What are the Five Dimensions of Impact?
The Five Dimensions of Impact are five questions you ask of any outcome: What changes, Who experiences it, How much of it happens, what Contribution the program made, and what Risk could make the result different from expected. They grew out of the Impact Management Project and are now maintained by Impact Frontiers.
They are a way of reading evidence, not a score, a certification or a template. Point them at one outcome and they show which questions your numbers answer and which they leave open.
THE SHORT VERSION
- Apply the five questions to one material outcome at a time, not to a whole organization at once.
- Tag each metric in your data dictionary with the one or two dimensions it answers, and write down the dimensions nothing answers yet.
- When a funder adds up partner results, the Who and How much tags decide how the total may be split and what it can claim.
What question does each dimension ask?
Each dimension asks one question and points to a different kind of evidence, so a single headcount can never answer all five. The table uses a fictional regional workforce fund, which funds four job-training partners (A, B, C and D), as the running example on this page.
| Dimension | Question | Evidence to seek | Workforce example |
|---|---|---|---|
| What | What changes, and does it matter? | A defined outcome and why it is significant | Placed in a job within 90 days of exit |
| Who | Who experiences it, and who is missing? | Characteristics, starting conditions, gaps | Enrolled, by gender, age and disability |
| How much | How many, how deep, for how long? | Scale, depth and duration | Unique people enrolled; wage by track; retained at 12 months |
| Contribution | What would have happened anyway? | A plausible comparison and its limits | Earlier cohorts; county data on the same definition |
| Risk | What could make the result differ? | Evidence gaps, adverse effects | Placements that do not last a year |
The video, The Five Dimensions of Impact (IMP Framework Explained), explains the framework that grew out of the Impact Management Project. Watch it with the table open, and ask which of the five questions your own metrics already answer.
What does the What dimension ask?
What asks you to name the change itself, not the activity that aimed at it, and to say why that change matters to the people involved. “Trainees served” describes reach. “Completers in paid work within 90 days of exit” describes an outcome you can check.
One activity can produce several outcomes, including unwanted ones: a job paying less than the trainee’s previous income is still a placement. Decide which effects are material and write that decision down. Impact Frontiers on the What dimension.
What does the Who dimension ask?
Who asks which people experience the outcome, what their starting point was, and who the numbers leave out. An average improvement can sit beside a smaller group that did worse, and only a breakdown shows it.
Choose the breakdown before collection, so every partner records the same categories. The fund’s dictionary splits enrollment by gender, age and disability; a partner that records age in different bands cannot join that split until someone maps its categories. Collect sensitive characteristics only when a decision needs them, and never publish a group small enough to identify a person. The Who dimension.
What does How much measure?
How much covers three things at once: scale (how many people), depth (how much each one changed) and duration (how long the change lasts). A large headcount answers scale and says nothing about the other two.
In the workforce example, unique people enrolled is scale, starting wage by training track is depth, and still in the same job at 12 months is duration. Keep follow-up coverage visible: a trainee with no 12-month answer is unknown, not retained and not lost. How much, in the current norms.
What does Contribution mean in the Five Dimensions?
Contribution asks what part the program played, compared with what would probably have happened without it. Change after a program is not the same as change caused by it: a strong local job market, other services and people’s own effort all move placement numbers.
Most teams answer Contribution with a comparison, not an experiment: an earlier cohort counted the same way, or an outside number on the same definition, such as the county median wage for the same job. Keep the enterprise’s contribution to outcomes separate from an investor’s contribution to the enterprise; providing capital does not make every observed outcome additional. Enterprise contribution guidance.
What is impact risk?
Impact risk is the chance that the real result differs from the one you expect, and the useful response is to name each risk with its evidence, the action available and who will review it. Typical risks are weak evidence, a change that does not last and effects nobody asked about.
In the workforce fund, the clearest risk is placements that do not last a year, which is why retention at 12 months is in the agreement. A confident rating should never hide an open gap; consult the current risk guidance when you use its formal categories.
How do you use the five dimensions in a data dictionary?
Tag each dictionary row with the one or two dimensions its metric answers, then read down the tags to see which questions no metric covers. The dimensions work as a guide to what kind of evidence a row is, not as five boxes to fill on every row.
A dictionary row carries the metric, its plain definition, the dimension tag, an optional standard code and the roll-up rule. The fund’s enrollment row is tagged How much and Who.
Tag all five metrics from the fund’s reporting agreement and the gaps appear on their own. The table below is the fictional fund’s tagging; copy its shape for your own metrics.
| Agreed metric | Dimensions it answers | What it leaves open |
|---|---|---|
| Enrolled (unique people, quarterly) | How much (scale) · Who | Whether anything changed |
| Completed training | How much (scale) | An output, not yet an outcome |
| Placed within 90 days of exit | What · How much | Job quality; what would have happened anyway |
| Retained at 12 months | How much (duration) · Risk | Why people left |
| Starting wage, hourly, by track | How much (depth) | Comparison with the local market |
Read down the middle column and nothing says Contribution. That is a known gap to write down, not a reason to add metrics nobody will collect; the chapter A shared data dictionary: metric, dimension, standard shows the full entry, including collection point, missing-value rule and owner.
Funders do this at the start. AHA Ventures, the investment arm of the American Heart Association, is growing its portfolio from 19 to more than 40 companies. Its head of impact drafts each company’s logic model and data dictionary during the onboarding call, from the call itself, with Sopact Sense.
PROMPT · PASTE INTO CLAUDE, CHATGPT OR YOUR AI TOOL
Below is the list of metrics we report, with their definitions. For each metric, tag the one or two Five Dimensions of Impact it answers (What, Who, How much, Contribution, Risk; for How much, say scale, depth or duration) and say in one line what it leaves open. Then list any dimension no metric answers. Rules: - Use only the definitions I give. Where a definition is too vague to tag, write "ASK: [question]". - Do not invent numbers, targets or benchmarks. - If a metric has no definition in my list, write "not in our data". [PASTE METRICS AND DEFINITIONS]
The walkthrough below takes one job-training program from its theory of change to IRIS+ metrics and the Impact Management Project’s dimensions. Watch for where each dimension lands as a field in the program’s definitions.
How do the five dimensions show up when a funder adds up results?
In a portfolio roll-up, How much decides what can be added, and Who decides how the total may be split. Both depend on every partner counting by the same dictionary row.
The fund’s placement metric shows it. Partners A, B and D reported placements within 90 days of exit: 42, 31 and 27, a portfolio total of 100. Partner C reported 55 counted within six months, a different definition, so its number is held until C confirms its 90-day count.
The Who tag works the same way. Because the dictionary row asks every partner to record gender, age and disability at enrollment, the portfolio’s enrolled count can be split by those categories; where a partner left a category blank, the split reports it as not recorded, never as zero.
The optional benchmark is the roll-up’s nearest approach to Contribution; the chapter Roll up and benchmark portfolio results covers benchmarks and held numbers in full.
Can you turn the five dimensions into a score, and how does IRIS+ fit?
You can build your own scoring rubric on the five dimensions, but it is your assessment method, not an official Five Dimensions score, and IRIS+ codes supply metrics rather than answers to the five questions. Both are useful when you are clear about what they are.
If you score, write down the rules and weights (the FAQ below has the checks). IRIS+, from the Global Impact Investing Network, catalogs metric definitions with codes; a code can supply evidence for a dimension, but choosing one does not settle importance, contribution or risk.
What can the five dimensions not tell you?
The dimensions organize questions; they do not supply evidence, and they cannot turn outputs into outcomes or a before-and-after change into proof of cause. Completed training is an output, whatever dimension you tag it with.
Placement and retention often rest on follow-up surveys, so they carry self-report and non-response: the trainees who answer may differ from those who do not. If an AI tool drafts your tags or your review, treat the draft as questions to settle and check it against the records; a person on your team decides what the evidence supports.
Start with one outcome this reporting cycle
Pick the outcome your funders ask about most and run the five questions on it before you touch the rest of your metrics.
- Choose one material outcome, such as placed in a job, and write its definition in one sentence with its time window.
- Write the five questions beside it and answer each from the evidence you already hold, marking “not in our data” where nothing answers.
- Tag the outcome’s dictionary row with the one or two dimensions its metric answers.
- Check the Who breakdown: are the same categories recorded by everyone who reports this outcome?
- Pick one comparison for Contribution, such as an earlier cohort counted the same way, and note its limits.
- List each risk with its evidence, the action available and the person who reviews it.
After the first cycle you have one outcome reviewed on five questions, a tagged dictionary row and a short list of evidence to collect next.
Frequently asked questions
Is the Impact Management Project the same as the Five Dimensions?
Not quite. The Five Dimensions of Impact were developed through the Impact Management Project, and Impact Frontiers now maintains the norms that describe them. When you cite the dimensions or use their formal risk categories, work from the current Impact Frontiers guidance rather than assuming an older summary still matches it. The five questions, What, Who, How much, Contribution and Risk, are the part most teams use day to day.
What is an example of the five dimensions of impact?
Take a job-training outcome: completers placed in paid work within 90 days of exit. What is the placement itself; Who is the trainees by gender, age and disability; How much is the count, the starting wage by track and retention at 12 months; Contribution is a comparison with earlier cohorts or county data; Risk is placements that do not last a year.
Can I add the five dimensions into one score?
You can design a rubric that scores each dimension, but it needs written rules, weights and judgment, and the result is your own method; the dimensions supply no universal additive score. Keep missing evidence visible instead of scoring it as poor, and test whether two reviewers reach the same score on the same case.
Are the five dimensions only for impact investors?
No. Investors use them to review portfolio companies, but foundations, government programs and the organizations they fund use the same questions to examine an outcome. A funded partner can tag its own metrics and see what its reports leave open. Investor contribution, meaning what the capital itself added, is a related but separate question that mainly concerns investors.
Does a high five-dimension score prove impact?
No. A score reflects the evidence you had and the rules you chose. It cannot remove uncertainty or show that the program caused a change. A causal claim needs a comparison designed for it, and even then self-reported follow-up and non-response limit what you can say. Report the score with its evidence and gaps beside it.
