Divide the present value of adjusted social outcomes by the present value of the inputs invested. First review the outcome quantities, valuation sources and contribution assumptions. Then calculate each year’s value, discount it to a common date and test how the result changes under different assumptions.
This is the calculation exercise in the optional valuation sequence. Bring your value map and proxy-selection notes. You will leave with a calculation that a colleague can reproduce, a sensitivity table and a clear statement of what the result does—and does not—show.
Keep three calculations separate
SROI ratio = present value of adjusted outcomes ÷ present value of inputs.
Net present value (NPV) = present value of adjusted outcomes − present value of inputs.
Net SROI ratio = net present value ÷ present value of inputs.
The conventional SROI ratio and the net SROI ratio are different measures. Name the one you report. Social Value International’s Guide to SROI, Stage 5, distinguishes these calculations and explains discounting and sensitivity analysis.
A hypothetical conventional ratio of 1.48:1 represents an estimated 1.48 units of social value per unit invested under the stated scope and assumptions. It does not mean the investor receives 1.48 units of cash, or that the organization has that amount available to spend.
Review the value map before doing the arithmetic
Confirm which stakeholders, activities and reporting period the calculation covers. Check that each outcome has a measure, a quantity, an appropriate valuation source and an explanation of contribution. Identify material negative outcomes and possible overlap between rows. A clean spreadsheet cannot compensate for a missing outcome or a proxy that values something different.
Review the denominator too. Keep a record of included resources, their timing and any allocation across activities. Explain treatment of relevant non-cash inputs instead of silently counting only the grant. Do not compare a whole-program benefit estimate with the cost of one small part of the program.
Label forecast outcomes separately from observed outcomes. A forecast can support planning, but anticipated results should not become reported achievements simply because they appear in the same worksheet. If the evidence is not ready, revisit when monetary valuation is credible.
Worked example: calculate one fictional outcome
All numbers below are invented to teach the arithmetic. They are not a customer result, recommended proxy, typical discount rate or valuation of any other lesson’s employment cohort. This deliberately simplified example contains one outcome; a real analysis must consider its material outcomes and harms.
Assume 10 outcome units, each valued at 1,000 currency units per year. Gross first-year value is 10 × 1,000 = 10,000. For the exercise, assume 20% would have happened anyway and 25% of the remaining change is attributed to others. Assume zero displacement explicitly for teaching; zero must not become an automatic setting in a real analysis.
Adjusted year-one value = 10,000 × (1 − 0.20) × (1 − 0.25) = 6,000.
Notice that the percentages are applied sequentially to the remaining value. This example does not subtract 20% and 25% from the original value as a combined 45%. In your own model, document precisely what each estimate means and avoid applying the same adjustment twice.
Lay out the years before discounting
Assume this outcome lasts three years, with 10% drop-off after the first year. The undiscounted adjusted values are therefore 6,000 in year one, 5,400 in year two and 4,860 in year three. Drop-off describes a reduction in continuing impact in this model; it is not automatically the percentage of people who stop answering a survey.
Assume inputs of 10,000 at the start, called year zero. Benefits occur at each year-end. Use a 5% discount rate solely for this teaching example. For an actual appraisal, select and document the rate appropriate to its purpose and governing guidance, including consistent treatment of prices and inflation.
Present value in year t = adjusted value in that year ÷ (1 + discount rate)t.
| Year | Adjusted value | Calculation | Present value |
|---|---|---|---|
| 1 | 6,000 | 6,000 ÷ 1.05 | 5,714.29 |
| 2 | 5,400 | 5,400 ÷ 1.05² | 4,897.96 |
| 3 | 4,860 | 4,860 ÷ 1.05³ | 4,198.25 |
| Total | 16,260 | Sum unrounded discounted values | 14,810.50 |
The input occurs at year zero, so its present value is 10,000. If inputs occur in later periods, record those dates and discount them consistently rather than assuming everything happens at the start. Keep unrounded values in the calculation and round only the displayed results.
Duration, drop-off and discounting answer different questions: how long the outcome continues, how its impact changes, and how values at different dates are compared. A missing duration assumption leaves a multi-year projection unsupported; it does not make division mathematically impossible.
Calculate the ratio and explain it
Conventional SROI = 14,810.50 ÷ 10,000 = approximately 1.48:1.
NPV = 14,810.50 − 10,000 = 4,810.50. The corresponding net SROI ratio is approximately 0.48:1. Using NPV as the numerator while calling the result the conventional SROI ratio would confuse two different calculations.
A suitable explanation is: “In this fictional three-year model, adjusted outcomes have a present value of approximately 14,811 against inputs of 10,000. The conventional ratio is 1.48:1, using the stated valuation, contribution, duration and timing assumptions.” Keep the assumptions beside that sentence.
A ratio above 1 does not independently prove causation, funding effectiveness or cash savings. A ratio below 1 means the included valued outcomes are below included inputs under the model; examine omissions, uncertainty and the decision context before judging the activity. Neither number describes everything stakeholders may value.
Test which assumptions change the conclusion
Start by changing one assumption at a time so the effect is visible. Then test combinations that are defensible for the real decision. The variations below are teaching scenarios, not an empirically justified range for a real program.
| Scenario | Change from central model | Conventional SROI |
|---|---|---|
| Central | Three years; original assumptions | 1.48:1 |
| One-year duration | Only the first year is included | 0.57:1 |
| Two-year duration | Only years one and two are included | 1.06:1 |
| Lower proxy | Value per unit reduced by 25% | 1.11:1 |
| Higher proxy | Value per unit increased by 25% | 1.85:1 |
| Higher deadweight | 40% rather than 20% | 1.11:1 |
In this example, the duration assumption materially affects whether the ratio exceeds 1. That gives the team a practical question: what evidence supports value continuing beyond the first year? It does not justify extending the duration until the number looks attractive.
A central estimate is useful when accompanied by its assumptions and sensitivity results. A low-to-high scenario range is not automatically a statistical confidence interval. Nor is a fixed plus-or-minus percentage always the right test: choose alternatives that reflect the actual uncertainty in your evidence.
Test outcome quantities, valuation choices, contribution, duration, drop-off and relevant input estimates. Report important alternatives even when they make the result less favorable. If one choice dominates the result, make that dependency visible and identify the evidence needed to improve it.
Build a worksheet another person can check
- Keep one row per defined outcome, with a source ID for each quantity and valuation.
- Separate original values, contribution adjustments and year-by-year projections.
- Record inputs and benefits against actual timing conventions.
- Show the discount rate, formulas and full-precision totals.
- Keep the central scenario separate from sensitivity scenarios.
- Label unresolved inputs and exclusions; do not convert blanks into zero.
- Save the reviewed version, reviewer and date alongside the report.
Ask a colleague to reproduce the small example above without seeing your final total. They should obtain 14,810.4956 for the present value of outcomes and approximately 1.48105 for the conventional ratio. If not, inspect year numbering, percentage definitions, rounding and whether an input was subtracted twice.
For a larger analysis, reconcile the report’s headline to its outcome rows. A changed source file should lead to a reviewed revision, not an unexplained change in a previously issued report.
Use AI to check the calculation, not invent the assumptions
Review this supplied SROI value map and input schedule. Use only the provided quantities, sources, assumptions and timing. List missing or conflicting fields before calculating. For complete rows, show adjusted annual values, drop-off, discount factors, present value of outcomes, present value of inputs, conventional SROI and NPV separately. Reproduce every formula. Test the alternative assumptions I supply and explain which changes affect the decision. Do not invent proxies, default missing percentages to zero or label the analysis independently assured. End with the evidence a human reviewer must check.
Verify the returned arithmetic in a spreadsheet or other calculation tool. Inspect the source references and the meaning of each percentage. A polished explanation can still multiply the wrong units or count the same benefit twice.
In a configured Sopact workflow, keep the calculation, underlying documents, definitions and review notes connected. Test whether your team can retrieve a source and distinguish a draft estimate from an approved reporting version. This supports accountable review; it does not make software a substitute for valuation expertise.
Watch the SROI explanation
Watch the video · 4 minutes 13 seconds. Review how the ratio relates to the decisions a team makes next. The arithmetic in this lesson is a separate teaching example, so its numbers need not match the video. Browse the video library.
Frequently asked questions
What is the SROI formula?
The conventional formula divides the present value of adjusted social outcomes by the present value of inputs. Define the scope, timing and adjustments before calculating, and identify this as the conventional ratio.
What is the difference between present value and NPV?
Present value expresses values at a common date. Net present value subtracts the present value of inputs from the present value of adjusted outcomes. They are not interchangeable numerators for the same ratio.
Does an SROI ratio of 3:1 mean a cash return?
No. It represents an estimated three units of social value per unit invested within the analysis. Some valued outcomes are not cash flows, and the ratio does not promise repayment to an investor.
What discount rate should I use?
Use a documented rate appropriate to the appraisal’s purpose and applicable guidance, with consistent prices and timing. The 5% rate in this lesson is a fictional teaching assumption, not a general recommendation.
How does drop-off affect SROI?
In a multi-year projection, drop-off reduces continuing impact in later periods under the model’s assumptions. It is separate from discounting and should not be equated automatically with survey nonresponse or participant dropout.
Must I report a range instead of a central ratio?
Report the central calculation with meaningful sensitivity tests and assumptions. Scenario results help show uncertainty but are not automatically statistical confidence limits. Select alternative assumptions for reasons you can explain.
Can I compare two organizations using their SROI ratios?
Do not treat the ratios alone as a ranking. Different stakeholders, outcomes, scope, valuation methods and time horizons can produce different results. Examine comparability and the underlying evidence first.
Can AI certify an SROI calculation?
AI can assist with calculations and review questions, but its output is not independent assurance. Check formulas, sources and assumptions, and use an appropriate review process for the decision and claims being made.
Finish the valuation exercise with a report and a next action
Bring the reviewed calculation into your donor report. State the decision, scope, observed and forecast outcomes, valuation method, central result, sensitivity tests and limitations. Explain whose outcomes remain outside the monetary estimate.
Then choose one practical improvement for the next collection cycle. It might be follow-up evidence on duration, a better outcome measure or a stronger source for a valuation. Assign an owner and review date. The ratio should help the team decide what to learn and change.
Get the impact-report writing ebook and explore report examples. For assistance preparing a draft from reviewed evidence, revisit writing a funder report with AI.
Lesson revised September 12, 2026. All calculation amounts and scenario variations are fictional teaching assumptions. This lesson is not an independently assured SROI analysis.