Sopact is a technology based social enterprise committed to helping organizations measure impact by directly involving their stakeholders.
Copyright 2015-2026 © sopact. All rights reserved.
SROI explained: the seven principles, the formula, the four adjustments, a worked example, and how to calculate SROI for a grantee or across a fund.
SROI, or social return on investment, is a framework for putting a dollar value on the changes a program produces for the people it serves and others affected. You collect evidence of what changed, assign each change a financial proxy that reflects what it is worth to the person who experienced it, subtract the share that would have happened anyway, and divide by what the program cost. The result is a ratio: every $1 invested produced $X of social value. The number is the headline; the evidence chain is the work.
The ratio itself is the least interesting part. Two SROI ratios are not comparable unless the studies used the same scope, stakeholder definitions, and adjustment rules — and they rarely do. What makes an SROI defensible is the discipline behind it: stakeholder-defined outcomes, sourced proxies, and four named adjustments that separate a gross number from net social value. Skip those and the ratio is a marketing figure, not evidence.
Key takeaways
Most SROI studies are won or lost at one stage: subtracting the value the program did not create. Some of the change would have happened anyway (deadweight), some was caused by other actors (attribution), some was pushed onto other groups (displacement), and some fades in years two and three (drop-off). Estimate each, cite the basis, and show the ratio before and after; skip any and a reviewer reads the whole study as un-rigorous.
Sopact calls the operating version Continuous SROI: the proxy library, the four adjustment rates, and the persistent participant ID are held in one place, so when a source updates or a fresh cohort completes, the ratio recomputes for every cohort rather than being rebuilt by hand. The outcome chain an SROI prices is validated first on the theory of change page, and the collection that feeds it on pre-and-post surveys.
The method is published by Social Value International as seven principles, and outcome definitions can be pulled from the IRIS+ catalog. What Sopact adds is the data layer that keeps the proxies, adjustments, and evidence connected so the study is repeatable, not a one-time consultant engagement.
SROI rests on seven principles published by Social Value International: involve stakeholders, understand what changes, value the things that matter, only include what is material, do not over-claim, be transparent, and verify the result. They are not a formality — six of the seven are about restraint, and most weak studies fail on principle five.
Read the principles as data requirements rather than ethics. Involving stakeholders means their words have to be captured and kept; being transparent means the scope, proxies, and assumptions must survive alongside the number; verifying means someone outside the study can re-run it. Each of those is a demand on the data layer, and each is what a Continuous SROI is designed to hold.
| Principle | What it demands in practice |
|---|---|
| 1 · Involve stakeholders | The people who experience the change define what changed and what it is worth |
| 2 · Understand what changes | Evidence the outcome chain, including unintended and negative outcomes |
| 3 · Value the things that matter | Assign a cited financial proxy to outcomes that have no market price |
| 4 · Only include what is material | Drop outcomes too small or too uncertain to change a decision |
| 5 · Do not over-claim | Apply deadweight, attribution, displacement and drop-off before claiming value |
| 6 · Be transparent | Disclose scope, sources, proxies and assumptions so the study can be checked |
| 7 · Verify the result | Have the analysis reviewed independently, including by the stakeholders themselves |
The SROI formula is the present value of net social benefits divided by the value of inputs. Sum the dollar value of every material outcome, subtract deadweight and attribution, apply displacement and drop-off, discount later years to present value, then divide by total program cost.
The arithmetic is a single afternoon's work. What takes a quarter is producing the inputs the formula needs — evidenced outcomes, cited proxies, and defensible adjustment rates. The walkthrough for each term is in the Academy: how to calculate the SROI ratio.
You calculate SROI in six stages: scope the analysis and its stakeholders, map the outcomes that actually changed, collect evidence and assign a cited financial proxy to each, apply the four adjustments, compute the present-value ratio, and report with a full audit trail. The first stage controls the rest — a wide scope with thin evidence is the most common reason an SROI fails external review.
The value map that carries stages one to three is built in how to build an SROI value map, and the proxy choice in stage three in how to pick a financial proxy.
The arithmetic of an SROI takes an afternoon; the data preparation takes a quarter. A workforce program has wages in payroll, confidence in an exit survey, and a deadweight estimate it cannot make without a comparison group — and the survey tool cannot link the same person across intake and follow-up because half the email addresses changed. The consultant spends three months matching CSV exports by hand before the calculation can begin.
That is where a persistent participant ID changes the economics. When intake, exit, and 90- or 180-day follow-up resolve to the same person, the outcome chain is reconciled at the source, the open-text confidence answers are extracted into variables, and the four adjustments are stored as named parameters. The first study still takes the same care; the second takes hours. Tracking a cohort across years for the drop-off curve is on longitudinal survey, and the wider frame on five dimensions of impact.
Four adjustments separate a gross outcome value from net social value: deadweight, attribution, displacement, and drop-off. Read the last column: skipping any one is the most common reason a published ratio fails an audit.
| Adjustment | What it subtracts | Common failure |
|---|---|---|
| Deadweight | The share that would have happened anyway | No comparison group, so 0% is assumed — read as zero rigor |
| Attribution | The share other actors or programs caused | 100% claimed, ignoring partner services and conditions |
| Displacement | The effect pushed onto other groups | Skipped entirely, even where it is the largest discount |
| Drop-off | How value fades in years two and three | A year-one peak reported as the headline, no curve shown |
Each adjustment needs an estimate, a cited source, and a sensitivity range, and the report should show the ratio before and after them. Estimating the two hardest is walked through in how to estimate deadweight and attribution. Holding the four as named parameters on one record is what lets a Continuous SROI recompute honestly when the evidence changes.
A 120-participant workforce program costing $480,000 produced an unadjusted ratio of 1:8.0 and an adjusted ratio of 1:4.7 — the four adjustments removed 41% of the headline value. That gap is the entire argument for doing them.
The cohort is 120 enrolled. Employment at six months was evidenced for 78 participants at an $18,400 earnings differential drawn from labor statistics; confidence and job-search persistence for 96 at $3,200 from a published social value bank; reduced benefit reliance for 41 at $6,900 from state benefit cost data. Figures are illustrative and the proxy sources are described by type rather than cited, because a real study must use proxies sourced for its own population and geography.
| Outcome | Gross value | After adjustments |
|---|---|---|
| Employment at 6 months | $1,435,200 | $869,193 |
| Confidence, job search | $307,200 | $172,032 |
| Less benefit reliance | $282,900 | $152,766 |
| Year-one total | $2,025,300 | $1,193,991 |
Year one net social value is $1,193,991. Carrying it forward with drop-off at 60% of year one in year two and 35% in year three, discounted at 3.5%, gives a present value of about $2,276,000 against $480,000 invested — a ratio of 1:4.7. Run the same arithmetic without the adjustments and the same program reports 1:8.0.
A reviewer will read the adjusted figure as the credible one and treat the unadjusted figure as a warning sign. This is why a published SROI should always show both.
Short-cycle training programs fail on drop-off, long-horizon education and health programs fail on attribution, and systems or advocacy work fails on deadweight and displacement. The adjustment most likely to break your ratio is a function of what your program is, not how careful you are.
Short-cycle programs end weeks after intake and rarely follow participants past exit, so persistence is assumed rather than measured and the ratio inflates. The repair is a follow-up wave on the same participant ID at 90 and 180 days, which is what pre-and-post surveys and the drop-off curve on longitudinal survey exist to produce.
Long-horizon programs touch a participant alongside schools, clinics, and other services, so a 100% attribution claim is not defensible; the repair is asking participants directly what else contributed, and holding the answer on the record. Systems and advocacy work has no observable counterfactual and often moves value between groups rather than creating it, so deadweight and displacement carry the analysis — and honest de-scoping, saying which outcomes you will not value, is stronger than a confident number. Where SROI is one instrument among several is set out on monitoring and evaluation, and the training-specific frame on the Kirkpatrick model.
When a funder asks an investee or grantee for an SROI, it is asking for a number that has to be priced off case-level evidence: intake forms, service events, case notes, attendance, and follow-up surveys for each participant. Most grantees hold that evidence in three or four systems that were never joined, so the request lands as a request for a research project.
Sopact calls the layer that answers it Case Intelligence: every service event, case note, and follow-up response resolved to one participant under a persistent Contact ID, so a grantee's SROI is priced from the case record rather than reconstructed from spreadsheets. The distinction matters because SROI values change in individual lives — a ratio built from program-level totals cannot show which participants moved, and cannot survive the question of who did not.
This is also what makes a grantee's number auditable by the funder. When each valued outcome resolves to a participant and a date, the funder can drill from the ratio to the case that produced it instead of accepting a summary. The record itself is on case management software; pulling outcomes out of grantee narratives is walked through in extract outcomes from a grantee report, and comparing one grantee against the rest in how to compute grantee variance.
A fund cannot produce a portfolio SROI by adding up its investees' ratios, because each investee scoped its study differently, chose different proxies, and applied different adjustment rates — summing them produces a number with no defensible meaning. The fix is not a bigger calculation. It is agreeing the definitions before anyone calculates anything.
Sopact calls that Portfolio Intelligence: one shared data dictionary across every investee, so each outcome, proxy, and adjustment rate is defined once and a fund-level SROI is comparable by construction rather than reconciled after the fact. Transparency is the first half — every investee reporting against the same published definitions. Traceability is the second, and it is the harder one.
What a limited partner or board actually tests is whether a fund-level figure resolves downward. Sopact calls a ratio that passes that test a Traceable Portfolio Ratio: every number in the fund's SROI drills from the portfolio total, to the investee, to the participant response and cited proxy behind it. A ratio that cannot be drilled is a claim; one that can is evidence. The fund-level view is on portfolio intelligence and the ongoing reporting cadence on portfolio monitoring software.
Blending financial and social return across a portfolio adds one more requirement — the two returns must be computed over the same period and the same set of investees, or the blend flatters whichever side has better data. The method is in how to calculate a blended portfolio SROI, the shared definitions in how to build a data dictionary, and the management practice around it on impact measurement and management.
The calculator below runs the same arithmetic as the worked example: enter your outcomes, the people affected, a proxy per person, and your four adjustment rates, and it returns the ratio before and after adjustments. It is an estimate to argue with, not a study — a published SROI still needs sourced proxies and a stated scope.
A forecast SROI published at proposal stage and never revisited is a projection, not evidence. Reading cohort data as it arrives lets the ratio move from forecast to evaluative and stay current as drop-off shows up in later years. That is the premise of the Loop, Sopact's method for continuous impact intelligence: collect clean at the source, analyze the moment data arrives, improve while you can still act.
The Loop is also what makes an SROI defensible. Every valued outcome traces back to the stakeholder evidence and the cited proxy behind it, so when a funder questions the ratio it resolves to its source. That standard has its own chapter in traceability and transparency.
One method, three moves that never stop
Then the cycle runs again, a little sharper each cohort. Read the method: the Loop methodology →
The fastest way to feel the difference is to value one outcome with a cited proxy and the four adjustments. Each prompt below pastes into Sopact Sense's Assistant, or reasons through with your team; the arrow above each links the Academy walkthrough that shows the expected output and the tips.
Academy walkthrough → Build the SROI value map
Build the SROI value map for this program: [PASTE PROGRAM + STAKEHOLDERS + INVESTMENT]. For each stakeholder group, list the outcomes that actually changed in their lives, mark each material or immaterial, and give the quantity and the evidence source. Include negative and unintended outcomes. Where evidence is missing, write NO EVIDENCE rather than assuming the outcome occurred. Return a table: Stakeholder / Outcome / Quantity / Evidence / Material?
Academy walkthrough → Pick a defensible financial proxy
For each outcome below, propose a financial proxy with its specific source, a conservative and optimistic range, and the population the source was drawn from: [PASTE OUTCOMES + GEOGRAPHY + POPULATION]. Flag any proxy whose source population differs materially from mine. Where no credible source exists, write NEEDS PROXY rather than inventing one. Return a table: Outcome / Proxy / Source / Range / Population match?
Academy walkthrough → Price a grantee's SROI from the case record
From this grantee's case-level data, build the SROI inputs: [PASTE CASE RECORDS OR GRANTEE REPORT]. For each participant-level outcome, give the count, the evidence in the record that supports it, and the date it was observed. Separate outcomes evidenced at the case level from those asserted only in narrative. Return a table: Outcome / Participants / Case evidence / Narrative only? / Observed date
Academy walkthrough → Roll up a portfolio SROI
Across these investees, test whether a portfolio SROI is defensible before calculating one: [PASTE INVESTEE RATIOS + THEIR SCOPES + PROXIES + ADJUSTMENT RATES]. Flag every place where scopes, proxy sources, adjustment rates, or time periods differ enough that summing would mislead. Then propose the shared definitions that would make them comparable. Return two tables: Incomparability / Investees affected / Why, and Metric / Shared definition / Proxy source.
Each walkthrough is short and practical: what to do, the prompt to run, the output to expect, and the tips that keep it reliable.
Watch: valuing outcomes with cited proxies and honest adjustments — the discipline behind a defensible SROI.
SROI, or social return on investment, is a framework for putting a dollar value on the changes a program produces for the people it serves. You collect evidence of what changed, assign each change a cited financial proxy, subtract the share that would have happened anyway, and divide by program cost, producing a ratio of social value to investment. In Sopact's framing, the operating version is Continuous SROI, where the ratio recomputes as fresh evidence arrives.
SROI is the full form of Social Return on Investment, sometimes written Social ROI. It means the social value a program creates expressed against what it cost, as a ratio such as 1:4 — four dollars of social value for every dollar invested. The framework was developed by the Roberts Enterprise Development Fund in the late 1990s, refined by the New Economics Foundation, and is now stewarded by Social Value International. Sopact operationalizes it on a persistent participant record so the ratio is repeatable rather than a one-time study.
The SROI formula is the present value of net social benefits divided by the value of inputs. In practice: sum the dollar value of every outcome, subtract deadweight and attribution, apply displacement and drop-off, then divide by total program cost. Multi-year studies discount future value to present terms. The arithmetic is simple; the credibility comes from the cited evidence and the four adjustments, which Sopact keeps parameterized on one record as a Continuous SROI.
The seven principles published by Social Value International are: involve stakeholders, understand what changes, value the things that matter, only include what is material, do not over-claim, be transparent, and verify the result. Six of the seven are about restraint rather than measurement, and most weak studies fail on 'do not over-claim' by skipping the four adjustments. Sopact treats the principles as data requirements, holding scope, proxies, and assumptions alongside the number so the study can be checked.
The four adjustments are deadweight, attribution, displacement, and drop-off. Deadweight is the share that would have happened anyway; attribution is the share other actors caused; displacement is the effect pushed onto other groups; drop-off is how value fades in later years. Skipping any is the most common reason an SROI fails external review. Sopact stores all four as named parameters so the ratio adjusts honestly as evidence changes.
There is no universal benchmark; ratios vary by sector, scope, and how the adjustments were applied, and reported nonprofit ratios commonly land between 1:2 and 1:10. A 1:3 with full adjustments is more defensible than a 1:30 with weak ones. Two ratios are only comparable if the studies used the same scope and rules. Sopact keeps the scope, proxies, and adjustments explicit so a ratio can be defended rather than merely quoted.
You price it off the grantee's case record: every service event, case note, and follow-up response resolved to one participant, so each valued outcome traces to a person and a date. Sopact calls that layer Case Intelligence, and it is what turns a funder's SROI request from a research project into a query. A grantee SROI built from program-level totals instead cannot show which participants moved, or answer the question of who did not.
Not by adding up investee ratios — each study scoped differently, chose different proxies, and applied different adjustment rates, so the sum has no defensible meaning. A portfolio SROI requires one shared data dictionary across investees, agreed before anyone calculates. Sopact calls that Portfolio Intelligence, and calls a fund ratio that drills from the portfolio total down to the participant response a Traceable Portfolio Ratio.
A forecast SROI predicts the value a program will create using planned outcomes and proxies from prior research; an evaluative SROI measures the value a cohort actually produced. Forecasts inform funding decisions; evaluations are stronger evidence and preferred for public reporting. Sopact runs both from the same outcome map on one record, so the forecast can be compared to the evaluation and the gap explained.
ROI compares financial return to financial investment; SROI extends the comparison to outcomes with no market price — improved health, increased confidence, avoided justice contact — by assigning each a cited financial proxy, and it requires beneficiaries to help define what counts. Sopact holds the proxy library and the stakeholder evidence together, which is what makes the social side of the ratio defensible.
A theory of change explains how a program produces outcomes; an SROI takes that validated chain and adds monetary values plus the four adjustments to produce a ratio. SROI builds on a theory of change rather than replacing it, and the cleanest workflow writes the theory of change first. Sopact keeps both on one record, so the SROI prices exactly the outcomes the theory of change validated.
A first SROI for a single program typically takes three to five months, most of it stakeholder mapping, evidence collection, and proxy research. Repeat studies are faster because the outcome map and proxy library carry over. With a persistent participant ID that keeps evidence linked across waves, the ratio can recompute in days rather than months, which is what Continuous SROI means in practice.
They can collect the survey data, but they cannot link a respondent's pre-program survey to the same person's post-program survey to the proxy library to the calculation, so most SROIs done in those tools end with a consultant rebuilding the link by hand from CSV exports. Sopact binds every response to a persistent ID at collection, so the integration the consultant usually rebuilds each cohort is already done.
Next: price a grantee's ratio from the case record on case management software, or roll it up across a fund on portfolio intelligence.