Sensitivity Analysis

This technique determines the sensitivity of an outcome (e.g. profit projection, value of a project, or valuation) to changes in key assumptions used in a base or primary case. Any projection or estimated value must be viewed as an estimate based on a number of inherent assumptions. Performing a sensitivity analysis is very useful to understand the dynamics of a particular projection or decision and to highlight the importance of testing assumptions. For example, in Table 3.4, we estimate the sensitivity of the stock price of Roberts Manufacturing Company to changes in key assumptions: the rate of sales growth and the level of profitability.

Sensitivity analysis will be used throughout this book. For the specific application of this sensitivity analysis, see Chapter 22, Business Valuation and Value Drivers.

Scenario Analysis

A projection or forecast version is based on a specific set of conditions and assumptions (e.g. economic recovery). Since projected performance will vary under different assumptions or scenarios, significant projections should be recast under different conditions, events, or scenarios. A scenario is an imagined sequence of future actions or events. Whereas sensitivity analysis provides insight into the impact of “flexing” or changing one or more assumptions, scenario analysis will examine the projection if a different course of action or chain of events occurs. For example, if the primary or base case of a projection assumes economic recovery, what would the results be if the recession continues? This requires the analyst to examine and revise all variables under a specific scenario. Scenario analysis requires substantial consideration to identify and change assumptions that are impacted under a different scenario. For example, a recession scenario could affect unit sales volumes, average selling prices, interest rates, inflation, labor costs, and commodity prices. One of the most important aspects of scenario planning is to identify the management actions that can be taken to mitigate unfavorable changes or to capitalize on potential upsides.

A simplistic form of scenario analysis is to develop best‐case and worst‐case scenarios. This will develop a range of potential outcomes that can be useful in evaluating a projection.

Probability and Expected Value

In developing financial projections and evaluating investment projects, probability analysis can be helpful. In the example in Table 3.5, executives had developed estimates of sales under five discrete scenarios. In addition to the base plan, two upside and two downside scenarios were identified. Managers had estimated the probability of each scenario. The simple illustration provides an expected value (EV), or outcome, based on the probabilities assigned to each of the scenarios.

Of course, the most critical part of this analysis would be to identify the scenarios, estimate sales, and also estimate the probability of each scenario. In some cases, prior experience may provide a basis. Industry trends and data can be useful, for example the probability of product development success. Even where the probabilities can't be developed scientifically, use of reasonable estimates can prove helpful. In this case, the base projection of $120,000 appears aggressive when compared to the EV of $117,700. There is more downside exposure than upside to the base plan, and, pending additional information, the base forecast should be revised down to $117,700. Probability and expected value are utilized in Part Three (Business Projections and Plans), in Chapter 15 (Revenue and Gross Margins), and also in Part Five (Valuation and Capital Investment Decisions).

Decision and Event Trees

Decision and event trees are useful to visualize, evaluate, and communicate various scenarios, especially future projects or other decisions where the outcomes are uncertain. For example, a firm may face a decision to replace an existing product with a new one or continue to sell the existing product. This is unlikely to be a single decision or event. How successful will the new product be? What will happen to the sales of the existing product if not replaced? What subsequent options will management have in optimizing the result?

The simple illustration in Figure 3.1 is a very effective way to lay out various management decisions and to describe potential outcomes resulting from each alternative choice. Each of the six potential outcomes will have probability of occurrence and an estimated value (e.g. NPV, sales, earnings per share). Each of these six outcomes will also involve a second level of management options or decisions.

A more comprehensive review of decision trees and examples with values and probabilities is included in Chapter 21, Capital Investment Decisions: Advanced Topics.

Performance Trees

While we are on the subject of trees, another useful way to analyze and present financial data is through the use of tree analysis. Essentially, the analyst will create a visual model to drill down to almost any element of financial performance. These can be used to model/illustrate return on capital or equity, value drivers, or drivers of accounts receivable DSO. An early application is the so‐called Dupont formula used to analyze or decompose return on equity (ROE). In Table 3.6, we have modeled out the three major drivers of ROE: profitability, asset turnover, and financial leverage. The model allows us to change any estimate of financial performance (e.g. improving margins) to compute the effect on ROE. In addition to the useful aspects of this model, it is an outstanding way to communicate the drivers of ROE.

Predictive Models

Predictive models are utilized frequently, especially on the planning side of FP&A. They can range from a simple projection of project costs in Excel to complex models used to predict the weather, volume of calls or retail traffic, or ultimately revenue levels. All predictive models build on the experience of the past to project future events, trends, or activity. Specific examples of predictive modeling are included in Chapters 13 (Budgets, Operating Plans, and Forecasts), 14 (Long‐Term Projections), and 15 (Revenue and Gross Margins).

Performance Improvement and Diagnostic Tools

In Chapter 1, we suggested that the objectives of financial analysis are not simply to report on events, outcomes, and future projections. Analysts should be prepared to identify the root causes of variances, shortfalls, and overruns, and be prepared to recommend possible corrective actions. Even better, analysts should identify improvement opportunities and recommend actions to realize these opportunities. This type of analysis requires an understanding of business processes, process improvement, and diagnostic techniques. In many cases, these skills will not be resident within the traditional FP&A department. Many of these skills are developed in consulting firms or in quality or continuous improvement organizations within the enterprise. In Chapter 5, Building Analytical Capability, we review ideas for evaluating, developing, or acquiring talent with these skills (or teaming with resources outside FP&A).

Figure 3.2 illustrates a simple but effective summary of an analysis of unfavorable accounts receivable trends and identification of the root causes of higher past‐due accounts. Accounts receivable balances reflect the overall effectiveness of the revenue process, including order entry, billing, shipping, quality, and customer service. The charts identify and summarize the specific reasons or root causes for a large number of receivables going unpaid owing to product quality problems. Armed with this data, the analyst can recommend specific corrective actions and measure progress toward implementing better practices.

For a more detailed explanation of the revenue process and accounts receivable, refer to Chapter 17, Capital Management and Cash Flow: Working Capital. These performance improvement tools can be applied across a large number of business processes, including planning and forecasting, supply chain management and inventory, cost of quality, manufacturing variances, and project cost overruns.

Activity‐Based Measures

In many types of analysis, including product, customer, and business unit profitability, many general costs and expenses must be assigned or allocated. The easiest and most common method is to allocate expenses on the basis of sales dollars. This may be acceptable if the amounts allocated are not significant, or if sales dollars are representative of the activities and drivers of that expense. For example, order processing, invoicing, and service activities may support all product lines within an enterprise. In preparing product line P&Ls, these costs are allocated on the basis of sales dollars. However, in most organizations, sales dollars are not reflective of the level of activity and time attributable to each product line. For example, the cost of processing and invoicing an order may be the same for a small‐dollar repair order as it is for a $100,000 system. Service costs will be driven by the level of product failures and returns, not sales dollars. Inevitably, certain product lines will consume a disproportionate share or these activities.

Table 3.7 presents a comparison of these activities and measures and an allocation of total costs on the basis of sales dollars.

Note that the activity measures selected as representative of the drivers of time and cost for the department, namely number of invoices generated and returns, vary significantly from the percentage of sales dollars for each product line.

Table 3.8 assigns the cost for order processing and billing (OPB) on the basis of the number of transactions (invoices), and assigns the customer service costs on the basis of product returns. The table also compares the costs allocated by sales dollars to the costs assigned by activity measures, highlighting a significantly different assignment of costs. The analysis also identifies some other notable concerns that should be pursued, for example the large number of returns for parts. What are the root causes of returns and possible corrective actions? The company should also look at the profitability for parts, to ensure that pricing covers the relatively high transaction costs.

Benchmarking and External View

Most traditional efforts by FP&A professionals are directed at the internal aspects of the enterprise. However, these professionals can extend and expand the value they add by looking outside the enterprise. FP&A can utilize benchmark and external information to add value in the following ways:

• Analysis of markets, customers, and competition.
• Using external information to benchmark and evaluate performance.
• Utilizing benchmark information to set enterprise goals for performance and value creation.

Benchmarking is more fully explored in Chapter 11, The External View: Benchmarking Performance and Competitive Analysis.