Hundreds of people and enterprises from around the world have been sharing their understanding of impact. There is consensus that impact has five dimensions: what, how much, who, contribution and risk.

How do people experience impact?

Research across Myanmar, Bolivia, Belgium, the US and the UK – with all kinds of different people – has shown that various things really matter to us when we think about the impact we want from engaging with an enterprise.

  • We generally want to participate in defining the impact that enterprises have on our lives.
  • We may describe impact in terms overcoming challenges or achieving certain outcomes. Whether framed as a challenge or an opportunity, we have a view about which products, services or jobs would be more effective than others. We also broadly know how vulnerable, disadvantaged or underserved we are in relation to different types of outcome – for example, being healthy but poorly educated – and feel improvements in vulnerable aspects of our life are more impactful than in aspects where we are already well-served.
  • Many of us agree that a product, service or opportunity that has a positive impact is most impactful if it is the only available option. It does not immediately lose its value if something else meets the same need but becomes less impactful if it is aggressively competing with other products or services that offer the same (or greater) benefit.
  • It matters whether or not a product, service or job creates the change we were looking for. We are often willing to take higher risks that our experience will differ from our expectation when the potential benefit is greater.

Understanding how people (and the planet) directly experience impact is the foundation of impact management.


What is our shared understanding of impact?

In finance, we have developed shared fundamentals, which include risk, return, liquidity, volatility – and asset classes, which group investments with similar characteristics – to describe and manage against our respective financial goals. Financial capital flows, and the investment management ecosystem has grown, not just because we have common accounting standards but because we have evolved a shared way of communicating our financial goals and understanding of performance. It would be impossible to uphold any notion of “fiduciary duty” if we didn’t.
For impact, we now have a shared understanding that impact refers to material effects experienced by people and planet, both positive and negative – and that to understand and manage impact, we need to consider five dimensions of any effect:


  1. We relate the effect we experience to a certain outcome, and we judge how positive (or negative) and how important that outcome is to us, based on our own preferences, on expert opinion or public consensus.
  2. EXAMPLE: We may relate our experience of secondary education to our ability to have financial security and we may think that financial security is a positive and important outcome. We may decide that an outcome is important based on our own opinion, or guided by professional experts, or through shared consensus like the Sustainable Development Goals or Social Progress Index.


  3. We assess how much of an effect occurs by considering its significance:
    • How deep the effect is, based on whether the effect is a big or small driver of the outcome
    • How many people the effect occurs for, based on data about the number of people experiencing the effect
    • How long the effect lasts for, based on data about the time from beginning of the effect to end of the effect
    • How quickly the effect occurs, based on data about the time it takes for an enterprise to generate the effect
    EXAMPLE: Data may show that secondary education is a big driver of financial security or that a statistically significant number of people perceive secondary education to contribute ‘a lot’. We may receive feedback from alumni that their ability to secure a good job with the skills they gained from school has endured for many years.


  4. We define who we are and we judge how well-served we already are based on whether we are already experiencing - or have the opportunity to experience - the outcome that the effect relates to. Since different effects can lead to the same outcome - and it’s the outcome that matters to us - this is what we focus on when thinking about whether we are well- or under-served.
  5. EXAMPLE: We may be a farmer’s son in Nigeria and be more or less underserved when it comes to our ability to have good health; or a girl who speaks English as a second language in the United States, with more or less access to financial security. For environmental issues, we may be thinking about certain species and how near or far they are from endangerment.


  6. We assess whether the effect is better or worse than the effect that would likely otherwise occur, based on whether the effect:
    • Leads to more or less important positive or negative outcomes, and/or
    • Is more or less significant (in terms of depth or the number of people it occurs for or how long it lasts for or how long it takes to occur) and/or
    • Occurs for people (or planet) who are more or less underserved than those currently experiencing it
    EXAMPLE: A new health service might contribute the same amount to our physical health as a service we are currently using but contribute less to negative greenhouse gas emissions for the planet. It might be likely to contribute more or less significantly to our physical health than if we continue to use our current service. It might reach other people who do not currently have access to the service at all.


  7. We assess the likelihood that the effect is different from our expectation.
  8. EXAMPLE: An innovative health service might make our physical health worse when we expected it would make it a little better – but it could also make our health significantly better than expected. Or, it might make our health better but lead to negative effects for other people or the planet.

These five dimensions help us all to understand our effects on people and planet.


Dimensions (2) v2@1.5x Created with Sketch. RISK Which risk factors are material, and how likely is the effect different from the expectation? WH A T What outcomes does the effect relate to, and how important are they to people (or the planet) experiencing it? CONTRIBUTION How does the effect compare and contribute to what is likely to occur anyway? HOW MUCH How much of the effect occurs in the time period? WHO Who experiences the effect, and how underserved are they in relation to the outcome?

Do we have to manage everything that people and planet experience?

No – impact refers to material effects experienced by people and/or the planet.

Material effects are those that relate to important positive or negative outcomes (what), are significant (how much) and occur for groups of people and/or the planet who are not already well-served in relation to the outcome (who).

Managing impact involves figuring out which effects are material — and then trying to prevent the negative and increase the positive.

For example, a business may assess that their current mix of employees does not reflect the diversity of the community in which they operate. They might decide that this is a material negative effect because, as a large business in the local community, they have a significant effect (how much) on inequality (what) among the local population (what). The same business may assess that their healthcare product for women who are currently suffering from ill-health is not having a material effect as these women will experience only a marginal effect (how much) on their physical health (what) and are easily able to afford alternative treatment to achieve the same level of good health (who).


What information do we need to understand impact?

We collect and analyse data about our effects on people and planet, so that we can assess which effects are material. Data, analysis and assessment are all forms of information that we use to understand and communicate impact and make decisions about how we run our enterprise (or portfolio of enterprises).

We collect information as a ‘set’ across all dimensions of impact so that we can compare our impact to those with the same impact goals, alongside financial performance. Information about impact experienced by people and the planet cannot be understood through one or two dimensions out of context of the others. Over time we might be able to build up enough evidence to illustrate how a difference in one or two dimensions of impact affects the result across other dimensions but, for now, information is useful for managing impact if kept in the context of all dimensions.

Dimensions (1) v2@1.5x Created with Sketch. WH A T HOW MUCH WHO CONTRIBUTION RISK • Type of outcome(s) • Importance of outcome(s) Well-served Underserved Much worse than what is likely to occur Much better than what is likely to occur Low risk High risk Marginal effect For few Short-term Slowly Deep effect For many Long-term Quickly Important negative outcome(s) Important positive outcome(s) Neutral outcome(s) • Depth of effect in time period • Number of people affected in time period • Time period effect lasts for • Time taken for effect to occur • Demographic data • Environmental data • Geographic data How much of the effect occurs in the time period? What outcomes does the effect relate to, and how important are they to people (or the planet) experiencing it? Who experiences the effect, and how underserved are they in relation to the outcome? How does the effect compare and contribute to what is likely to occur anyway? Which risk factors are material, and how likely is the effect different from the expectation? Benchmarked performance across WHO, WHAT, and HOW MUCH Risk factors, e.g. evidence risk Assessment Data Analysis

Imagine two investment opportunities, both aiming to improve educational outcomes for kids by increasing the student transition rate from one level of school to the next. Both provide counselling to 11-16 year olds and they even work in the same city. But Investment A is a proven intervention to re-motivate teenagers who are skipping school, while Investment B is testing a new approach to helping teenagers with learning disabilities. A’s transition rate shoots up from 50% to 85% and the investment generates attractive financial returns, while B’s moves from 50% to 65% and the financial return doesn’t justify the risk taken. If you benchmark what A achieves versus B (the student transition rate and the financial risk-adjusted return), what have you really learned? That A is more efficient than B – and resources should therefore be diverted away from B? Or just that B’s goal was to take a higher level of risk to impact a harder-to-reach demographic? By comparing these two opportunities without data across the other dimensions – for example data on who is being affected – we would diminish our impact. It could lead us not to make that investment B at all, pulling resourcees away from our toughest problems.

When do we collect information to understand our impact?

Impact management is the process of collecting, analysing and assessing data about the actual impact that people and planet experience when they engage with an enterprise (or portfolio of enterprises). Based on our understanding of the impact we are having, and our intentions and constraints, we can set financial and impact goals and design an enterprise, or portfolio of enterprises to meet them. Since the initial impact goals we set are often based on incomplete information and are different to the impact that people and planet go on to experience, we often redesign an enterprise, or portfolio of enterprises, to achieve our goals or decided to re-set our goals


Who bears the cost of collecting information about impact?

Much of the data we collect to understand the five dimensions of impact is data that is also commercially useful. For example, demographic information (who), an understanding of the significance of effects we are having on someone’s life (how much) and how our what we are doing compares to what the market is doing (contribution) can help us to tailor, improve and differentiate our product or the quality of our employment.

EXAMPLE: Demographic information (who), an understanding of the significance of effects we are having on someone’s life (how much) and how our what we are doing compares to what the market is doing (contribution) can help us to tailor, improve and differentiate our product or the quality of our employment.

We can therefore embed much of the collection of new information about our effects on people and the planet into our general management approach and the cost is a legitimate use of capital from investors or funders. When the field as a whole will benefit from a deeper understanding of an enterprise’s impact, philanthropic or government sources can play an important role in helping fund studies. We might choose to do a deeper study of impact to understand what impact achieved can be directly attributable to the enterprise, or whether customer feedback is sufficiently unbiased, or how the enterprise interacts with a wider system.

Where we are unable to collect new information about all of our material effects, we have to rely on existing information about effects that an approach has had historically, taking into account any differences in context.  If this is what we are doing, it helps everyone if we are transparent about what assumptions we are making and how we are seeking to reduce the evidence risk we are therefore taking (and/or explain why the level of risk is justified by our other goals). This helps others we partner with (e.g. investors or funders) to assess their own level of impact risk.

EXAMPLE: If we are assessing our material effects on people and planet based on information about the effects that a comparable approach has had historically, it helps everyone if we are transparent about what assumptions we are making, what existing evidence we are relying on and how this affects the impact risk we are taking.

What is ‘good’ information?

There are some rules of thumb that help us to identify good information:

  • Completeness: consider whether the data is good quality, and whether it has been collected from a range of relevant stakeholders.
  • Accuracy: consider whether the information has been substantiated by a counterfactual, and if it has been subject to external review or was the product of an independent assessment.
  • Relevance: consider whether the information directly corresponds to the same impact goals across all the dimensions of impact.

When we examine an existing evidence base, the more numerous the individual studies (or sets of information) that demonstrate these features and share the same findings, the stronger your evidence base.

These rules of thumb can be converted into what are known as ‘Standards of Evidence’ to help us determine what level of confidence we can have that the information we collect is telling us about impact.

  1. At a minimum, we collect information that describes the chain of events that we intend to occur, what material effects result and what impact is therefore expected

    EXAMPLE: “I have access to maternal healthcare services, which enables me to receive treatment and advice, giving my unborn child and I the best chance of being healthy, which is essential for us to have full and happy lives”.

    However, recognising that people and planet will likely experience other impact too, both positive and negative, we also draw on existing information about the material effects that comparable approaches delivered in comparable contexts to help us judge other effects we might want to manage to mitigate impact risk or improve the impact we can have.

    EXAMPLE: By consulting existing research on the effects of offshore wind farms, we might learn that some marine animals experience negative impact due to the noise and regular disturbance. This information helps us to dedicate resource to trying to mitigate these negative outcomes.
  2. Collecting information directly from all people and planet experiencing material effects gives us a much greater level of certainty that positive impact is occurring, and therefore what we need to manage, without knowing who or what caused it.

    If there is a strong evidence that our activities and outputs lead to the desired outcomes in comparable contexts, we can often rely on activity and output data as proxies to help us to learn about our current performance and improve it. But good impact management at some point goes beyond proxy measures of the expected impact to check that people are experiencing what proxies suggest they are. Direct customer feedback on the impact they are experiencing is a good example of how enterprises collect this type of data cost-effectively.

    EXAMPLE: Maternal health clinics have tracked that 470 mothers received treatment in 2008, 89% of whom safely delivered a healthy child and 97% of whom reported satisfaction with the service.
  3. Using a control, or comparison, group helps to establish causality (where possible)

    EXAMPLE: A survey of a similar group of mothers in a neighbouring region reported a 65% safe, healthy delivery rate in the same period, while the maternal mortality rate is 43% below the regional average.
  4. An independent evaluation helps to confirm (or contest) the assessment, including removal of bias, and often enables a greater understanding of the outcomes experienced.

    EXAMPLE: An evaluation shows that local schools report better average attendance and attainment rates for the children born since provision of the service.
  5. Being clear about how context affects the delivery of these outcomes enables consistent replication.

    EXAMPLE: There are multiple drivers of positive maternal health outcomes - such as access to good nutrition or a sanitary delivery environment - and people in different contexts will be underserved in different ways in relation to this outcome. This will influence the effectiveness and/or applicability of post-natal care.


How do we share information with others?

Enterprises collect information that enables them to make decisions about whether and how to improve performance or re-allocate resources. When this information across all dimensions is communicated as a “set” to an investor or funder, they will be able to use the same set of information to make decisions and improve their own allocation of resources. There is a growing number of software platforms that enable investors and funders to collect and analyse either data 'sets' (to support their investees and improve their allocation of resources) or to aggregate individual metrics (for communication purposes). See resources below.



Sometimes there are too many sets of information for us to analyse and use for decision-making. For example, a passive asset owner with many underlying investments may choose to conduct due diligence on the strength of an intermediary’s impact management processes instead. Being transparent about what our impact management process involves is therefore very helpful for everyone and we anticipate accredited assessments of this kind in future.

Some investors, funders and asset owners may look for aggregated figures across only one or two dimensions –for example the total number of people, or number of low-income people,accessing services that relate to good health and well-being. This sort of information can give a ‘flavor’ of a type of outcome or number of people affected. Since it typically includes a mix of performance across the other dimensions (for example, the depth of improvement in people’s health), it cannot help us to learn whether one enterprise, or portfolio of enterprises, is performing better than another, and therefore improve. It is ideal if enterprises can therefore extract this sort of information from the data sets they collect anyway, so that they are always focusing their limited resources on collecting data that enables us all to learn and improve our impact.

See here for more about reporting to investors.

As we understand what different people want and need – and which enterprises do and do not work – we contribute to existing bodies of information (including our own). By contributing to existing bodies of information, we either reinforce the validity of existing information (does the information we have support what others have found?), or challenge it. This helps us all to set and re-set our goals and design or identify approaches with the greatest likelihood of success.


What does this mean for investors?

Each section of this website has commentary aimed at investors. This is not because the website is aimed solely at investors. It is the combined efforts of many different disciplines and is therefore relevant to everyone. But many investors – and the professional services firms that support them – are newer to managing their impact than the wide range of charities, social enterprises and businesses, foundations, social scientists, civil servants and evaluators, who have been doing this work for so many years. We have therefore included information specifically for investors where useful.

The type of impact information that investors rely on typically comes from the enterprise generating the impact and will therefore vary according to what level of evidence the enterprise finds is necessary and proportional to understanding their effects on people and the planet. The types of information investors can rely range from:

  • Activity or output data that serves as a proxy for outcomes, because there is an existing evidence base that the activities or outputs translate into the outcome in comparable circumstances
  • EXAMPLE: Number of patients receiving insulin for diabetes
  • Direct measurement of change in outcomes
  • EXAMPLE: % change in Body Mass Index,
  • Direct feedback from people experiencing effects
  • EXAMPLE: A survey that captures customer feedback that “my quality of life has improved significantly”
  • External evaluation that removes bias, considers attribution and/or uncovers other effects.

Investors collecting all these types of information will consider how good the information is and how applicable it is for their context.


Give me some examples

See how the risk dimension has been helpful to Acumen and NPC to understand the risks they are willing to take, which risk factors matter most to them and how they manage to risk. If you are an investor or funder you may be interested to see how risk can be managed at a portfolio level

Read more from our partners on our blog