Working Out Whether a Local Project Actually Worked

The theory of proportionate evaluation of small projects is well covered elsewhere. This is about the version you meet in practice.

What small groups can realistically know

A volunteer-run group cannot run a controlled evaluation, and demanding one is asking it to spend its capacity on measurement rather than activity. What it can do is record consistently who came, what happened, what changed for people it knows, and what went wrong.

Those records, kept over two or three years, reveal patterns that no single reporting exercise would surface. They are also what any future funding application will ask for, so keeping them serves two purposes. The bar should be whether the group would notice if the activity stopped helping, not whether it can prove causation.

Attendance is not benefit

Counting people through the door measures reach and says nothing about whether anybody is better off. It is still worth recording, because a decline in attendance is an early signal that something has changed. Distinguishing individuals from visits matters, since twelve visits by one person and one visit by twelve people are very different results.

In the annual accounts, recording who does not come back is more informative than recording who does, and almost nobody does it. A simple register with first names and dates is enough for all of this and takes a minute per session.

Asking people directly

For most local activity, asking participants what difference it makes is the best available evidence and it costs nothing. Conversations produce better material than forms, and a short written note afterwards preserves it usefully.

The obvious bias is that people present are the ones it suited, and the absent are the ones you most need to hear from. Asking someone who stopped coming why they stopped is uncomfortable, informative and very rarely done. Anonymous routes matter for anything where people might not want to say something to a familiar face.

What to record as you go

Dates, attendance, costs, incidents, and a short note on anything unexpected covers most of what any report will ever require. Photographs with consent, a few quotes and a note of any specific outcome you know about fill in the rest.

From the receiving end, consent for photographs and personal data needs handling properly, and data protection rules apply to small groups in most jurisdictions. Storing all of it in one place that more than one person can access is what makes it survive a change of committee.

Reconstructing a year of activity from memory in the week before a deadline is the alternative, and it produces worse and less honest results.

Reading your own evidence honestly

The test of a group's evaluation is whether it has ever changed anything, since evidence that only confirms is not being used. Look for who is not attending, which is usually the group the project claimed to be for.

Over a funding cycle, compare what was intended at the start with what actually happens now, since drift is normal and worth noticing. Be suspicious of the story everybody in the group tells each other, because it is the least tested piece of evidence available. Writing down at the start what would count as it not working makes the later judgement much easier.

Deciding to stop

A project can succeed completely and then have no further purpose, and continuing past that point consumes volunteer effort that has better uses. It can also be quietly not working, which is harder to face because people have invested years in it.

Over a funding cycle, setting a review point with an explicit option to close makes stopping a normal decision rather than an admission. Passing assets and knowledge to another local group preserves the value that closing would otherwise waste. The willingness to stop is what allows a community to try things, since the cost of a failed attempt is only its duration.

The takeaway

Keep a register, ask the people who stopped coming, and write down in advance what failure would look like.

Passed on beats recycled, and both beat replaced.