Twelve

SROI Checks

A practical audit guide for defensible SROI

Short, usable, and culturally grounded, this guide helps you design credible SROI, claim only your fair share of impact, and keep the numbers accountable. Download it, read it online, or use the checklist to stress-test your next analysis.

Image of the cover of a guide to perform SROI audit

Part A — Design & Scope

  • Theory of Change — Begin with purpose, relationships, and duties; make the logic testable.

  • Stakeholder Mapping & Inclusion — Map participants, whānau, funders, partners; document who was included and why.

  • Materiality — Decide what matters using both decision and values lenses.

  • Outcomes ≠ Outputs — Track change, not activity; balance quantitative and qualitative evidence.

Part B — Causality & Counterfactual

  • Attribution — Allocate impact shares across contributors; don’t assume 100%.

  • Deadweight — Estimate what would have happened anyway first.

  • Displacement — Test for “same total, new split” effects.

  • Duration & Drop-off — Keep only the retained portion each year; set an end-point.

Part C — Valuation & Reporting

  • Discounting — Convert multi-year values to present value using an appropriate social rate.

  • Monetisation & Valuation Sources — Use credible proxies with sources and base years; avoid double-counting.

  • Headline Results (NPV & SROI) — Report NPV and the SROI ratio with context.

  • Sensitivity, Transparency & Assurance — Show what moves the result and make the work auditable.

Why this guide

Strong SROI rests on clear design, fair causal claims, and careful valuation. These checks help you hold space for context and culture (including whakapapa and tikanga), keep claims modest, and keep results repeatable and transparent. 12 SROI CHECKS

Who it’s for

  • Boards & executives wanting defensible results they can stand behind.

  • Analysts & evaluators building SROI models and reports.

  • Funders & partners reviewing results and assurance.

How to use it

  • Read straight through to sharpen practice.

  • Dip into a specific check when you need a quick decision aid.

  • Run a status audit (Applied / Partial / N/A) across all twelve, then improve the weakest links. 12 SROI CHECKS

THE GUIDE, UNPACKED — ALL 12 CHECKS

Everything in the PDF, translated into scannable cards. Open any card for the summary, detailed guidance, and a status selector.

THEORY OF CHANGE

  • 𝗣𝗨𝗥𝗣𝗢𝗦𝗘 𝗕𝗘𝗙𝗢𝗥𝗘 𝗦𝗣𝗥𝗘𝗔𝗗𝗦𝗛𝗘𝗘𝗧𝗦.


    SROI 101 (1/12) – 𝘛𝘩𝘦𝘰𝘳𝘺 𝘰𝘧 𝘊𝘩𝘢𝘯𝘨𝘦

    A solid SROI starts with a Theory of Change: a testable chain that links what you invest to the changes you expect, for whom, and why. It turns values into a falsifiable logic model, so decisions can be challenged and improved—not just documented.

    The chain

    • Inputs — people, relationships, assets, capital, place.

    • Activities — programmes, services, governance choices.

    • Outcomes — specific changes in behaviour, status, or condition (with indicators).

    • Long-term impact — sustained effects at person, community, or system level.

    Minimum spec

    • Clear purpose and beneficiaries.

    • Stated assumptions (“because…”) and key risks.

    • Outcome indicators with baselines and targets.

    • Time horizon for when change is expected.

    • Distribution: who gains, who bears costs.

    Decision test
    Use the chain as a filter for every proposal:
    If we do X, we expect Y for Z within T, because assumption A. We’ll know it worked if measure M moves from baseline B to target R.

    Common failure modes

    • Vague outcomes (really outputs in disguise).

    • No counterfactual thinking baked in.

    • Beneficiaries unspecified; distribution ignored.

    • Missing measures, so nothing is falsifiable.

    Make it operational

    • Keep a one-page ToC that teams recognise.

    • Tie each outcome to evidence sources and data collection.

    • Revisit quarterly: retire links that didn’t hold; strengthen those that did.

    Audit your chain this quarter: surface the weakest link, name the assumption behind it, and set the next piece of evidence you’ll collect.

STAKEHOLDER MAPPING

  • 𝗦𝗧𝗔𝗥𝗧 𝗪𝗜𝗧𝗛 𝗧𝗛𝗘 𝗪𝗛𝗢, 𝗡𝗢𝗧 𝗧𝗛𝗘 𝗛𝗢𝗪 𝗠𝗨𝗖𝗛.

    SROI 101 (2/12) - 𝘚𝘵𝘢𝘬𝘦𝘩𝘰𝘭𝘥𝘦𝘳 𝘔𝘢𝘱𝘱𝘪𝘯𝘨 & 𝘐𝘯𝘤𝘭𝘶𝘴𝘪𝘰𝘯

    True impact starts by asking: ko wai mātou? — who are we, and who travels with us?

    𝗠𝗔𝗣 𝗧𝗛𝗘 𝗪𝗛𝗔𝗡𝗔𝗨 𝗪𝗘𝗕
    • Ā roto – kaimahi, trustees, shareholders
    • Ā waho – whānau, hapū, iwi, mana whenua
    • Taiao – whenua, awa, mahinga kai, taonga species
    • Mākete – customers, suppliers, funders, regulators

    Each node carries its own mana, risks, and aspirations. Think of them as concentric circles rather than a linear list. This helps to illustrate reciprocity: value flows both ways.

    𝗜𝗡𝗖𝗟𝗨𝗗𝗜𝗡𝗚 𝗧𝗛𝗘𝗜𝗥 𝗩𝗢𝗜𝗖𝗘𝗦
    • Korero tuku iho – gather stories, not just stats.
    • Kanohi-ki-te-kanohi – face-to-face hui build trust and nuance.
    • Co-design metrics – let stakeholders define what “good” looks like.

    Because an SROI is only as credible as the voices woven into it.


MATERIALITY

  • Not every outcome should be included in a social impact report.
    Materiality is a way to filter what is significant enough to measure and report.

    SROI 101 (3/12) - Materiality

    𝗧𝘄𝗼 𝗹𝗲𝗻𝘀𝗲𝘀.

    Financial materiality
    • Magnitude and likelihood of effects on performance or risk
    • Compliance and reputation with markets, funders, regulators

    Kaupapa materiality
    • Does the outcome enhance mana and uphold tikanga?
    • What is the effect on the mauri of taiao (land, water, species)?
    • Equity and distribution for whānau/hapū/iwi
    • Inter-generational oranga (long-run wellbeing)

    The sweet spot is where these lenses overlap.

    𝗛𝗼𝘄 𝘁𝗼 𝗮𝗽𝗽𝗹𝘆 𝗶𝘁
    From your stakeholder map (earlier post), list candidate topics.
    Dual-lens test for each topic (1–5 scale):
    - Significance to stakeholders
    - Consequence if ignored (mana/mauri or $/risk)
    - Strength of evidence

    Prioritise the high-scoring items.
    Document the boundary: what’s in, what’s out, and why.

    𝗖𝗼𝗺𝗺𝗼𝗻 𝘁𝗿𝗮𝗽𝘀
    • Measuring what’s easy, not what’s valued
    • Using only the financial lens
    • Spreading effort across too many “priorities”

Outcomes

  • Count changes - not outputs.
    Outputs ≠ oranga (enduring wellbeing).

    𝘚𝘙𝘖𝘐 101 (4/12) – 𝘖𝘶𝘵𝘤𝘰𝘮𝘦𝘴

    Outputs are what you deliver. Outcomes are what shift in people, taiao and systems.

    𝗢𝘂𝘁𝗽𝘂𝘁𝘀 (examples)
    Workshops held, seedlings planted, hours mentored, whānau attending.

    𝗢𝘂𝘁𝗰𝗼𝗺𝗲𝘀 (examples)
    - Whānau & work: % in sustained employment at 6/12 months; median wage uplift; improved financial security.
    - Cultural: increased participation in tikanga/te reo in workplace or community contexts.
    - Taiao: improved water quality (E. coli/clarity), native seedling survival rate, enhanced mahinga kai access.

    𝗛𝗼𝘄 𝘁𝗼 𝗽𝗶𝗰𝗸 𝗴𝗼𝗼𝗱 𝗶𝗻𝗱𝗶𝗰𝗮𝘁𝗼𝗿𝘀
    - Start from material topics (Post 3), not from what’s easy to count.
    - Define a baseline and the direction of change (↑/↓).
    - Time-bound the window (3, 6, 12 months; multi-year for taiao).
    - Triangulate evidence: admin data + surveys/interviews (kōrero tuku iho) + observation.
    - Disaggregate where it matters (age, rohe, hapū, gender).
    - Keep a balanced kete: quantitative + qualitative, near-term + long-term.

    𝗖𝗼𝗺𝗺𝗼𝗻 𝗽𝗶𝘁𝗳𝗮𝗹𝗹𝘀
    - Vanity metrics (activities dressed up as outcomes).
    - One-off snapshots without follow-up.
    - No counterfactual/baseline, so “change” is guesswork.

ATTRIBUTION

  • Claim only YOUR impact..

    SROI 101 (5/12) – 𝘈𝘵𝘵𝘳𝘪𝘣𝘶𝘵𝘪𝘰𝘯

    Attribution asks a simple question: what proportion of the change can your kaupapa fairly claim?

    Impact is usually shared between your mahi, partners, whānau effort, public services and wider context.

    𝗛𝗼𝘄 𝘁𝗼 𝗲𝘀𝘁𝗶𝗺𝗮𝘁𝗲 𝗶𝘁 (fast, defensible)

    1. Map contributors (from Post 2/12): You - Partners/providers - Whānau effort - Public services/policy - Market/context.

    2. Gather evidence of contribution:
    - Participant surveys (e.g., “Which supports most helped this change?”)
    - Admin data on dosage/intensity (hours, touchpoints)
    - Hui/interviews with partners; document causal roles
    - Simple comparison where possible (e.g., similar group not receiving your mahi)

    3. Allocate shares that sum to 100 %. Be conservative; justify each % in one sentence.

    4. Write the rule you used (so others can repeat it).

    A little example:
    - Observed improvement = 10 percentage points.
    - After deadweight 30 % and displacement 10 %, remaining change = 6 pp.
    - Your evidenced share = 50 % → claim = 3 pp.
    - (Only the 3 pp enters your SROI.)

    𝗪𝗵𝗮𝗶 𝘄ā𝗵𝗶 𝘁𝗶𝗸𝗮 - 𝗳𝗮𝗶𝗿 𝘀𝗵𝗮𝗿𝗲 𝗿𝘂𝗹𝗲𝘀
    - Never assume 100 % unless you ruled out other causes.
    - Count whānau/self-help—it’s real contribution.
    - Keep a paper trail: data used, method, rationale.
    - Sense-check with partners; adjust if they hold key pieces of the puzzle.

    𝗖𝗼𝗺𝗺𝗼𝗻 𝗽𝗶𝘁𝗳𝗮𝗹𝗹𝘀
    - Double-counting the same change across organisations
    - Confusing attribution with deadweight or displacement (upcoming posts)
    - Inflating claims because “we funded it”

DEADWEIGHT

  • 𝗦𝗨𝗕𝗧𝗥𝗔𝗖𝗧 𝗧𝗛𝗘 𝗜𝗡𝗘𝗩𝗜𝗧𝗔𝗕𝗟𝗘.

    SROI 101 (6/12) – 𝘋𝘦𝘢𝘥𝘸𝘦𝘪𝘨𝘩𝘵 (the counterfactual)

    Deadweight is the share of change that would have happened anyway, without your kaupapa.

    If employment has been rising across your rohe, some job gains will occur regardless. SROI counts only the lift above that tide.

    How to estimate it (fair, fast):
    - Baseline trend: what was happening before the mahi? (Use pre-intervention data if you can.)
    - Comparator: a matched group/rohe or waitlist on business-as-usual.
    - Benchmarks: published rates (e.g., natural regeneration, relapse, typical progression).
    - Informed judgement: triangulate kaimahi and whānau views with data; express a range (low/central/high).

    Treat deadweight as your first subtraction. Then consider attribution (who gets the credit) and displacement (benefit shifted from elsewhere). Keep the time window explicit — many outcomes need a 6–12 month lens; taiao often needs multi-year.

    Apply it well:
    - State the period and source for every estimate.
    - Subtract deadweight first, then apply attribution and displacement.
    - When evidence is thin, lean conservative and run sensitivity tests.

    Common traps:
    - Assuming zero because the outcome feels unique.
    - Using one blanket percentage for everything.
    - Confusing deadweight with attribution or displacement.

DISPLACEMENT

  • 𝗦𝗔𝗠𝗘 𝗧𝗢𝗧𝗔𝗟, 𝗡𝗘𝗪 𝗦𝗣𝗟𝗜𝗧.

    SROI 101 (7/12) – 𝘋𝘪𝘴𝘱𝘭𝘢𝘤𝘦𝘮𝘦𝘯𝘵

    Displacement is when your measured gains are reallocated from elsewhere rather than created anew. One side of the seesaw rises because the other side falls. The outcome count looks good; the net impact may be small.

    Where it shows up
    - Hiring in a static local labour market (your hires replace others who would have been hired).
    - Contract wins inside a fixed buyer budget (spend shifts provider, not total demand).
    - Training cohorts with capped seats (your entrants move others below the cut-off).
    - Place-based activity when regional totals are flat (share changes, not totals).
    To separate shift from growth, start with the pool. If the total opportunity is effectively constant over your measurement window, assume some portion of your uplift is displacement and show it explicitly.

    Estimating it (credible, fast)
    - Capacity check: is the pool fixed over the period (jobs, seats, budget)?
    - Totals unchanged: before/after data show constant aggregate outcomes despite your rise.
    - Comparators: matched area/provider declines as yours increases.
    - Decision evidence: buyer/employer surveys indicate substitution (“we would have contracted/hired anyway”).
    - Decompose the uplift: Observed ↑ = net new + displaced (report both).

    Apply displacement in the right order. Subtract deadweight first (what would happen anyway). Then subtract displacement (what came from others). Then apply attribution (your fair share of what remains). Be clear about unit, boundary, and the time horizon you’re using.

    Pitfalls to avoid
    - Claiming redistributed outcomes as impact.
    - Assuming 0% displacement in fixed-capacity contexts.
    - Mixing up displacement with attribution or deadweight.

DURATION & DROP-OFF

  • Durability of Impact Outcomes

    SROI 101 (8/12) – 𝘋𝘶𝘳𝘢𝘵𝘪𝘰𝘯 & 𝘋𝘳𝘰𝘱-𝘰𝘧𝘧

    Year-one gains make good headlines, but SROI asks a tougher question: how long do these gains really last? Duration tells us how many years an outcome continues; drop-off tells us how much it declines each year. Ignoring these factors risks overstating your impact and directing resources towards short-lived gains.

    Outcomes rarely stay at full strength forever:

    - Skills and confidence typically erode without reinforcement.
    - Employment and income impacts may start strong, but stability over time is critical.
    - Environmental improvements often build slowly, persist longer, but still require ongoing care.

    Knowing these patterns helps to plan and prioritise investment. Each domain’s unique drop-off profile informs resource allocation and long-term strategy.

    Getting this right requires thoughtful evidence:

    - Track cohorts by following the same people or places at regular intervals (e.g., 3, 6, 12 months, and annually thereafter).
    - Identify retention signals such as sustained employment, ongoing participation, survival rates, or repeat usage.
    - Benchmark against published data on persistence, relapse, or natural recovery, where available.
    - Use stakeholder insights and informed judgement to validate assumptions and fill evidence gaps.
    - Define a clear endpoint. Stop counting when the outcome plausibly returns to baseline or becomes negligible.

    Evidence doesn’t need to be perfect, but it must be credible. Triangulating multiple sources ensures your estimates of duration and drop-off remain defensible under scrutiny.

    To reflect this clearly in your model:

    - Start from your net Year-1 outcome (after adjusting for deadweight, displacement, and attribution).
    - Apply the annual drop-off rate each subsequent year, keeping only the retained portion.
    - Use your chosen social discount rate to reflect time value.
    - Test sensitivity with different durations and drop-off rates, and document the rationale behind your choices.

    Clearly showing your assumptions strengthens your model. Transparent logic around duration and drop-off improves the rigour and reliability of your overall SROI calculation.

DISCOUNTING

  • Discounting: Future to Present

    SROI 101 (9/12) – 𝘋𝘪𝘴𝘤𝘰𝘶𝘯𝘵𝘪𝘯𝘨

    Future benefits are not worth the same as benefits received today. Discounting converts a stream of future costs and outcomes into their value now, so options are comparable on equal terms. In practice, you’ll use the New Zealand Treasury’s social discount rate, or a clearly justified iwi-defined rate, then test the sensitivity of your results.

    Why it matters
    - Makes different time horizons comparable (1-year vs 10-year effects).
    - Reflects opportunity cost and risk: later benefits are less certain.
    - Prevents over-claiming from long tails with small annual effects.

    Discounting doesn’t diminish the importance of long-run outcomes; it simply expresses them in today’s units so decision-makers can weigh timing and scale together. It also forces transparency about whose time preferences you are using.

    Choosing the rate (be explicit)
    - Default: apply the current Treasury social discount rate (real).
    - Contextual choice: where appropriate, use an iwi-defined rate that reflects intergenerational horizons—document rationale and governance sign-off.
    - Sensitivity: publish low/central/high cases (e.g., ± 2 percentage points) to show robustness.

    Once the rate is stated, the rest is mechanics. What matters is that readers can trace the assumptions and see how much they drive the result.

    How to apply to your model
    - Start with net Year-1 benefits after deadweight, displacement, attribution, and drop-off.
    - For each later year, keep only the retained portion, then discount it back to today.
    - Do the same for costs, ensuring all figures are in the same price base year.
    - Sum discounted benefits and costs to get NPV, then present the SROI ratio.
    - Small changes in the rate can move long-horizon projects noticeably. That’s why publishing a sensitivity sweep is as important as the central estimate.

    Checks before you publish
    - Price base and inflation treatment are declared and consistent.
    - Rate source, rationale, and any departures from Treasury guidance are documented.
    - A clear sensitivity table shows how SROI changes with different rates and durations.

Sensitivity and Scenario Analysis

  • Testing the assumptions that influence your results

    SROI 101 (12/12) – Sensitivity and Scenario Analysis

    Sensitivity and scenario analyses assess how changing key assumptions. The discount rate, duration, drop-off, deadweight, and attribution affect the robustness of your SROI results. The radar chart visualises alternative assumptions, displaying how much each parameter affects the final ratio.

    What to vary (one at a time)
    - Discount rate: Treasury’s recommended rate compared against alternative rates defined by iwi or sector-specific guidelines.
    - Duration & drop-off: Adjust the length of outcome persistence or the rate at which outcomes decrease.
    - Deadweight: Vary the assumptions around what would have happened without the intervention or programme.
    - Attribution: Alter the share of outcomes credited to partners, funders, or external influences.

    Sensitivity analyses provide transparency. Clearly document the rationale behind each assumption adjustment and record the resulting changes in SROI and NPV.

    How to present sensitivity results
    - Rank assumptions by their impact on SROI or NPV.
    - Illustrate shifts in the payback period under various scenarios.
    - Highlight combinations of assumptions that maintain or fail to meet a desired SROI ratio.
    - Radar charts (as illustrated) clearly and effectively compare baseline and adjusted scenarios.

    Presenting sensitivity findings clearly helps stakeholders understand the confidence in your results. The objective is not simply to show a range, but to communicate which assumptions significantly affect outcomes and where further evidence or clarity may be required.

    Include alongside sensitivity visuals:
    - A detailed assumption table (central, low, high scenarios) with documented sources and rationales.
    - Clearly stated discount rate & price base year presented consistently in real terms.
    - Reporting of NPV and SROI ratio together, with explicit acknowledgement of outcomes not monetised.
    - A brief statement on assurance procedures, double-counting checks, and stakeholder validations undertaken.

Net Present Value and SROI Ratio

  • 𝗪𝗛𝗘𝗥𝗘 𝗧𝗜𝗠𝗘 𝗠𝗘𝗘𝗧𝗦 𝗩𝗔𝗟𝗨𝗘: 𝗡𝗣𝗩 & 𝗦𝗥𝗢𝗜

    SROI 101 (11/12) – Net Present Value and SROI Ratio.

    Once outcomes are monetised, we still need to compare the value that lands in different years. Discounting brings each year’s additional benefits and costs back to today using a social discount rate, so we’re comparing like with like.

    In the image, the yellow portion is the present value that counts today; the white portion is the value shaved off by time.

    What to calculate
    - PV Benefits: After deadweight, displacement, attribution; then apply duration & drop-off.
    - PV Costs: Include capital, operating, in-kind/opportunity costs, and renewals; keep financing flows separate from social value.
    - Net Present Value (NPV): PV(Benefits) − PV(Costs) — the welfare gain in today’s dollars.
    - SROI Ratio: PV(Benefits) ÷ PV(Costs) — dollars of value created per $1 invested.

    These steps impose discipline: only additional outcomes count, and everything is placed on the same temporal footing. They also keep welfare economics distinct from cash accounting, so funding mechanics don’t distort conclusions.

    How to interpret (without hype)
    - Report NPV and the SROI ratio together; a ratio above 1:1 signals benefits exceed costs, but magnitude and context matter.
    - Show ranges (low/central/high) and a payback year when cumulative PV turns positive.
    - Disaggregate by stakeholder/domain so readers see where value sits, not just the headline.
    - Be explicit about what’s excluded (un-monetised outcomes, negative externalities) and about any culturally specific valuations and governance/sign-off.

    Ratios are a headline, not a verdict; they help prioritise but don’t capture distribution, risk, or cultural weight. Use the results to illustrate where value is present and when it is realised; use the ratio to convey efficiency to busy readers.

    Assurance & transparency
    - State the discount rate and price base year; present in real terms.
    - Publish the valuation table and netting assumptions; avoid double-counting across domains.
    - Stress-test big movers (valuation inputs, duration/drop-off, attribution, discount rate) and show how results shift.

MONETISATION & VALUATION

  • 𝗩𝗔𝗟𝗨𝗘 𝗕𝗘𝗬𝗢𝗡𝗗 𝗣𝗥𝗜𝗖𝗘.

    SROI 101 (10/12) – 𝘔𝘰𝘯𝘦𝘵𝘪𝘴𝘢𝘵𝘪𝘰𝘯 & 𝘝𝘢𝘭𝘶𝘢𝘵𝘪𝘰𝘯 𝘚𝘰𝘶𝘳𝘤𝘦𝘴 (non-market)

    Many outcomes that matter in SROI carry no market price. Connection, safety, ecological integrity, cultural access, and yet they have real welfare value. Monetisation is a translation: from an outcome to a welfare change to a monetary metric, so that options can be compared on a standard scale.

    What belongs in non-market value
    - Use value: benefits directly experienced (e.g., safer housing, improved wellbeing).
    - Option value: maintaining the ability to benefit in future.
    - Non-use value: existence/bequest/cultural heritage (e.g., access to mahinga kai, protection of taonga species).
    - Distributional stakes: who gains? Whānau, community, environment, public services.

    Because there is no single “correct” price, transparency and fit matter more than elegance, aim for values that are credible for your context and explicit about uncertainty.

    Evidence principles
    - Context relevance first, then rigour; if you transfer values, adjust and justify.
    - Document the source and base year; express all figures in real terms.
    - Publish ranges (low/central/high) rather than a single point when evidence is thin.
    - Avoid double-counting across domains; state boundaries should be clearly stated.
    - Where culturally specific values are used, record the governance/sign-off process.

    Once you have candidate values, assemble them into a documented set that readers can audit. Bring them into the model only after you have netted out what is not additional.

    Applying values in the model
    - For each material outcome, map Outcome → Indicator → Valuation source (e.g., CBAx line, revealed/stated preference, replacement cost, or customary valuation with local endorsement).
    - Net adjustments: deadweight, displacement, attribution.
    - Apply duration & drop-off, then discount to present value.
    - Allocate by stakeholder and report the composition of benefits by domain.
    - Run sensitivity tests on key values and show how results move.