A loyalty programme can look busy without changing a single buying decision.
Customers enrol. Points are issued. Rewards are redeemed. The dashboard moves.
But a retailer recently asked the question that matters:
"Did you actually see repeat customers increase, or is it more of a nice to have?"
That question separates programme activity from commercial impact.
A useful loyalty programme should change something the retailer values. It might bring a customer back sooner, increase the chance of a second purchase, protect a relationship during a quiet period or make a useful service easier to access.
If the same customers would have bought the same products at the same time without the reward, the programme may be recording loyalty rather than creating it.
The answer is not to abandon loyalty. It is to define what success means before adding more points, tiers or discounts.
Four numbers that should not be confused
Enrolment, participation, redemption and repeat purchase belong to the same journey, but they do not prove the same thing.
1. Enrolment
This is the number of customers who joined.
Enrolment tells you whether the invitation was visible and acceptable. It does not tell you whether members later noticed, valued or used the programme.
2. Active participation
This is the number of enrolled customers who earned, checked or used a benefit during a defined period.
The distinction matters. [Deloitte's 2025 Consumer Loyalty Program Survey](https://www.deloitte.com/us/en/insights/industry/retail-distribution/reshaping-customer-loyalty-programs.html) found that respondents were enrolled in eight programmes on average but active in five. The survey covered 5,564 US adults and also found that 40% sometimes forgot to redeem.
Those figures should not be treated as a promise for every small retailer. Deloitte notes that several behavioural questions concerned each respondent's preferred programme, so the results should not be generalised to every membership.
The practical lesson is narrower: a member list can be much larger than the group receiving regular value.
3. Redemption
Redemption shows that a customer used a reward.
It may indicate relevance and low friction. It may also show that the retailer discounted a purchase the customer had already decided to make.
A high redemption rate can be useful, expensive or both.
4. Repeat purchase
This is the proportion of eligible customers who bought again within the chosen period.
It moves closer to the real goal, but one question remains:
Would those customers have returned without the programme?
That is the difference between repeat purchasing and incremental repeat purchasing.
Choose one behaviour before choosing more rewards
A programme becomes difficult to evaluate when it tries to increase visits, basket value, referrals, category discovery and customer data collection at the same time.
Choose one primary behaviour for the test.
For example:
- complete a second purchase within 45 days
- return within the normal replenishment cycle
- make one additional full-price visit during a quiet month
- try a relevant category after the first purchase
- remain active for six months
- use a service that reduces checkout or collection friction
Write the behaviour as a measurable sentence:
`Among customers making a first purchase this month, increase the percentage making a second purchase within 45 days.`
That is clearer than "improve loyalty".
It gives the retailer a customer group, an event and a time window. It also makes inappropriate reward ideas easier to reject.
Establish a baseline before changing the programme
The retailer needs to know what normally happens without the new intervention.
A simple baseline might answer:
- How many first-time customers normally return?
- How long does the second purchase usually take?
- Which categories have naturally short or long repurchase cycles?
- What was the average basket and gross margin?
- Which other promotions were running?
- Did seasonality, a new location or unusual stock availability affect the period?
Suppose 200 eligible first-time customers bought during a typical month. If 40 made a second purchase within 45 days, the baseline repeat rate was 20%.
After introducing a reward, 54 of 200 comparable customers returned. The observed repeat rate was 27%.
The visible difference is seven percentage points.
That is useful evidence, but it is not automatically proof that the reward caused all 14 additional visits. Customer mix, season, product availability and other campaigns may differ.
The retailer should preserve that uncertainty rather than presenting a precise causal claim the test cannot support.
Use the fairest comparison the business can manage
Large organisations may run controlled experiments. A small retailer can still make a better decision without pretending to have a laboratory.
Option 1: Random comparison
Where lawful, fair and practical, divide eligible customers into comparable groups. One receives the tested offer and the other follows the normal journey.
This is the strongest simple design because both groups experience the same trading period.
It still needs safeguards:
- do not disadvantage customers unfairly
- avoid confusing staff or customers
- keep the offer rules consistent
- check privacy and direct-marketing requirements
- decide in advance which result counts as success
Option 2: Matched customer cohorts
Compare customers with similar starting conditions:
- first purchase period
- product category
- location
- basket range
- new or existing customer status
- normal repurchase cycle
Matching cannot remove every difference, but it is better than comparing all members with all non-members.
People who choose to join may already be more engaged.
Option 3: Before-and-after periods
Compare the same behaviour before and after the change.
This is easier, but it is vulnerable to seasonality, promotions, price changes, weather, stock availability and changes in customer mix.
Use comparable periods where possible and record what else changed.
The goal is not perfect science. It is a decision that is more reliable than looking at points issued.
Calculate the economic result, not only the visit count
A programme can increase visits and still weaken margin.
Track at least:
`Repeat purchase rate = customers who repurchased / eligible customers`
`Observed lift = test repeat rate - comparison repeat rate`
`Incremental gross profit = incremental revenue x gross-margin rate - reward cost - programme operating cost`
The operating cost may include:
- subscription or integration charges
- staff setup and support time
- printing or replacement cards
- customer-service exceptions
- fraud or duplicated rewards
- discounts applied to demand that already existed
Imagine the programme appears to create 14 additional visits. If each visit produces GBP30 of revenue at a 40% gross margin, the gross profit before programme cost is GBP168.
If rewards cost GBP90 and the monthly operating cost is GBP100, the programme has not yet paid for itself in that period.
The conclusion may be to improve the design, target a different behaviour or measure over the full customer cycle. It should not be hidden behind a growing redemption count.
Paper and digital programmes face the same test
Paper can be a sensible starting point for a small shop.
It is visible, inexpensive to understand and requires little customer setup. It can also be lost, forgotten, duplicated or difficult to connect to customer-level behaviour.
A digital programme can connect participation across visits and make measurement easier. It may also create subscription cost, integration work, staff training and customer-data responsibilities.
Competitor systems demonstrate how configurable the mechanics can become. [Square's current loyalty documentation](https://squareup.com/help/us/en/article/3952-create-a-loyalty-program-with-square) describes earning by spend, item or visit, along with different reward types and expiry settings.
Those controls are useful only when they serve a defined behaviour.
Choosing "one point per visit" because the setting exists is not a strategy.
For either format, test:
1. Can staff explain the programme in one sentence?
2. Can customers understand how they earn and redeem?
3. Does redemption work during a normal busy checkout?
4. Can exceptions be resolved without arbitrary manual points?
5. Can the retailer measure the target behaviour?
6. Can reward cost and margin be explained?
7. Can the programme be paused or changed without surprising customers?
Reward friction can hide behind enrolment
A programme may offer genuine value but make that value difficult to use.
Customers may forget the card, miss an expiry message, lose track of points or discover at checkout that the expected reward does not apply.
This is not merely a customer-experience issue. It changes measurement.
If customers cannot see or use a benefit reliably, low redemption does not prove the reward lacks value. It may prove that the journey is unclear.
Track:
- enrolment completion
- active-member rate
- time from earning to redemption
- unused and expired value
- failed or abandoned redemption
- staff overrides
- customer questions and complaints
Fix obvious friction before judging the reward concept.
Customer data creates a value exchange
A digital loyalty programme may connect identity, purchases, preferences, visits and communications.
That can help the retailer provide continuity and measure behaviour. It also creates responsibilities that should be designed into the programme rather than added later.
[ICO guidance for wholesale and retail businesses](https://ico.org.uk/for-organisations/data-protection-fee/paying-a-data-protection-fee-what-do-you-need-to-know/wholesale-and-retail-sector/) makes a useful distinction: holding customer details is not the only question; what the business does with those details matters.
Before collecting another field, ask:
- What customer benefit or operational purpose requires it?
- How will the purpose be explained?
- Who can access the data?
- How long will it be retained?
- Which communications has the customer agreed to receive?
- Can the business honour access, correction or deletion rights where required?
- Does a third party receive or process the information?
Use the minimum data needed for the programme and the measurement design.
Personalisation should not become an excuse to collect information without a clear purpose.
A practical 30-minute loyalty measurement setup
Before the next campaign, write down:
Target behaviour
One action, one eligible group and one measurement window.
Baseline
The normal repeat rate, purchase timing, revenue and gross margin for a comparable group or period.
Programme treatment
The exact earning rule, benefit, cost, expiry and customer communication.
Comparison
The fairest available untreated group, matched cohort or prior period.
Friction measures
Failed enrolment, forgotten rewards, failed redemption, staff overrides and customer complaints.
Commercial measures
Incremental repeat purchases, incremental revenue, gross profit, reward cost and operating cost.
Decision rule
Define what result means:
- keep the programme
- change the reward
- simplify the journey
- test for longer
- stop the intervention
Set the decision rule before seeing the result. Otherwise it is too easy to move the goalposts when activity looks encouraging.
Where EzyCarto fits
EzyCarto is designed as a unified retail operations platform rather than a standalone points counter.
Its current public product scope connects loyalty and rewards with checkout, customer relationships and analytics. That connected context can make it easier to observe the journey from participation to purchase and to keep the customer experience consistent.
The platform does not decide whether a programme succeeded.
The retailer still needs to define:
- the behaviour worth changing
- the eligible customer group
- the comparison
- the reward economics
- the data-governance boundary
- the decision after the test
Explore [EzyCarto Loyalty and Rewards](https://ezycarto.com/loyalty) when you are ready to connect the programme to the wider retail journey. Use [EzyCarto CRM](https://ezycarto.com/crm) to understand how customer context fits that journey.
If checkout friction is already causing customers to leave silently, fix that foundation before adding another incentive. Our guide to [the customers lost at checkout](https://ezycarto.com/blog/the-customers-you-lose-at-checkout-never-complain-first) explains the warning signs.
Loyalty programme decision checklist
Before launch or renewal, confirm:
- [ ] We have named one primary behaviour to change.
- [ ] We know the normal baseline.
- [ ] We have a fair comparison method.
- [ ] We distinguish enrolment, activity, redemption and repeat purchase.
- [ ] We calculate incremental revenue and gross-margin impact.
- [ ] We include reward and operating costs.
- [ ] Staff can explain and operate the programme during a busy period.
- [ ] Customers can understand and use their rewards.
- [ ] We collect only the customer data needed for a clear purpose.
- [ ] We have a decision rule for keeping, changing or stopping the programme.
A loyalty programme earns its place when it creates useful, profitable behaviour for the customer and the retailer.
Activity is evidence that the programme exists.
Changed behaviour is evidence that it works.