To compare loyalty reward use across enrollment periods, group members by when they joined, then measure the share of eligible members who redeem within the same number of days after becoming eligible. Compare only cohorts that have completed that window. A newer group may look weaker simply because its members have had less time to use their rewards.
The useful result is not just a percentage. You need to know who was eligible, what they could redeem, and what happened to rewards that were not used. That lets you decide whether to change an offer, improve how guests learn about it, or wait for more complete data.
Choose the cohort and the clock
An enrollment cohort is a group of members who joined during the same period, such as January or February. The enrollment date places a guest in a group. The reward eligibility date is the date that guest first has a reward they can actually redeem under the program’s rules.
Those dates may be weeks apart. A points balance or a dollar-equivalent estimate is not necessarily a redeemable reward. If a guest needs more points, a minimum purchase, or another condition before using the offer, do not count that guest as eligible until those conditions are met.
For this comparison, use a member-based rate: the number of eligible members who redeemed at least one reward within 30 days of their first eligibility date, divided by the number of members eligible for that full 30-day observation window. Count each member once in both the numerator and denominator, even if they redeem more than once.
Start with records that identify the member, enrollment date, eligibility date, reward terms and redemption date. A loyalty report, transaction record or export may hold some of those fields. Ask the vendor where each field comes from and whether a transaction can be tied back to a member and reward. Your restaurant POS system may provide transaction details, but confirm how those details reach the loyalty report.
Separate what the records show from assumptions you make. For example, a redemption timestamp is observed data; an assumption that every eligible guest saw an offer is not. Record any missing dates or unmatched transactions before treating the rate as a basis for a decision.
Use a worksheet to compare two outcomes
The following figures are hypothetical. Assume each cohort joined in the named month, the same reward rules applied, and all listed members completed 30 days after first eligibility. Assume each redemption was valid and matched to one member. The figures measure members, not rewards issued or sales.
| Measure | January cohort | February: outcome A | February: outcome B |
|---|---|---|---|
| Eligible members with a complete window | 40 | 60 | 60 |
| Members redeeming within 30 days | 32 | 30 | 45 |
| Member redemption rate | 80% | 50% | 75% |
| Difference from January | Reference | 30 percentage points lower | 5 percentage points lower |
In outcome A, February’s rate is 30 divided by 60, or 50%. That is a large enough gap to investigate before extending the same offer without changes. Check whether February members could see and understand the reward, whether staff could locate it at checkout, and whether the terms differed in practice.
In outcome B, 45 divided by 60 is 75%. The five-point gap is smaller, so the owner might keep the offer as it is while checking another complete cohort. Neither result proves what caused the difference. A change in guest mix, promotion timing or reward terms could matter, and these small fictional groups should not be treated as a forecast.
Set the decision threshold before looking at the report. For example, an owner could decide that a gap of at least 20 percentage points triggers an operational review, not an automatic cancellation. Under that rule, outcome A triggers the review and outcome B does not. The threshold is a management choice, not a statistical guarantee.
If the offers were different, label that difference alongside the rates. Comparing a free-item reward with a discount can still be useful, but the result cannot isolate whether enrollment timing or the offer itself drove the gap. The loyalty program guide can help frame the reward rules before you make that comparison.
Account for outstanding and expired rewards
A nonredeeming member is not necessarily a lost member. At the end of the observation window, a reward may still be available, may have expired, or may need investigation because its status cannot be matched. Keep those outcomes separate from the redeemed count.
In hypothetical outcome A, 30 February members did not redeem within 30 days. Suppose 18 still had an outstanding, redeemable reward on day 30 and 12 had a reward that expired unused by day 30. The two groups add up to 30. The 18 outstanding rewards might be redeemed later, but a later redemption does not change the 30-day rate. The 12 expired rewards need a different follow-up: check whether the expiration was clear and whether the terms left enough opportunity to use them.
The shift manager should flag a checkout problem if staff repeatedly cannot find a guest’s eligible reward. The person responsible for loyalty reporting should then reconcile the affected member and transaction records before issuing a cohort comparison. If records cannot be resolved, report the affected count as unknown rather than quietly placing it among expired rewards or treating it as a confirmed redemption.
Also decide how to handle reversals and duplicate transactions. A voided redemption should not automatically remain a successful use, while two records for the same member should not count as two members. Document the rule and apply it to every cohort. If a vendor cannot show those exceptions, ask for an export that lets the reporting owner investigate them.
Set up a fair reporting routine
Run the report on a schedule, but do not force a comparison before the data is ready. A January enrollee who first qualifies late in the month may complete the 30-day window after a February enrollee who qualifies quickly. The eligibility date, not the cohort month alone, determines whether each member has had the full opportunity.
For a clean comparison, include only members whose first eligibility date is at least 30 days before the report cutoff. Show the number excluded because their windows are still open. When you revisit the report later, those members can enter the completed-window measure. Keep the cutoff date and rules with each report so an owner can explain why a prior result changed.
Check that the reward and channel rules are comparable. If eligible orders earn points only under certain conditions, confirm those conditions before combining records. Ask whether the report can distinguish an eligible reward from a points balance and whether it records a redemption made at each checkout channel you intend to include. Do not assume every order or channel earns points.
Access to a guest’s contact details is also separate from permission to send marketing. If the next action is a reminder campaign, check the available send controls and your permission records before contacting anyone. For the economics of an offer, use a separate cost calculation; a redemption rate alone does not tell you the cost per reward or the sales attributable to the program. The reward cost guide is a useful companion.
Ask for a pass-or-fail demonstration
Before choosing a reporting workflow, give the vendor a small fictional dataset with two enrollment months, different eligibility dates, a duplicate redemption, an expired reward and an outstanding reward. Ask the vendor to show the steps using a simulation or other non-live setup. Practice must not create a real payment, send a guest message or dispatch a courier.
- Filter members into the correct enrollment cohorts and show the report cutoff.
- Exclude members who have not completed 30 days after first eligibility, and show how many were excluded.
- Count each eligible member once and identify whether that member redeemed within the window.
- Separate outstanding, expired and unresolved nonredemptions. Show how a duplicate or void is handled.
- Export the underlying fields and totals so the restaurant’s reporting owner can reconcile them.
Pass: the vendor can produce and explain the correct cohort counts and rates from the fictional records, including the exceptions, without live actions. Fail: the output mixes incomplete windows into the denominator, counts rewards instead of distinct members without making that distinction, or leaves exceptions untraceable. If a manual export is part of the workflow, test that step too and decide whether the work is manageable for your team.
After the simulation passes, a live service pilot is a separate decision. Agree on who reviews real records, how discrepancies are logged and when the report will be checked. Do not treat a simulated result as proof that every live transaction will be classified correctly.
Bring your cohort definition, sample fields and demonstration checklist to a vendor conversation. Compare those requirements with Fleksa’s restaurant marketing tools, and confirm current support for the reporting and export workflow before choosing it.
Frequently asked questions
Should I compare everyone who joined in each month?
Not for this rate. Include members who became eligible and completed the full observation window. Track other enrollees separately so you can see how many have not yet qualified or are still within their window.
Does an expired reward count as a redemption?
No. An unused expired reward is a nonredemption. Show it separately from a reward that remains available and from a status you cannot resolve.
Can a high redemption rate prove the program increased sales?
No. It shows reward use among the defined eligible members. Evaluating sales effects requires a separate comparison and careful attention to other changes that may influence purchases.
