The actual application of financial modelling in the commercial world is not well understood, even by those who regularly develop financial models. The typical misunderstanding is that the key use of a financial model is to generate a number – a value, a return, a profit – that the decision-maker can use to guide action. The real value of a financial model is not the output number but the process of thinking through the model that is required to build it: to make assumptions explicit, to understand the logical links between business drivers and financial consequences, and to understand which variables are the most critical for the decision’s success or failure.

• Financial modelling for investment analysis generates insights that are not available through intuition: specifically, the systematic identification of which assumptions are the most critical – the variables where a 10% change in the assumption results in a 30% change in the output – and which can be used to focus analytical attention and risk management effort on the most critical uncertainties rather than spreading it across all the inputs.
• Modelling for financial forecasting supports strategy by providing an organisational common language for effective strategic discussion. When the management team is considering a decision to expand into a new market, to launch a new product line, to acquire a business, having a financial model that quantifies the economic consequences of the decision under various scenarios changes the nature of the debate from a qualitative debate about the merits of competing opinions to a quantitative debate about the merits of competing assumptions: a debate that is more likely to lead to a well-informed, well-argued decision.

Modelling for business planning is not modelling for reporting – it is predictive and conditional, not descriptive and historical. A financial report describes the past; a financial model describes the future under certain conditions. This is the strength of a model: it enables the decision-maker to explore the entire space of possible futures without being locked into the most recent history or the best-case scenario.

• Finance tools for better decision-making work by translating the complexity of the real world into a simple yet structurally valid model that can be analysed. The simplification is deliberate and necessary – a model that incorporates all possible variables in a real business would be too complicated to build, too difficult to analyse, and too hard to explain. The art of modelling is the art of getting the right simplifications – those that capture the essential structure of the decision, filtering out the distractions.
• The most valuable insights from a financial model are not the base-case projections but the sensitivity analysis and scenario analysis that define the model’s limits: the circumstances under which the decision is no longer right (is it still right if revenue growth is 20% below the base case?), the factors that have the most influence on the decision (a 1% change in WACC has a greater impact than a 5% change in operating costs), and the irreversible decisions that should be postponed until certain uncertainties are resolved.

The use of financial modelling for decision-making is valuable in more business contexts than many people think. The most well-known – valuation for mergers and acquisitions – is only one of many. Capital investment appraisal, business planning and budgeting, pricing decisions, restructuring analysis, debt capacity assessment, and scenario analysis are all examples of decision contexts where a sound financial model provides insights that intuition or rules of thumb can’t.

Table 1: Financial Modelling Applications — Decision Contexts and Key Model Types

Better decision-making through financial analysis means choosing the right model for each decision, rather than using a standardised approach regardless of the decision context. A company that uses an acquisition DCF model for a pricing decision, or a sensitivity table for a restructuring decision, is misusing its model. Knowing what model type is suitable for what decision context – and why – is one of the most important decisions in the practice of data-driven business decisions in finance.

• The most widely applicable model type is the scenario comparison model. This model structure explicitly develops two or three different sets of assumptions (representing different commercial scenarios). It maps their financial consequences, enabling the decision-maker to understand how the decision shifts under different future scenarios. This approach to scenario analysis is valuable for all important business decisions, regardless of the quantitative method applied.
• The most misused model type is the DCF valuation, which is extremely useful for decisions where the value of future cash flows is the major value driver (ongoing businesses, capital investments, long-term contracts) but misleading for decisions where optionality, competitive positioning and relationships are the major value drivers. It is critical to understand the limits of each method’s application, just as it is critical to understand how to build the method.

Knowing how to use financial models in practice means understanding what makes a model useful for decision-making, rather than simply providing numbers for a report. It is not the model’s complexity; it is its structure and analysis that make its logic visible, its assumptions testable, and its results directly relevant to the decision question.

• A decision-useful model is based on a decision question – a specific question that the model answers. ‘Should we buy this business for $200 million, creating value for our shareholders?’ is a decision question. ‘What is the value of this business?’ is not – it is too vague to guide model development. Writing the decision question explicitly before designing the model results in a more targeted and useful model.
• All assumptions in a decision-useful model are sourced, documented and verifiable. This means not only documenting the number used for an assumption (8% annual revenue growth rate) but also why that number was used (the average of the last three years’ growth rates of the comparable peer group as sourced from Bloomberg), what the range of alternative plausible numbers is (the comparable peer group ranges from 4% to 14% annual growth) and what the model’s conclusion is sensitive to, given that range.

Finance tools for better decisions are effective when they are designed from the perspective of the reader – when the model’s structure, naming conventions and presentation of results anticipate the questions that the decision-maker will want to ask and present the requested information as directly as possible. A model that forces the reader to search through 15 tabs, infer results from underlying data, and request additional calculations to answer the next set of questions is unlikely to have an impact on the decision-making process.

The most critical aspect of designing a decision-useful model is separating assumptions from calculations. When assumptions are baked into formulas – as literal constants in the calculation sheets – the model can’t be updated, can’t be audited, and can’t be stress-tested. When assumptions are collected on a single input sheet, clearly identified, and called from the calculation sheets, the model becomes an adaptable tool that can be readily re-configured to explore different scenarios.

• Scenario design – how a model transitions between sets of assumptions – is the model design feature that is most directly linked to decision-usefulness. A model with a scenario toggle (a single cell that switches the entire model between Base, Upside and Downside assumption sets) lets the decision-maker quickly explore the full range of possible outcomes without having to reconfigure the model.
• Sensitivity analysis tables – in particular, the two-variable data tables in the What-If Analysis tool in Excel – are the most decision-useful output of a financial model, because they provide a direct visual representation of the change in the key output metric across the entire range of combinations of the two most significant assumptions. This is usually a table of enterprise value across combinations of WACC and terminal growth rate for a valuation, and NPV across combinations of revenue growth and margin for an investment.

The process of financial modelling for decision making breaks down into a five-step sequence, from decision to model delivery. Knowing this process – and the specific analytical skills required to get the job done – provides an operational guide to building models that support better decisions, rather than generating analytical results that are not relevant to the decision-making process.

Step 1 — Define the Decision, Not Just the Model

The first – and most critical – step in developing a decision-useful financial model occurs before the model is ever created: the explicit definition of the decision that the model will be used to support, the intended users, and the information needed to make that decision.

• Specifying the decision means answering three questions: What is the decision? Who is making the decision, and what do they need to know? And what might change the recommendation – what alternative scenario or assumption would lead to a different decision? The answers to these three questions shape the model’s structure, the level of detail, and the outputs required.
• Building a model before the decision question is fully defined is the most common error in modelling for business planning – it usually results in a model that is too complex, that answers questions that the decision-maker does not need answered, at a level of detail that is not required, and does not address the uncertainties that make the decision complex.

Step 2 — Map the Business Logic Before Opening Excel

Data-driven business decisions in finance mean the model should reflect the business reality of the decision, not a financial model template. The analyst should draw a diagram or write out the causal chain from business drivers (number of customers, price, cost per unit, market share) to financial results (revenue, margin, cash flow) before creating formulas.

• Mapping the business logic is important for two reasons: first, it ensures that the model reflects the correct causal relationships (avoiding the common pitfall of modelling revenue as a top-line growth rate, but not connecting it to the business drivers), and second, it highlights the most important value drivers – those elements in the causal chain where the business reality is most uncertain and where the financial results are most sensitive.
• The model architecture decisions at this point – the level of detail in the revenue build, treatment of working capital, tax and depreciation – will be costly to alter in the future without rebuilding the model. Time spent on architecture before adding content is the most effective use of time in the model development process.

Step 3 — Build the Assumptions Layer with Documented Rationale

The assumptions layer is the analytical core of any model for financial forecasting for strategy – it is where the analyst’s commercial judgement is translated into quantitative terms, where the model’s quality is made or broken. Each assumption should be traced to its source, benchmarked against evidence (if possible) and documented in enough detail to allow a sceptical reader to evaluate its reasonableness.

• The quality of assumptions is measured along four lines of questioning: is it evidence-based (historical, industry, management)? Is it consistent with other assumptions in the model? Does it make sense in the circumstances of the decision? And is it clearly documented so that another analyst would make the same decision?
• The budget accuracy test – comparing the previous year’s assumptions with the actual results – is the most effective way to test the quality of assumptions in recurring models (such as annual budgets or strategic plans). Any assumptions that have been consistently optimistic or pessimistic should be adjusted, and the adjustment documented.

Step 4 — Construct the Analysis and Stress-Test the Conclusions

With the assumptions on the table, the analytical construction of the model turns them into financial outcomes that answer the decision question. The construction should follow the logic detailed in Step 2, from drivers to revenues, costs, earnings, and cash flows, in a logical, arithmetically verifiable sequence.

• When financial models are used for investment analysis, the analytical construction should be followed by a stress test – running the model under a downside case to test the decision’s resilience to adverse conditions. The downside case is not the worst case (which would paralyse all decision-making) but a plausible adverse case that the decision-maker should be prepared to deal with.
• Sensitivity analysis should be run on the two or three assumptions that most affect the output, determined by a simple one-variable sensitivity analysis that measures the change in the output for a given change in each assumption. The assumptions that cause the biggest changes in the output are the ones that should be most closely scrutinised, benchmarked and managed.

Step 5 — Communicate the Outputs for Decision Action

An analytically sophisticated model cannot be used to improve decision quality if its outputs are not communicated in a way the decision-maker can easily understand, challenge, and put into action. Finance decision support tools for business that are technically sound but poorly communicated are one of the most frequent – and avoidable – sources of value destruction in finance practice.

• Communication of decision support should present the conclusion first (the model’s recommendation, and why it should be followed) before the analysis. If a 20-page model output is presented to a decision-maker before he or she knows what the model recommends, then the decision-maker will have already developed an opinion about what is best and is less likely to be persuaded that the model is correct.
• The optimal format for communicating decision support is a one-page summary that includes the recommendation, the three most important assumptions underlying it, the sensitivity of the recommendation to changes in those assumptions, and the risks that could change the recommendation. It saves the decision-maker’s time and provides the information required to make a decision.

To support better decision-making with financial analysis, it is essential to develop a modelling capability that integrates technical Excel skills, commercial judgement and communication skills. The journey is not a straight line – technical skills can be developed rapidly through practice. Still, commercial judgment develops over time as experience is gained in decision-making and in observing the dynamics of business. The best approach to development is a combination of technical skills development with exposure to the commercial environment in which models are used.

Table 3: Financial Modelling Capability Development — A Progression Framework

The most effective single practice activity to build the real use of financial modelling decision support skill is to dissect and rebuild publicly available models – the financial models in scheme booklets, independent expert reports and equity research reports from investment banks and advisory companies. These documents contain models built by professionals for real, critical decisions and examining these models shows the analytical decisions, model structures and communication strategies used by practitioners much more clearly than any classroom example.

• Building a simple model for all major decisions you come across (even though they may not be part of your official job description) is the best way to develop the business acumen that sets the experienced practitioner apart from the technically competent novice. The model for your personal investment, career choice, or hypothetical business uses the same skill as the model for a corporate acquisition.
• Soliciting feedback from non-finance colleagues on the outputs of models you build – specifically, whether the executive summary effectively communicates the recommendation and whether the supporting analysis answers the questions they want answered – builds the communication skill component of modelling competency that is impossible to develop through technical skill-building alone.

Decision-making financial modelling is not most valuable as a technical skill in building spreadsheets, but as a process of structured thinking about the future – a way to make the decision assumptions explicit, the risks transparent and the range of possible outcomes apparent. The use of financial models in practice, as required by today’s finance professionals, is not about how to build models but about why certain model structures and analytical approaches are used to shape decisions, rather than simply provide a quantitative context for decisions that have already been made.

The practitioners who become best known as business decision support tools in finance are those who combine their financial modelling skills with commercial judgement, integrity and communication skills – those who can tell the decision makers what they don’t want to hear, when they are justified in doing so, in a way that gets them to change their decision. Data-driven business decisions: Finance at this level is one of the most highly valued skills in the modern business world, and one that grows in value with every model constructed, every decision supported, and every problem solved.

• Focus on the decision question – write it down first. The question determines the scope of the model, the usefulness of the results and the likelihood that the analysis will inform the decision.
• Always start with the assumptions layer and make all material assumptions visible, sourced, and documented. If you can’t find and change the assumptions, you don’t have a decision-support model; you have an unreliable calculation.
• Always run and report a credible downside scenario. Strategic financial forecasting that only presents the best case leads to overconfident decision-making; the downside case is the insurance policy that stops the worst decision-making.
• Conclusion before method: Models are not read to be understood – they are ready to make decisions. Organise all communications of outputs to answer the decision question as quickly as possible.
• Build commercial judgement as well as technical ability – using financial analysis to improve decisions requires an appreciation of the business environment in which the model is being used. Look for opportunities to work on real decisions, real business problems, and real commercial discussions, and view every engagement as an opportunity to develop the analytical judgement that makes technical modelling useful for decision-making.