Most CSI reports are, on close reading, activity reports wearing the language of impact. They tell you how many learners attended a programme, how many rand were disbursed, how many events were held. All of that is legitimate information. None of it, by itself, tells you whether anything actually changed for the people the programme was meant to serve. The confusion between the two is one of the most persistent and most fixable problems in South African corporate social investment.
Fixing it does not require a research department or an academic evaluation unit. It requires a results chain that is thought through before the spending starts, not reconstructed afterwards to justify it.
Outputs are not outcomes
An output is something a programme did: learners enrolled, workshops delivered, materials distributed. An outcome is something that changed as a result: learners who gained a qualification, found employment, or improved a skill they can demonstrate. The two are easy to conflate because outputs are simple to count and outcomes are not — and a large output number is genuinely seductive in a year-end report. "Two thousand learners trained" reads impressively on a slide. It says nothing about how many of those learners finished the training, applied what they learned, or are better off a year later. A company that reports only outputs may be running an excellent programme or a mediocre one; the number alone cannot tell the difference.
This is not an argument against tracking outputs — they are necessary operational data. It is an argument against mistaking them for evidence of impact, which is a different and harder thing to produce.
Start with a Theory of Change
The fix begins before implementation, with a Theory of Change: an explicit statement of the if-then logic connecting what the programme will do to what it expects to result. If we provide structured mentorship alongside bursary funding, then dropout rates should fall, because a documented driver of dropout — lack of academic and personal support — is being addressed. Written out plainly, a Theory of Change forces a programme design team to be honest about assumptions that are usually left implicit, and it gives everyone downstream — implementers, funders, evaluators — a shared reference point for what success is supposed to look like, agreed before anyone has an incentive to move the goalposts.
Programmes designed without this step tend to discover their assumptions only when something goes wrong, at which point it is too late to correct course cheaply.
A practical measurement stack
From a Theory of Change, a practical measurement stack follows in a fairly predictable sequence. A baseline, captured before the intervention starts, so later change can be measured against a real starting point rather than assumed. A small set of indicators tied directly to the Theory of Change — not everything that could plausibly be measured, but the handful of things that would actually confirm or challenge the logic. A defined method of data collection, applied consistently so results are comparable across cohorts and years. A verification step, so reported figures are not simply self-reported by whoever has an interest in them looking good. Periodic evaluation, to step back from the indicator-level detail and ask whether the programme as a whole is achieving what it set out to. And reporting that presents this honestly to the board or funder, including where results fell short of expectation.
Each of these steps is achievable without specialist tooling, provided it is planned for from the outset rather than bolted onto a programme already underway.
Honesty as strategy
The temptation in impact reporting is to lead with the best-looking number and let it stand in for the whole story. This is usually a mistake, and not only an ethical one. Boards, funders and increasingly the public are more sophisticated readers of impact claims than they once were, and inflated or selectively reported figures tend to be recognised as such over time, at real cost to credibility. Publishing what was actually verified — including modest results, and including the occasional programme that did not work as intended — builds a track record that stakeholders learn to trust precisely because it is not always flattering.
That trust compounds. A CSI function known for honest, evidenced reporting finds it easier to defend its budget, attract co-funding, and be taken seriously in strategic conversations than one known only for producing impressive-sounding numbers no one quite believes.