Which volunteer numbers are worth collecting

Seven in ten volunteer programs run a feedback survey. Fewer than a third can tell you what share of last year’s volunteers came back this year.

Those two figures come from the same survey of the same practitioners, and between them they describe the whole measurement problem. We collect what is easy to produce and pleasant to read. We skip the numbers that would tell us whether the program is gaining ground or quietly draining.

What programs actually count

The source is the 2023 Volunteer Management Progress Report from Tobi Johnson & Associates, built on 1,247 completed responses across 36 countries, 79 percent of them in the United States. Respondents were asked what their organization tracks.

What they track	Share of programs
Volunteer feedback surveys	71.3%
People served	47.0%
Dollar value of volunteer hours	43.4%
Event participation rate	38.0%
Annual retention rate	31.9%
New volunteer conversion rate	25.7%
Monthly churn	15.4%

Read down that column slowly. Feedback surveys, the most reported item, produce a document. People served produces a sentence for the annual report. Dollar value produces a line for the board. All three are outputs you can show someone.

Retention, conversion and churn produce something less comfortable: evidence about whether the program is working. Those three sit at the bottom.

The practice everyone skips is the same one it has always been

This is not new, and it is not a failure of the current generation of coordinators. When the Urban Institute surveyed charities and congregations on nine volunteer management practices, measuring the outcomes and impact of volunteers came dead last. Fourteen percent did it to a large degree. That work was published in 2004, so it is old, and the honest reading is that the ordering has been stable for twenty years rather than that nothing has improved.

I trained as a systems engineer before I ever ran a volunteer program, and the habit that stuck from that training is a narrow one: only instrument what you intend to act on. Every reading you take has a cost, and the cost is not the collecting. It is the attention. A dashboard with fourteen numbers on it gets glanced at. A dashboard with four gets read.

So the useful question is not what could be measured. It is which four or five numbers earn their place, and what you can stop doing on Monday.

The study that should shorten your list

Piatak and Carman surveyed 323 volunteers between January and April 2020 and published the results in the Journal of Public and Nonprofit Affairs in 2023. They tested which parts of the volunteer experience predicted two outcomes: whether volunteers would recommend the organization to someone else, and whether they intended to volunteer again.

Appropriate training was the strongest predictor of both. Volunteers who felt they had been properly trained were far more likely to recommend the organization and far more likely to come back. The inclusion and diversity climate of the organization predicted both outcomes as well. Separately, 13.4 percent of the open-ended suggestions volunteers wrote in asked the organization to be more inclusive.

Now the part that should change your tracking. Logistics did not predict either outcome. Choice of project, ease of registration, the number of activity options available: none of them reached significance. These are exactly the things most volunteer software is built to optimize and most best-practice guidance tells you to fix first.

I want to be careful about what that means. It does not mean a broken registration form is fine. It means logistics behave like plumbing. Nobody stays because the sign-up flow was pleasant. Plenty of people leave over a form that eats their evening. Fix it once and stop measuring it, because the measurement will not move anything you care about.

Four numbers, five if you run events

Annual return rate. Take last year’s volunteer list and this year’s, and calculate what share came back. This is the single most informative number in volunteer management and under a third of programs have it. It is also the number that makes every other decision arguable, because it gives you a before and an after. For what the answer usually reveals, see why a third of your volunteers don’t come back.

First-to-second conversion. Of the people who did one shift, what share did a second? This is the number that separates a recruitment problem from a welcome problem, and the two require completely different responses. If conversion is healthy and return rate is poor, you are losing people at the year boundary and should look at what happens in the quiet months. If conversion is poor, your problem is in the first ninety minutes of someone’s experience.

Hours, split by type of work. Not one total. Two or three buckets, sorted by how much judgment the role carries. The aggregate hides the thing you most need to see, which is whether you are running a program of consequential roles or a rota of interchangeable ones. The arithmetic for turning those hours into a defensible dollar figure is in the board wants a number.

Share who are very satisfied. Not average satisfaction. The share who pick the top box. NCVO’s Time Well Spent 2023, with YouGov fieldwork among 7,006 UK adults, found that 90 percent of very satisfied volunteers expected to continue against 81 percent of merely satisfied ones. An average score of “good” conceals which side of that nine-point gap your people are sitting on. More on how to get that number honestly in a volunteer survey that tells you something.

If you run events, add event participation rate, meaning the share of people who said yes and then turned up. Thirty-eight percent of programs track it, and for anyone who has ever built a crew list around a 30 percent no-show, it is worth the trouble.

What I would not bother with

The dollar value of volunteer hours is a communication number, not a management number. Calculate it once a year, put it in front of the board, then close the file. It does not go up when you get better at this and it does not go down when you get worse. It goes up when you get bigger. Forty three percent of programs track it, which is more than track whether their volunteers come back, and that ordering is backwards.

People served is a program metric wearing a volunteer metric’s coat. It tells you about the service, not about how the volunteers are doing, and you can serve more people next year while losing half your volunteers. Track it for the organization. Do not let it stand in for anything here.

Monthly churn belongs to big programs only. With forty active volunteers, one person missing a month moves the figure by a couple of points and means nothing. The 15.4 percent of programs tracking it are mostly not the ones who should be.

And drop the logistics questions from your survey. Ease of registration, number of options, choice of project. On the Piatak and Carman evidence you are spending your volunteers’ patience collecting data about the parts of the experience that do not predict whether they stay.

Someone will object that this ignores impact

They will be right, and I do not have a clean answer.

None of the four numbers above measures whether the work did any good. A program can have excellent retention, high conversion, hours weighted toward consequential roles, and be doing something that helps nobody. Volunteer Canada’s Canadian Code for Volunteer Involvement makes continuous evaluation of the impact and contribution of volunteers one of its twelve organizational standards, and it is the standard I see honored least often.

What I would say is that impact measurement is genuinely hard, expensive, and usually specific to your mission in a way no article can shortcut, whereas the four numbers above take a spreadsheet and an afternoon. The 14 percent figure from the Urban Institute is not evidence that coordinators are lazy. It is evidence that impact measurement is the one item on the list that a person with 30 percent of their week cannot simply start doing.

The limits of everything above

The VolunteerPro figures are what practitioners say they track, self-reported, and there is no audit behind them. Programs that respond to a benchmarking survey are also probably more measurement-minded than average, which would mean the real tracking rates are lower than the table shows.

The Piatak and Carman study is 323 volunteers surveyed at a single point in time, with self-reported outcomes rather than observed behavior. Cross-sectional work of that shape can tell you what travels together. It cannot tell you what causes what. Volunteers who felt well trained may simply be volunteers who were doing better in every respect.

The direction still seems safe to act on, because the cost of acting is low. Moving effort from counting registration friction to counting who came back does not require a budget. It requires deleting some rows.