the assumptions are tested differ.

2.5 Building the Excel model#

Objective#

The student builds a working forecasting and scenario analysis model in which all the formulas of section 2.2 are implemented: the growth rate; the simple moving average and exponential smoothing with measurement of the forecast error; linear regression with the coefficient of determination; the weighted expected value of three scenarios together with its dispersion; and sensitivity analysis with elasticity coefficients. The logic of the model follows the sequence of the solution to the first case in section 2.4: from the assumptions to sensitivity.

The architecture of the model#

The model consists of two layers. This division is not a technical convenience: the formula sheets are intended for understanding one instrument at a time in isolation, while the working sheets are for applying them together to a single decision.

Every sheet is built from the same six blocks: A the heading and the description of the variables, B the problem statement (why this calculation is needed), C the data, D the calculations, E the visualisation, F the interpretation. For the student this means that the same things can be looked for on every sheet: where the inputs are, where the formula is, where the conclusion is.

Named ranges. All inputs and results are named; the formulas refer to names, not to cell addresses. The model contains 113 of them, and all of them begin with CH02_ (for example, CH02_CAGR_Result, CH02_SCN_psuma). The rule is simple: if a value is going to be used in another formula, it must have a name. The formula =(CH02_Y2025/CH02_Y2021)^(1/CH02_CAGR_n)-1 states what it computes; =(C16/C12)^(1/C24)-1 does not. (In the published model the sheet and range names are in Lithuanian: Įvadas — introduction, Prognozavimas — forecasting, Regresija — regression, Scenarijai — scenarios, Jautrumas — sensitivity; psuma — sum of probabilities, vidurkis — mean, alfa — alpha, maza/vidutine — low/medium, stiprus/vidutinis — strong/moderate, Metai_pirmi/Metai_paskutiniai — first/last year.) When a row is moved, a named range remains correct, whereas a direct reference may silently point to the wrong cell.

Step-by-step instructions#

At the end of every step there is a check point — a control question whose negative answer means an error in the model rather than bad data.

Step 1. The growth rate (the “CAGR” sheet)#

Revenue for five years (CH02_Y2021–CH02_Y2025) and the boundaries of the period (CH02_Metai_pirmi, CH02_Metai_paskutiniai) are entered. The number of periods is calculated, not typed in: CH02_CAGR_n = CH02_Metai_paskutiniai − CH02_Metai_pirmi. The result is CH02_CAGR_Result under formula (2.1). Alongside it, the arithmetic mean of the annual changes (CH02_CAGR_vidurkis) is calculated, and benchmarks are given: the sector average (CH02_CAGR_sektorius) and GDP growth (CH02_CAGR_bvp).

Check. Where the data fluctuate, CAGR must be lower than the arithmetic mean. If the two coincide, the wrong cell is probably being calculated, or n is wrong.

Step 2. The forecast and its accuracy (the “Prognozavimas” sheet)#

The controls are set: the moving average window CH02_PRG_k, the smoothing coefficient CH02_PRG_alfa and the initial forecast CH02_PRG_F1. Two forecasts are calculated — CH02_SMA_Prognoze (formula 2.2) and CH02_EG_Prognoze (formula 2.3). The mean absolute error is measured for both: CH02_PRG_mae_sma and CH02_PRG_mae_eg.

Measuring the error here is not an additional ornament. The accuracy of a forecast is assessed by comparing the method with a past that has already happened: a method that reproduces history less well has no grounds for laying claim to the future.

Check. When α is raised from 0.2 to 0.5, the forecast must move closer to the most recent actual figure. If it does not move, the formula is using a hard-coded constant rather than CH02_PRG_alfa.

Step 3. The trend and its strength (the “Regresija” sheet)#

The regression coefficients CH02_REG_b (the slope of the regression line) and CH02_REG_a (the intercept) are calculated, together with the coefficient of determination CH02_REG_r2 and the forecast CH02_Regresija_Prognoze for the planned value of the independent variable (CH02_REG_planas). The thresholds for interpretation are held separately (CH02_REG_r2_stiprus, CH02_REG_r2_vidutinis) so that they can be changed without touching the formulas.

Check. If CH02_REG_r2 is lower than CH02_REG_r2_vidutinis, the forecast is presented with a reservation: the relationship is weak, so the figure indicates a direction rather than a magnitude.

Step 4. Scenarios (the “Scenarijai” sheet)#

A baseline value (CH02_SCN_R0), the growth rates of three scenarios (CH02_G_opt, CH02_G_baz, CH02_G_pes) and their probabilities (CH02_P_opt, CH02_P_baz, CH02_P_pes) are entered. The sum of the probabilities CH02_SCN_psuma is calculated separately. The results are the weighted expected value CH02_EV_2030 (formula 2.5), the standard deviation CH02_Sigma_2030 (formula 2.6) and the coefficient of variation CH02_SCN_cv.

Check. CH02_SCN_psuma must equal 1. If it does not, the expected value is meaningless — it weights by shares that do not add up to a whole.

Step 5. Sensitivity (the “Jautrumas” sheet)#

The step of change CH02_JAU_delta and the number of steps CH02_JAU_n are set. The baseline value (CH02_Jautrumas_Bazinis) is taken from the scenarios sheet, so that when an assumption is changed there, the sensitivity table recalculates by itself. The elasticity coefficients (formula 2.7) are calculated separately for each assumption: CH02_JAU_eps_g (the baseline growth rate), CH02_JAU_eps_gopt, CH02_JAU_eps_gpes and CH02_JAU_eps_p (the probability).

Check. Elasticity is measured by changing one assumption at a time. If two are changed at once (as in a scenario table where both price and volume vary), the resulting figure shows the combined effect rather than the influence of a single factor — the error illustrated by the first case in section 2.4.

How to check the whole model#

1. One input in a blue cell is changed — all the results that depend on it, including the sensitivity sheet, must recalculate.
2. No hard-coded number should remain in the formulas, apart from mathematical constants: every quantity comes from a named range.
3. Block F of every sheet must contain a conclusion expressed in a sentence, not merely a number. If a conclusion cannot be formulated, the model is calculating something that nobody asked for.

File: Excel/CH02/Ch02_Model_Master.xlsm Instructions with the blocks and formulas of all 13 sheets: Excel/CH02/Ch02_Formula_Guide_v2.md (the earlier version, Ch02_Formula_Guide.md, describes only the five working sheets and the “Įvadas” sheet)

2.6 Questions and exercises#

2.6.1 Facts and concepts (recognition level)#

Question 2.1 [recognition]. The student lists the four planning paradigms set out in Table 2.1 and gives one author and one period for each of them.

Question 2.2 [recognition]. The student names the five stages of the planning process (section 2.1.2) and indicates in which of them the assumptions are formulated and in which the decision is taken.

2.6.2 Explanations (comprehension level)#

Question 2.3 [comprehension]. The student explains how ex ante, in itinere and ex post evaluation differ, and why all three are needed in the planning cycle.

Question 2.4 [comprehension]. The student explains why the compound annual growth rate (CAGR) is better suited to measuring growth than the arithmetic mean of the annual growth rates.

2.6.3 Computational exercises (application level)#

Question 2.5 [application]. CAGR. A company’s revenue in 2020 was EUR 1 200 thousand and in 2025 EUR 1 850 thousand. The student calculates CAGR. If this trend continued, what would revenue be in 2030? What assumptions does the student make in extrapolating?

Question 2.6 [application]. Scenario analysis. The investment in a new product is EUR 150 thousand. Three scenarios:

• Optimistic (p = 0.25): net profit over 3 years — EUR 250 thousand
• Base case (p = 0.50): net profit over 3 years — EUR 120 thousand
• Pessimistic (p = 0.25): net profit over 3 years — EUR 30 thousand

The student calculates: (a) the weighted expected value, (b) the standard deviation, (c) the coefficient of variation. Is the investment justified?

Question 2.7 [application]. Sensitivity analysis. A project’s NPV = EUR 200 thousand at a selling price of EUR 50 per unit. If the price rises by 5 % (to EUR 52.50 per unit), NPV rises to EUR 260 thousand. The student calculates the elasticity coefficient. What does it show about the importance of price to the project?

Question 2.8 [application]. Exponential smoothing. Quarterly sales (EUR thousand): Q1 = 85, Q2 = 92, Q3 = 88, Q4 = 95. The initial forecast for Q1 = 80. Applying α = 0.3, the student calculates the forecasts for Q2–Q5 and compares them with the SMA (k = 3) forecast for Q5.

2.6.4 Comparative analysis (analysis level)#

Question 2.9 [analysis]. How does the process of preparing a business plan differ for a start-up and for a mature company? Which paradigms (rational planning, lean, agile) are better suited to each case, and why?

Question 2.10 [analysis]. Which cognitive biases, in the student’s view, occur most often in the Lithuanian business context (farms, small businesses, start-ups)? What safeguards could the student propose?

Question 2.11 [analysis]. An audit of the planning process. The student chooses a real organisation (it may be the student’s own workplace, a family business or a publicly described case) and identifies: (a) at which level of planning (strategic/tactical/operational) the organisation plans most strongly and most weakly; (b) which of the five stages of the planning process are carried out formally, which informally and which not at all; (c) whether the organisation applies ex ante, in itinere and ex post evaluation; (d) what improvements the student would propose. An analysis of 2–3 pages is submitted.

Question 2.12 [analysis]. A comparison of forecasting methods. The student finds publicly available data (the State Data Agency; Eurostat; the Bank of Lithuania) on the dynamics of an indicator of a chosen sector of the Lithuanian economy (for example, construction output, retail trade or agricultural product prices) over the past 10 years. The student applies three forecasting methods — SMA (k = 3), exponential smoothing (α = 0.2) and linear regression — and compares their forecasts for the next 2 years. Which method appears the most reliable, and why? What assumptions arise from each method?

2.6.5 Synthesis and original design exercises (synthesis level)#

Question 2.13 [synthesis]. An Excel scenario model. Using the instructions of section 2.5, the student builds an Excel model for a business case of the student’s own choosing (it may be hypothetical). The model must include: the assumptions area, projections for 3 scenarios (5 years), the weighted expected value, the standard deviation and the sensitivity table. A short interpretation (1 page) is submitted: which assumptions are critical, and why?

2.6.6 Open discussion questions (evaluation level)#

Question 2.14 [evaluation]. Mintzberg argues that “planning is not the making of strategy but the programming of strategy”. Is the claim correct? The student gives arguments for and against, drawing on the material of section 2.1.

Question 2.15 [evaluation]. Arie de Geus (1988) calls planning learning, and corporate planning institutional learning. What practical implications does this thesis have for an organisation’s planning process? How does it change the logic of evaluation?

Question 2.16 [evaluation]. Does the business plan appraisal system of the National Paying Agency (NMA) encourage “good” planning (clearly stated assumptions, scenarios, sensitivity analysis) or “attractive” planning (optimistic figures without alternatives)? The student supports the answer with arguments.

2.6.7 Decisions with AI#

The earlier exercises of this section test what the student is able to do alone. These two test something else — whether the student is able to formulate a problem, hand tasks over to a tool, check the result obtained and take public responsibility for it. The procedure for the tasks is set out in Appendix A: the extended task workflow (A.1), the verification protocol (A.2), the handover record (A.3), the verdict (A.4), defence and critique (A.5), the assessment criteria (A.6), the conditions (A.7). The appendix needs to be read before starting the task.

Question 2.17 [basic task]. Situation. UAB “Girių skonis” (a hypothetical private limited company) produces mixes of dried berries and mushrooms. Its revenue in 2021–2025 (EUR thousand) was: 2 000; 3 400; 4 000; 4 300; 4 400. The 2021 result was sharply reduced by the breakdown of the sales chains to open-air markets and catering establishments; those chains recovered in 2022. The owner is preparing an expansion plan and is applying to a bank for a loan; the plan provides for revenue of EUR 8 million to be reached in 2028. The bank’s analyst asked an AI assistant to assess whether the target is realistic. Before reading the AI summary, protocol steps 0a and 0b are written down (A.2).

This is a task at openness level 2: where to look for the weak point is not stated. The growth rate must be recalculated — for at least two different base periods (formula 2.1) — with a statement of why that particular period was chosen. The summary set out below is of the kind that an AI assistant prepares in a few minutes. It contains no arithmetic errors.

An appraisal of the expansion plan of UAB “Girių skonis”. Revenue in 2021–2025 grew from EUR 2.0 million to EUR 4.4 million. The compound annual growth rate: CAGR = (4.4 / 2.0)^(1/4) − 1 = 21.8 %.

Applying this rate to 2026–2028: 4.4 × 1.218³ = EUR 7.95 million. The EUR 8 million target set out in the plan practically coincides with the company’s historical growth trend and is therefore to be regarded as well founded. The company is growing faster than the average of the food processing sector. Recommendation: grant the loan.

The verification protocol is applied to the summary, and the question whether the conclusion withstands its checks is answered. The calculations are approximate — what is assessed is not precision but what has been noticed.

What is submitted: up to two pages — the records of the eight protocol steps (0a, 0b and 1–6), a verdict with a review condition, and answers to three questions: what growth rate the period 2022–2025 shows and why it differs from the one given in the summary; what the growth dynamics of the last two years say about how well founded the plan is; and what the summary lacks if it is to be relied on in taking a financing decision.

Question 2.18 [extended task]. This task is intended for answering independently when each way of planning is appropriate: the student formulates a research question, hands part of the analytical work over to AI tools, checks the results obtained critically and defends the resulting conclusion in the seminar. The general principles are set out in Appendix A; what is detailed here is the mechanics of the task specific to this chapter. The basic task requires no agent environment; for the extended task, the AI agent working environment described in Appendix B is recommended — its files are then submitted together with the work (A.3).

The conditions of the task. The lecturer assigns each student a planning conception from the development of the paradigms discussed in section 2.1.1 (Table 2.1). The object — the business, the conditions and the level of planning at which that conception is the most appropriate way of planning — is found by the student; this search is precisely what forms the core of the task. On the scale of levels of openness in A.1 this corresponds to level 4; if the lecturer also specifies a sector or a company, to level 3; in the open variant, where the student chooses the conception as well, to level 5.

The addressee of the decision is the head or the owner of the organisation, who is deciding not what to do but how to plan: over what horizon, with a document of what degree of formality, and how often the assumptions are to be reviewed.

What the defender must prove. What is defended is not the correctness of the conception in general — all four paradigms are alive, and they complemented rather than replaced one another (the first proposition of section 2.3). What is defended is its suitability under specific conditions, and the work must therefore contain three things:

1. The area of suitability. A specific business or type of activity, the conditions (the level of uncertainty, capital intensity, the reversibility of the decision, the cost of an experiment) and the level of planning — strategic, tactical or operational (section 2.1.2). The claims are grounded in real examples found by the student, not in general references to the literature.
2. Evidence about the environment. It is shown that suitability is determined by measurable properties of the environment and not by the preferences of whoever chooses (see the quantitative requirement).
3. The limits of the conception. What the conception does not see, and when it begins to mislead, is stated. The limits are sought not only in the criticism offered by another paradigm but also in the assumptions of the conception itself; without this part the work is regarded as unfinished.

Quantitative justification (mandatory). The level of uncertainty is the principal criterion of a conception’s suitability (the second proposition of section 2.3), and it must therefore be substantiated, not declared. The quantitative part measures not uncertainty itself but observed variability and predictability — what can be calculated from past data. Historical calm does not remove uncertainty: a structural break occurs without warning (section 2.1.3). The quantitative estimate must therefore be accompanied by a qualitative part: the reversibility of the decision, the cost of an experiment, and whether the possible outcomes are known at all. The methods of this chapter and public data are applied — the State Data Agency, Eurostat, the Bank of Lithuania and sector reviews. The measures are: the fluctuation of the indicator over 8–10 years (the descriptive standard deviation and the coefficient of variation — see the note to formula (2.6)); the stability of the growth rate over different periods (formula 2.1); the clarity of the trend (formula 2.4); or the forecast error obtained by comparing the result of SMA or of exponential smoothing (formulas 2.2–2.3) with the actual figure. One measure is sufficient if its choice is justified.

Organisation of the work and cross-analysis. Within a subgroup, each member defends a different conception. Because the conceptions compete with one another, the seminar naturally turns into a debate rather than a sequence of presentations: the same business decision looks different in different paradigms. Where the number of students is large, the lecturer may direct several subgroups to different sectors.

Example: the student who defended administrative planning shows that at the operational level of a bread bakery, norm-setting and the schedule remain irreplaceable. The student who defended adaptive planning shows that experiment suits the new product line of the same bakery, because its cost is low. The student who defended strategic planning shows that for the decision to build a second production hall neither of the two earlier logics is appropriate, because the investment is irreversible. Three different conclusions about the same company are not a contradiction: they speak about different levels of planning, and it is precisely this conclusion that the students must reach for themselves.

The open variant (e). If the lecturer assigns the task in its open variant, the student chooses both the conception and the case.

The conceptions and the research questions that correspond to them are set out below. The questions in the table indicate a direction; at stage E1 it is narrowed down to one specific question. The third column indicates where to look for the limits — this is a starting point, not an exhaustive list.

The workflow of the task. The task is carried out in accordance with the E1–E7 workflow described in Appendix A (A.1). The examples for this chapter follow; stage E4 needs no separate example (see A.3).

• E1. Formulating the research problem.

  A formulation to be avoided: “The application of adaptive planning in start-ups.”

  A recommended formulation (conception d): “For a rural tourism homestead that is considering both expanding its range of services and constructing a new building, does the logic of adaptive planning suit both decisions or only one of them, and what determines that boundary?”

• E2. Analytical objective and format of the result.

  Example (qualitative): “Three features are identified — the fluctuation of demand, the reversibility of the decision and the cost of an experiment — by which it is decided for which level of planning the conception is suitable; the conditions under which the conclusion would change are stated.”

  Example (quantitative): “The fluctuation of demand in the sector in 2016–2025 is measured (the standard deviation and the coefficient of variation), and the error of the moving average forecast for the same period is compared with the actual figure. The result is an estimate of uncertainty expressed as a range, on which the suitability of the conception is grounded.”

• E3. Decomposition into subtasks and assignment to AI. Examples:

  • the search for the statistical series of the sector and for examples of real companies — assigned to AI as a source finder; every figure and every example is checked in the primary source;
  • the calculation of the fluctuation and of the forecast error — not assigned to AI: it is done independently in the Excel model, because the whole conclusion depends on this result;
  • the formulation of the limits of the conception — not assigned to AI: this is the core of the research, for which the student is responsible;
  • the search for critical literature (for example, Mintzberg’s critique of formal planning, the empirical study by Brews and Hunt) — assigned to AI as a source finder; every claim is checked in the original, because AI tends to simplify the critique into a dichotomy.

• E5. Verifying the results.

  Example (protocol step 0a): “The hypothesis is put forward that demand in the sector fluctuates strongly (the coefficient of variation will exceed 20 %), and that the assigned conception of adaptive planning will therefore be suitable at least at the tactical level.”

  Example (protocol step 0b): “The hypothesis is regarded as refuted if the coefficient of variation turns out to be below 10 % and the error of the moving average forecast below 5 %: in such an environment, experimenting costs more than the information it yields.”

  Example (protocol step 4 — sensitivity): “The result is sensitive to the choice of period: when 2020–2021 are included, the fluctuation almost doubles. Both variants are therefore presented, with a statement of which is to be regarded as typical and why.”

• E6. Research conclusions and review condition.

  Example: “The assigned conception is suitable for the tactical and operational levels of this activity but not for the investment decision: the latter is irreversible, and the cost of an experiment exceeds the possible benefit of the information. The boundary runs not between businesses but between types of decision. The conclusion is to be reviewed if a way of staging the investment were to emerge — the logic of experiment would then apply to it as well.”

• E7. Defence, critique and reflection. The lecturer forms the pairs so that the conceptions collide (A.8): administrative planning against adaptive planning (the logic of stability against the logic of experiment), or strategic planning against strategic thinking — the historical dispute between Ansoff and Mintzberg (section 2.1.1) transferred to a specific case.

  Example: a student who defended strategic planning is opposed by an analyst of adaptive planning: “In the sector of your case the coefficient of variation of demand is 34 %, and the error of the five-year forecast exceeds 30 %. What, then, is a five-year plan for, if none of its figures survives the second year? Does your conclusion withstand this argument?”

Submission. In accordance with A.1: up to 4 pages (shorter for remote work, see A.5); in an annex, the handover record and the unedited prompt log; after the seminar, the defender’s reflection and the opponent’s summary (A.5).

References#

• AB SEB bankas. (n.d.). Investicijų finansavimas [Investment financing]. https://www.seb.lt/verslui/finansavimas/investiciju-finansavimas
• Ackoff, R. L. (1970). A Concept of Corporate Planning. Wiley.
• Akerlof, G. A. (1970). The market for “lemons”: Quality uncertainty and the market mechanism. The Quarterly Journal of Economics, 84(3), 488–500. https://doi.org/10.2307/1879431
• Andrews, K. R. (1971). The Concept of Corporate Strategy. Dow Jones-Irwin.
• Ansoff, H. I. (1965). Corporate Strategy: An Analytic Approach to Business Policy for Growth and Expansion. McGraw-Hill.
• Ansoff, H. I. (1975). Managing strategic surprise by response to weak signals. California Management Review, 18(2), 21–33. https://doi.org/10.2307/41164635
• Ansoff, H. I. (1979). Strategic Management. Macmillan.
• Anthony, R. N. (1965). Planning and Control Systems: A Framework for Analysis. Harvard Business School Press.
• Argyris, C., & Schön, D. A. (1978). Organizational Learning: A Theory of Action Perspective. Addison-Wesley.
• Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140. https://doi.org/10.1016/0749-5978(85)90049-4
• Arrow, K. J. (1963). Uncertainty and the welfare economics of medical care. The American Economic Review, 53(5), 941–973.
• Bank of Lithuania. (2021). Makroekonominės prognozės (2021 m. gruodžio mėn.) [Macroeconomic projections (December 2021)]. https://www.lb.lt/
• Bank of Lithuania. (2023). Metų ataskaita (2022 m.) [Annual report (2022)]. https://www.lb.lt/
• Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120. https://doi.org/10.1177/014920639101700108
• Beck, K., Beedle, M., van Bennekum, A., Cockburn, A., Cunningham, W., Fowler, M., … Thomas, D. (2001). Manifesto for Agile Software Development. https://agilemanifesto.org/
• Brews, P. J., & Hunt, M. R. (1999). Learning to plan and planning to learn: Resolving the planning school/learning school debate. Strategic Management Journal, 20(10), 889–913. https://doi.org/10.1002/(SICI)1097-0266(199910)20:10<889::AID-SMJ60>3.0.CO;2-F
• Bricklin, D. (2009). Bricklin on Technology. Wiley.
• Courtney, H., Kirkland, J., & Viguerie, P. (1997). Strategy under uncertainty. Harvard Business Review, 75(6), 67–79.
• De Geus, A. P. (1988). Planning as learning. Harvard Business Review, 66(2), 70–74.
• Doerr, J. (2018). Measure What Matters: How Google, Bono, and the Gates Foundation Rock the World with OKRs. Portfolio/Penguin.
• Dreyer, H. C., Kiil, K., Dukovska-Popovska, I., & Kaipia, R. (2018). Proposals for enhancing tactical planning in grocery retailing with S&OP. International Journal of Physical Distribution & Logistics Management, 48(2), 114–138. https://doi.org/10.1108/IJPDLM-01-2017-0018
• Drucker, P. F. (1954). The Practice of Management. Harper & Row.
• Eisenhower, D. D. (1957, 14 November). Remarks at the National Defense Executive Reserve Conference. The American Presidency Project. https://www.presidency.ucsb.edu/documents/remarks-the-national-defense-executive-reserve-conference
• Fayol, H. (1949). General and Industrial Management (C. Storrs, Trans.). Pitman. (Original work published 1916)
• Flyvbjerg, B. (2003). Delusions of success: Comment on Dan Lovallo and Daniel Kahneman. Harvard Business Review, 81(12), 121–122.
• Flyvbjerg, B. (2014). What you should know about megaprojects and why: An overview. Project Management Journal, 45(2), 6–19. https://doi.org/10.1002/pmj.21409
• Flyvbjerg, B., & Budzier, A. (2011). Why your IT project may be riskier than you think. Harvard Business Review, 89(9), 23–25.
• Gartner. (2020). 2020 strategic roadmap for cloud financial planning and analysis solutions. Gartner, Inc. (The report is paywalled and Gartner’s site offers no public access, so the primary text has not been read; the report title, the 2020 date and the origin of the term xP&A are reported consistently by several industry publications.)
• Gică, O. A., & Negrusa, A. L. (2011). The impact of strategic planning activities on Transylvanian SMEs: An empirical research. Procedia – Social and Behavioral Sciences, 24, 643–648. https://doi.org/10.1016/j.sbspro.2011.09.084
• Gudonavičius, L., Bartosevičienė, V., & Šaparnis, G. (2009). Imperatives for enterprise strategists. Inžinerinė ekonomika – Engineering Economics, 61(1), 75–82. https://doi.org/10.5755/j01.ee.61.1.11593
• Hope, J., & Fraser, R. (2003). Beyond Budgeting: How Managers Can Break Free from the Annual Performance Trap. Harvard Business School Press.
• Hyndman, R. J., & Athanasopoulos, G. (2021). Forecasting: Principles and Practice (3rd ed.). OTexts. https://otexts.com/fpp3/
• ILTE. (n.d.). Kaip parengti verslo planą [How to prepare a business plan]. https://ilte.lt/kaip-parengti-verslo-plana/654 (accessed 26 September 2026)
• ILTE. (2025). Finansinė priemonė „Startuok“: priemonės aprašymas [The Startuok financial instrument: description of the measure] (version of 11 December 2025). https://ilte.lt/
• Ittner, C. D., & Larcker, D. F. (2001). Assessing empirical research in managerial accounting: A value-based management perspective. Journal of Accounting and Economics, 32(1–3), 349–410.
• Ivert, L. K., & Jonsson, P. (2014). When should advanced planning and scheduling systems be used in sales and operations planning? International Journal of Operations & Production Management, 34(10), 1338–1362. https://doi.org/10.1108/IJOPM-03-2011-0088
• Janis, I. L. (1972). Victims of Groupthink: A Psychological Study of Foreign-Policy Decisions and Fiascoes. Houghton Mifflin.
• Johnson, G., Whittington, R., Scholes, K., Angwin, D., & Regnér, P. (2017). Exploring Strategy: Text and Cases (11th ed.). Pearson.
• Kahneman, D., & Lovallo, D. (1993). Timid choices and bold forecasts: A cognitive perspective on risk taking. Management Science, 39(1), 17–31. https://doi.org/10.1287/mnsc.39.1.17
• Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291. https://doi.org/10.2307/1914185
• Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecard—measures that drive performance. Harvard Business Review, 70(1), 71–79.
• Kaplan, R. S., & Norton, D. P. (1996). The Balanced Scorecard: Translating Strategy into Action. Harvard Business School Press.
• Kaplan, R. S., & Norton, D. P. (2001). The Strategy-Focused Organization: How Balanced Scorecard Companies Thrive in the New Business Environment. Harvard Business School Press.
• Kiechel, W. (2010). The Lords of Strategy: The Secret Intellectual History of the New Corporate World. Harvard Business Press.
• Koller, T., Goedhart, M., & Wessels, D. (2020). Valuation: Measuring and Managing the Value of Companies (7th ed.). Wiley.
• Kraus, S., Harms, R., & Schwarz, E. J. (2006). Strategic planning in smaller enterprises: New empirical findings. Management Research News, 29(6), 334–344. https://doi.org/10.1108/01409170610683851
• Learned, E. P., Christensen, C. R., Andrews, K. R., & Guth, W. D. (1965). Business Policy: Text and Cases. Irwin.
• Lin, P.-C., & Uzsoy, R. (2016). Chance-constrained formulations in rolling horizon production planning: An experimental study. International Journal of Production Research, 54(13), 3927–3942. https://doi.org/10.1080/00207543.2016.1165356
• Makridakis, S., Spiliotis, E., & Assimakopoulos, V. (2020). The M4 Competition: 100,000 time series and 61 forecasting methods. International Journal of Forecasting, 36(1), 54–74. https://doi.org/10.1016/j.ijforecast.2019.04.014
• Makridakis, S., Spiliotis, E., & Assimakopoulos, V. (2022). M5 accuracy competition: Results, findings, and conclusions. International Journal of Forecasting, 38(4), 1346–1364. https://doi.org/10.1016/j.ijforecast.2021.11.013
• Makridakis, S., Wheelwright, S. C., & Hyndman, R. J. (1998). Forecasting: Methods and Applications (3rd ed.). Wiley.
• Markin, S., Sinha, A., Chandna, S., & Foster, J. (2021). SAP Integrated Business Planning: Functionality and Implementation (3rd, updated and revised ed.). Rheinwerk Publishing (SAP PRESS).
• McGrath, R. G. (2013). The End of Competitive Advantage: How to Keep Your Strategy Moving as Fast as Your Business. Harvard Business Review Press.
• Meal, H. C., Wachter, M. H., & Whybark, D. C. (1987). Material requirements planning in hierarchical production planning systems. International Journal of Production Research, 25(7), 947–956. https://doi.org/10.1080/00207548708919887
• Ministry of Agriculture of the Republic of Lithuania. (2024). Ūkio subjektų, siekiančių pasinaudoti parama pagal Lietuvos žemės ūkio ir kaimo plėtros 2023–2027 metų strateginio plano intervencines priemones, ekonominio gyvybingumo nustatymo taisyklės [Rules for determining the economic viability of entities applying for support under the intervention measures of Lithuania’s Strategic Plan for Agriculture and Rural Development 2023–2027] (Order No 3D-440, consolidated version of 30 August 2024). https://www.e-tar.lt/
• Ministry of Finance of the Republic of Lithuania. (2021). Ekonominės raidos scenarijus 2021–2024 metams. 2021 m. gruodis [Economic development scenario for 2021–2024. December 2021]. https://finmin.lrv.lt/
• Mintzberg, H. (1994). The Rise and Fall of Strategic Planning: Reconceiving Roles for Planning, Plans, Planners. Free Press.
• Mintzberg, H., & Waters, J. A. (1985). Of strategies, deliberate and emergent. Strategic Management Journal, 6(3), 257–272. https://doi.org/10.1002/smj.4250060306
• Mintzberg, H., Ahlstrand, B., & Lampel, J. (1998). Strategy Safari: A Guided Tour Through the Wilds of Strategic Management. Free Press.
• National Audit Office of Lithuania. (2025). VĮ Turto banko veikla teikiant valstybės nekilnojamojo turto valdymo ir priežiūros paslaugas [The activity of the state enterprise Turto bankas in providing state real estate management and maintenance services] (State audit report No VAE-3). https://www.valstybeskontrole.lt/
• National Paying Agency under the Ministry of Agriculture. (2023). Artėja metas kreiptis dėl paramos ūkiams modernizuoti [The time to apply for support to modernise farms is approaching]. https://nma.lrv.lt/
• National Paying Agency under the Ministry of Agriculture. (2025). Smulkūs ir vidutiniai ūkininkai paramos paraiškas teiks nuo lapkričio 3 d. [Small and medium-sized farmers will submit applications for support from 3 November]. https://nma.lrv.lt/
• Niven, P. R., & Lamorte, B. (2016). Objectives and Key Results: Driving Focus, Alignment, and Engagement with OKRs. Wiley.
• Nyman, H., Tschirch, F., & Hongell, L. (2025). Linking strategy to tactical planning – using integrated business planning (IBP) as a strategy implementation process. International Journal of Physical Distribution & Logistics Management, 55(11), 213–239. https://doi.org/10.1108/IJPDLM-06-2024-0233
• Nørreklit, H. (2000). The balance on the balanced scorecard—a critical analysis of some of its assumptions. Management Accounting Research, 11(1), 65–88. https://doi.org/10.1006/mare.1999.0121
• Nørreklit, H. (2003). The balanced scorecard: What is the score? A rhetorical analysis of the balanced scorecard. Accounting, Organizations and Society, 28(6), 591–619. https://doi.org/10.1016/S0361-3682(02)00028-0
• Porter, M. E. (1980). Competitive Strategy: Techniques for Analyzing Industries and Competitors. Free Press.
• Porter, M. E. (1985). Competitive Advantage: Creating and Sustaining Superior Performance. Free Press.
• Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard Business Review, 68(3), 79–91.
• Ries, E. (2011). The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses. Crown Business.
• Sarasvathy, S. D. (2001). Causation and effectuation: Toward a theoretical shift from economic inevitability to entrepreneurial contingency. Academy of Management Review, 26(2), 243–263. https://doi.org/10.5465/amr.2001.4378020
• Schwenk, C. R., & Shrader, C. B. (1993). Effects of formal strategic planning on financial performance in small firms: A meta-analysis. Entrepreneurship Theory and Practice, 17(3), 53–64. https://doi.org/10.1177/104225879301700304
• Senge, P. M. (1990). The Fifth Discipline: The Art and Practice of the Learning Organization. Doubleday.
• Simon, H. A. (1960). The New Science of Management Decision. Harper & Row.
• Taleb, N. N. (2007). The Black Swan: The Impact of the Highly Improbable. Random House.
• Taylor, F. W. (1911). The Principles of Scientific Management. Harper & Brothers.
• Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319–1350. https://doi.org/10.1002/smj.640
• Tetlock, P. E., & Gardner, D. (2015). Superforecasting: The Art and Science of Prediction. Crown.
• Weber, M. (1978). Economy and Society: An Outline of Interpretive Sociology (G. Roth & C. Wittich, Eds.). University of California Press. (Original work published 1922)