Published in Partnership with TriplePundit and in advance of the 2013 Skoll World Forum. Watch the live stream April 10-12 by clicking here.

Article Highlights:

  • Many charities are spending time & money measuring their effect, but this is pointless.
  • It fails, because it doesn’t help donors to choose between them – the single most useful thing which it could do. Impact assessment is fundamentally a comparative exercise.
  • Such comparisons between charities are possible, and do happen. But they’re still rare.

Since there are so many ways that charities and funders can use their finite resources, they must make choices: choices between competing goals (reduce homelessness or reduce child poverty), between competing approaches (reduce homelessness by providing more housing or by lobbying), and between competing target groups (provide more housing in Hull or in Halifax?)

So understanding the impact (or even potential impact) of various programmes or organisations is fundamentally a comparative exercise. The important question is: ‘what does the impact and associated risk of Programme A relative to that of Programme B suggest about how we should allocate our limited resources?’

The normal question of ‘what is the impact of Programme A’ is pretty unhelpful. Often this question is in principle impossible to answer, but even when it isn’t, it’s not helpful. Let’s suppose that Programme A is one whose impact is known: it’s an education programme whose effect is evident in children’s scores in year-end tests, and that it increases those scores by 10% on average. Let’s further suppose that we know the unit cost, which is £50 per child. (By the way, this is a great deal more clarity on impact than is normally available.)

picture1

Well great, but this still doesn’t help because the choice wasn’t between doing Programme A and doing nothing. The choice is always between competing alternatives, so at a minimum you need data on two alternatives. The key question for impact-istas is ‘how does our impact (or that of a programme we might run/fund) compare to what our impact could be if we did something else? That is, are we taking the best approach, or are we leaving value on the table?’ Therefore we need to figure out what information do we need in order to see if we’re on the best track or if we could improve.

Like most good insights, this sounds rather obvious – but you wouldn’t guess it from much of the discourse about impact.

For instance, I talk to many charities and donors who want to know what their total impact is. Suppose they knew, and let’s say that they prevented 274 people from becoming homeless, which cost £10,000 per person. So what? What does that suggest about where they could improve, which parts of their brilliant model they should share with others, and which parts of their work are superfluous (or even counterproductive) and should be dropped?

Nothing. It’s operationally useless. There is not one management decision for which those impact data are useful.

{I often wonder why people ask this question. They sometimes imply they want to report to some external entity – a funder or board perhaps. But either that entity is going to use the information, in which case it needs to be compared with something for the reasons discussed, or they’re not, in which case reporting on it is pointless. Maybe I’m missing something. }

picture2

How shall I compare thee…?

The trick is comparisons – between options which are comparable(!) Do other programmes prevent homelessness of comparable people more cheaply? Can other approaches improve students’ ability at tests more cheaply or quickly? These are useful questions because they can be used in management decisions: they address the central question above, by showing whether we’re on the best track or if (where) we could improve. Hence in several spheres – such as international development, health and education – ‘multi-arm studies’ are common, which compare the effect of several programmes.

The comparison could just be between one organisation’s performance over time. But again, it’s not useful simply to know whether impact is rising or falling: rather, what’s useful is to know how a particular change which the organisation made (changing its operational model, for instance) affected impact, because this informs the decision about whether to continue with that change or to revert.

picture3

In many spheres, much information about effects, comparative effects and comparative costs of competing approaches is available – publicly and freely – for instance from academics or other independent sources. Charities and funders looking to see whether they’re on the best track and if/where they could improve do well to use it.

Take me to your ruler

The problem of course is that often options are not comparable. The units of improved education are totally different to those of better air quality or preserved heritage or good medical research. Hence we cannot use the same ‘ruler’ for them all: each social impact ‘ruler’ only works for a few activities. [The various attempts to squash multiple dimensions into single units - to develop a ruler for any sector - are incoherent, I suspect.]

picture4

So for an organisation which runs or funds multiple types of work, the answer to ‘what’s our impact?’ will normally include several pieces in different units.

This should be no surprise. Consider the question of ‘how’s the economy doing?’ You’d expect the answer to include various quantitative measures – GDP/capita, growth, unemployment, and perhaps social mobility, inequality and even happiness – as well as some narrative assessment. Such as we find in most newspapers on most days.

As impact-istas, we need better rulers and better ways of measuring. But no ruler of itself shows whether one line is longer than another, nor how any line could make itself longer or better over time. Understanding impact may be possible in isolation: increasing and maximising impact is fundamentally a comparative exercise.