How do you learn sensitivity analysis quickly and correctly? Start with a financial decision that depends on a forecast. A project might look worthwhile if customer sales grow, equipment costs stay controlled and the required return remains moderate. But what happens if one of those assumptions changes? Sensitivity analysis asks that question systematically. It changes an input while keeping specified other inputs fixed, then calculates how the result responds. This eduKateSG Finance Quick Notes lesson uses an easy net-present-value example to show why testing assumptions is more helpful than treating a single optimistic number as certainty.
- Choose one output: a project NPV, estimated company value, profit margin or another result.
- Identify key inputs: cash receipts, costs, discount rates, growth or timing.
- Set a base case: write the starting assumptions clearly.
- Change a selected input: increase or decrease it while keeping others unchanged for a one-variable test.
- Recalculate: measure both the direction and size of the output change.
- Use realistic ranges: a mathematically possible input is not necessarily plausible.
- Compare scenarios separately: a coordinated downside case can change several assumptions together.
Teaching paragraph 1 — begin with one base-case project. Imagine the fictional investment from our net-present-value lesson. It costs S$250 today and is expected to deliver S$100 at the end of year one, S$110 at the end of year two and S$120 at the end of year three, with nothing afterward. At a hypothetical 10% annual discount rate, the present value of future receipts is about S$271.98, producing a base-case NPV of +S$21.98. This is the result we will stress-test. We are not testing whether S$21.98 is a guaranteed profit; we are asking how sensitive that modelled value is to different assumptions.
Teaching paragraph 2 — change only the discount rate. Keep all cash flows and their dates unchanged. At an 8% discount rate, NPV rises to about +S$32.16. At 10%, it is about +S$21.98. At 12%, it falls to about +S$12.39. The pattern is logical: for positive future cash receipts, a higher required return makes delayed amounts less valuable today. Because the project cash flows are identical in this comparison, any change in NPV can be traced to the discount-rate input.
Teaching paragraph 3 — change a cash-flow assumption instead. Reset the discount rate to the base-case 10%. Now suppose year-two expected cash falls from S$110 to S$90, while the other cash flows stay unchanged. NPV drops from about +S$21.98 to approximately +S$5.45. This is a separate one-variable sensitivity test, not a simultaneous rate-and-sales scenario. It reveals that a S$20 reduction in nominal year-two receipts reduces present value by about S$16.53 at the stated 10% rate. A reader can understand the risk without needing to master an advanced statistical model.
The Four-Step Sensitivity Method
- Step 1: calculate and record the base case.
- Step 2: choose one input to test, such as the annual discount rate.
- Step 3: set a lower and higher plausible value while keeping the other inputs fixed.
- Step 4: compare the new output values with the base and explain the consequences.
The core skill is controlling what changes. If sales, costs and the discount rate all change simultaneously, you may learn about a downside scenario but you will not know how much of the result came from each individual input. One-variable sensitivity analysis provides clarity by isolating an assumption.
Worked NPV Sensitivity Table
- Base project: −S$250 today, +S$100 in year 1, +S$110 in year 2, +S$120 in year 3.
- Discount rate 8%: NPV ≈ +S$32.16.
- Discount rate 10%: NPV ≈ +S$21.98.
- Discount rate 12%: NPV ≈ +S$12.39.
- Discount rate 15%: NPV ≈ −S$0.97.
- At 10%, with year-two receipt reduced to S$90: NPV ≈ +S$5.45.
All values are hypothetical and calculated using end-of-year cash-flow timing. Small display rounding differences can arise. The 15% result is particularly instructive: the same project changes from positive to slightly negative NPV when the discount rate is sufficiently high. That is a model decision boundary, not a claim that market rates will reach that figure.
What Is the Dependent Variable?
The dependent variable is the output you are studying, such as NPV. It depends on the selected inputs. In our example, when the discount rate changes from 8% to 12%, NPV moves from approximately S$32.16 to S$12.39. The output changes because it is calculated from that input. A useful explanation always names the output rather than reporting isolated percentages that leave readers unsure what is being measured.
What Is an Independent Input?
An independent input is a factor you deliberately vary in the model, such as discount rate, unit selling price, costs or next year’s cash flow. ‘Independent’ in this exercise does not mean that the factor is always independent in the real economy. Inflation, prices, wages and discount rates can move together. We hold other inputs fixed temporarily to understand the mathematical effect of each assumption before building more realistic combined scenarios.
Why Start with a Base Case?
If you do not define the original assumptions, a statement like ‘NPV fell by S$10’ has no clear meaning. A base case gives a reference point with explicit cash-flow dates, prices, costs and rates. It also helps catch hidden changes. If an analyst unknowingly changes the project duration as well as the discount rate, the two resulting NPVs may not answer the same question.
Sensitivity Is Not the Same as Forecasting
A sensitivity analysis does not predict which future assumption will occur. It says what a model would produce if that assumption were changed while other specified things were unchanged. A realistic range may be supported by historical variation, market evidence or expert estimates, but the table does not assign probabilities by itself. Avoid turning an illustrative upside case into a promised outcome.
Scenario Analysis Is Related but Different
Scenario analysis usually combines several internally consistent assumptions to describe a possible situation, such as lower customer demand, higher energy costs and a more expensive borrowing environment occurring together. Sensitivity analysis often isolates one driver at a time. Both are useful. A company facing a recession might want a combined downside scenario, but a one-variable sensitivity table can help show which specific assumptions have the largest mechanical effects.
A Simple Downside Scenario
Return to the hypothetical three-year project. A downside scenario could assume year-two receipts fall to S$90 and the discount rate rises to 12%. That is not our one-variable sensitivity result because two inputs changed. Under those combined assumptions, NPV is lower than it is with just one adverse change. Do not attribute the entire difference to sales or the discount rate alone. Model a full scenario separately and label every changed assumption.
Why the Range Must Be Reasonable
A sensitivity table can test any number a computer accepts, including economically impossible assumptions. But a rate that makes a growing perpetuity formula undefined is not a valid valuation result. An extreme sales increase may exceed factory capacity or market demand. A useful analyst chooses ranges that relate to actual risks and makes exceptional stress cases explicit. The goal is to understand decision robustness, not generate spectacular numbers.
What Is Break-Even Sensitivity?
A break-even analysis finds the input level that makes an outcome hit a decision threshold, such as NPV equal to zero. For our example, NPV is positive at a 12% rate and slightly negative at 15%, so the zero-NPV discount rate lies between them. That zero-NPV rate is also the project’s internal rate of return if the solution is unique. See How to Learn Internal Rate of Return Quickly and Correctly for the mathematical definition.
One-Variable Tables Versus Two-Variable Tables
A one-variable table changes one input along a row or column. A two-variable table might show how NPV changes for combinations of cash-flow growth and discount rates, using one input on each axis. The latter can expose interactions and a broader range of results, but it is important to label exactly what stays fixed. A two-variable table still does not cover every possible economic relationship or prove the most likely future result.
The Tornado Chart Idea
A tornado chart ranks the impact of varying different inputs within stated ranges. You can test the discount rate, revenue and equipment cost separately and compare their effects on NPV. The inputs causing the largest output changes appear most influential under those ranges. But rankings depend on how widely you allow each input to vary. If one input is tested over an unrealistic range while another is barely changed, the chart may exaggerate the first one’s importance.
Which Financial Decisions Benefit from Sensitivity Analysis?
- A company evaluating a new machine or store location.
- A lender testing a borrower’s ability to pay under higher interest rates.
- A household comparing mortgage payments under different permitted loan-rate assumptions.
- An investor valuing a business under different cash-flow and growth estimates.
- A manager examining how labour costs or product prices change profit margin.
- A school or education service planning costs against uncertain enrolment.
The same educational method works across personal and business finance. The inputs, obligations and level of risk change, but the learning question remains: ‘What happens to my result when one important assumption changes?’ An honest answer includes both the arithmetic and what the calculation cannot guarantee.
Sensitivity and DCF Terminal Value
In a company DCF, terminal growth and discount rates can produce especially large valuation differences. For example, the perpetual-growth formula divides the next year’s cash flow by the discount rate minus the terminal growth rate. If that difference gets small, the modelled terminal value can rise sharply. That is why it is not enough to publish one terminal value with an unexplained growth assumption. Read How to Learn Terminal Value Quickly and Correctly for a worked test.
Why Correlation Matters
In the real world, inflation may raise both prices and operating expenses, while changes in interest rates can affect borrowing costs and valuation discounts. A model that increases revenue but assumes no associated costs may overstate upside. One-variable analysis isolates the effect on purpose, but combined scenarios should incorporate plausible relationships among drivers. The user should know whether they are looking at an isolated mathematical change or a realistic business situation.
How to Use a Spreadsheet Correctly
Place assumptions in clearly labelled cells, calculate the base case once and reference those cells in the output formula. Build a small table of alternative inputs and record the resulting NPVs. Avoid hard-coding the same discount rate in many formulas, because changing only one copy can make the workbook inconsistent. Verify at least one row by manual calculation, and check that your table returns to the original result at the base-case input.
How to Explain Findings to a Beginner
Use plain sentences: ‘At a 10% required rate, our modelled NPV is about S$22. At 12%, it falls to about S$12. At 15%, it becomes slightly negative.’ Then add a limitation: ‘These are model outcomes based on forecast receipts, not assured returns.’ Readers learn more from that explanation than from a large table without context.
Common Beginner Mistakes
- Changing several inputs while claiming only one was tested.
- Treating a sensitivity table as a probability forecast.
- Forgetting the base-case assumptions or which year cash arrives.
- Using improbable ranges only to produce an attractive outcome.
- Reporting extra decimal places as though they remove uncertainty.
- Failing to check the formula returns the correct base case.
- Ignoring interactions among prices, costs, growth and discount rates.
Ten-Minute Sensitivity Practice
- Minutes 1–2: list project receipts of S$100 at the end of year one and S$121 at the end of year two, with S$180 paid today.
- Minutes 3–4: calculate the NPV at a 10% annual discount rate.
- Minutes 5–6: hold cash flows fixed and recalculate at a 5% discount rate.
- Minutes 7–8: restore the 10% rate and reduce year-two receipts from S$121 to S$110.
- Minutes 9–10: say which assumption changed in each test and why the answer moved.
Mini Quiz and Answers
Question 1: At a 10% rate, what is NPV for −S$180 today, S$100 next year and S$121 in year two? Answer: 100 ÷ 1.10 + 121 ÷ 1.21 − 180 ≈ +S$10.91.
Question 2: If the discount rate alone increases, do positive future receipts normally have higher or lower present value? Answer: Lower, all else equal.
Question 3: If lower sales and higher interest rates are both assumed at once, is that still a one-input sensitivity test? Answer: No. That is a multi-variable scenario or combined sensitivity case.
Continue Learning Finance
Read How to Learn Net Present Value Quickly and Correctly to understand the output, How to Learn Weighted Average Cost of Capital Quickly and Correctly for a possible discount-rate input, and How to Learn Terminal Value Quickly and Correctly for a particularly sensitive DCF assumption. These lessons help turn a model into an explanation rather than a mysterious single answer.
Trusted Sources and Final Idea
See Corporate Finance Institute — Sensitivity Analysis and OpenStax Principles of Finance 2e — NPV Method and NPV Profiles for the technique and discount-rate sensitivity. All values are hypothetical. Remember: change a clearly identified assumption, recalculate one output, explain the difference and never call the model a prediction.