BW Staff RewardsBetter Work, rewarded

HomeBlogMeasuring whether it works

Measurement

How To Tell Whether Your Rewards Program Is Working

Four numbers worth watching, three that look useful and are not, and one question your engagement survey is not going to answer.

Published 14 July 2026 Reading time 8 minutes Category Measurement

At renewal someone will ask whether the program is working. The honest version of that answer has two parts: whether people are using it, which is easy to measure, and whether it changed anything, which is not.

Do the easy part properly first. Most programs fail the easy part and never get to the hard one.

The Four That Matter

Track monthly, per site
MetricDefinitionHealthy
ActivationAccounts set up, as a share of the team fileAbove 80% by day 30
Monthly activeOpened the app at least once in the monthAbove 50%
Recognitions per headRecognitions sent ÷ headcount2 or more per month
Redemption ratePoints redeemed ÷ points issuedAbove 80%

Track them per site, not just company-wide. The average hides the thing you most need to know, which is that two of your eleven locations stopped using it in March and nobody noticed.

A company-wide average of 62 per cent can be nine sites at 80 and two sites at zero. Those are completely different problems.

The Three That Mislead

  • Total points issued. Goes up when you spend more. Says nothing about whether the spend did anything. It is a measure of your generosity, not of the program.
  • Logins. Inflated by password resets and by people checking a balance they cannot spend. Monthly active users who did something is the better version.
  • Satisfaction with the program. Almost always positive, because the people who answer are the people who use it. The ones who never activated are not in your sample.

The Harder Question

Whether the program changed retention, safety or service is a genuinely difficult question, and anyone offering you a clean causal number is selling something.

What you can do is compare in a way that is honest about its limits:

  1. Compare sites, not periods. If the program rolled out to some locations before others, that is the closest thing to a natural comparison you will get.
  2. Compare high-usage to low-usage sites within the same period. Confounded — good managers produce both high usage and good retention — but still informative if you say so out loud.
  3. Look at twelve-month retention for staff who received recognition in their first quarter versus those who did not. Same caveat, same value.

None of those establish causation. All of them are better than a satisfaction score, and all of them are honest about what they are.

The Question A Survey Will Not Answer

Ask the people who left. Not in an engagement survey — they are gone — but in the exit conversation, where the question is usually about pay and management and rarely about anything else.

The version worth asking is not "did you like the rewards program". It is "did you feel your work was noticed here". The answer to that is what the program was bought to change, and it is not something a platform dashboard can report on.

A Reporting Cadence That Survives Contact With Reality

Who looks at what, and when
FrequencyWhoWhat
MonthlyProgram ownerThe four metrics, per site. Ten minutes.
QuarterlyFinanceCommitted, issued, redeemed. Accrual check.
Half-yearlyHR leadershipSite comparison and retention cut
AnnuallyEveryoneWhether to continue, and at what budget

The monthly review is the one that matters and the one that gets dropped first. Ten minutes a month is what stops a site quietly going dark for a quarter.

What Good Looks Like At Twelve Months

Activation above eighty per cent and holding. Monthly active above half. Recognition volume flat or rising rather than decaying after the launch spike. Redemption above eighty per cent, meaning people can reach something they want. No site sitting at zero.

If all of that is true, the program is doing what a rewards program can do. Whether that is worth the budget is a judgement, not a metric — but at least it is a judgement made on numbers that mean something.

Want a read on the program you already have?

Send the four numbers and we will tell you what we would look at next.