A customer loyalty program gives customers a clear benefit for a defined action, such as repeat purchases or referrals. It works only when customers understand the rules and the expected reward cost fits the business. Enrollment alone does not prove that the program created more profitable repeat buying.
Start with the customer behavior you want to encourage. Then compare the cost and complexity of a reward with the value of a genuine incremental visit or order.
Choose a reward customers can understand
| Program structure | How it works | Main design question |
|---|---|---|
| Points | Eligible purchases earn points that can later be redeemed | How many points are earned, and what is one point worth? |
| Stamp or visit reward | Customers earn progress toward a set benefit | Which purchases or visits count, and how is progress tracked? |
| Tiered program | Benefits change after customers meet clear thresholds | Are the tiers achievable and worth the operating cost? |
| Referral reward | A customer receives a benefit after a qualified referral | What event counts as a successful referral, and how are abuse and consent handled? |
| Member price or benefit | Eligible customers receive a defined price or service benefit | How will customers qualify, and will the price still cover costs? |
Do not add points, tiers, expiry dates, or exclusions simply because another program uses them. Every rule creates customer questions and staff work.
Calculate the cost before launch
List the reward’s product cost, discount amount, delivery or fulfilment cost, payment fees, tax treatment, and the system or staff work needed to administer it. Estimate how many rewards could be earned and redeemed under conservative, expected, and high-participation scenarios.
Reward cost per redemption = product or discount cost + fulfilment + payment and other variable costs.
Required eligible sales to fund a reward = reward cost Ă· contribution rate on those sales.
Suppose an offer costs KES 300 to provide and the related eligible sales contribute 30% after variable costs. Those sales would need to generate KES 1,000 in contribution-bearing revenue to cover the reward cost, before fixed overhead. This is an illustration of the calculation, not a forecast or universal target.
If rewards expire, state when and how clearly before a customer joins. Do not depend on unused points to make the program profitable. Make it easy for customers to check their balance and understand what can be redeemed.
Define fair, complete terms
Write the rules before enrollment. Explain:
- Who can participate and how to join.
- Which purchases or referrals qualify.
- How progress is earned and when it is recorded.
- What rewards are available and what each requires.
- Whether a reward can be combined with another offer.
- Expiry, exclusions, return, cancellation, and correction rules.
- How a customer can ask a question or leave the program.
- What personal data is collected and how it is used.
Apply rules consistently and provide a simple process to correct an error. If a returned purchase created reward points, explain how the points are adjusted and give the customer a record of the change.
Test whether the program changes behavior
Before launch, define the outcome you expect: higher repeat-purchase rate, shorter time to a second visit, higher customer retention, or increased contribution from a clearly defined group.
Compare enrolled customers with a suitable baseline or comparison group. Enrolled customers may already have been more loyal than other customers. A simple before-and-after comparison can overstate impact if seasonality, pricing, or another campaign changed at the same time.
Track:
- Enrollment and active participation.
- Eligible purchase frequency and value.
- Redemption rate and reward cost.
- Contribution after discounts and reward cost.
- Customer retention over a defined period.
- Returns, complaints, opt-outs, and rule corrections.
- Staff time and program administration cost.
Define each metric and cohort consistently. A high redemption rate can show that customers value the reward, but it can also mean the reward is too generous. A low redemption rate can mean customers do not understand the rules or the threshold is unrealistic.
Example: calculate a points reward
Suppose a program gives 1 point for every KES 100 spent and a customer can redeem 100 points for a product that costs the business KES 250 to provide. The apparent reward is not free: include the product cost, packing, delivery, payment charges, and any tax treatment in the estimate. Also define whether refunded purchases remove the points and how customers can see the adjustment.
Then estimate the purchase contribution required to fund the reward. If the contribution rate is 25%, KES 250 in reward cost requires KES 1,000 in contribution-bearing sales before fixed overhead. This arithmetic is only a planning example. A real calculation should use the business’s actual costs and rules.
Compare the behavior of members who redeem with similar customers who do not. If many redemptions come from purchases that would have happened anyway, the program may be rewarding existing behavior without changing retention. Consider a smaller benefit, a service improvement, or no program at all.
Avoid incentives that damage trust
Do not hide restrictions in fine print, change reward value without clear notice, or make customers spend more than the published threshold to redeem. Explain expiry and exclusions at enrollment. Keep communication relevant and respect marketing preferences.
Set a frequency limit for loyalty messages. Transactional information about a balance or redemption should be clear and separate from promotional content where practical. Use customer data only for defined purposes and protect access to program records.
Compare a program with a baseline
Before launch, record how often a defined customer group buys, the average contribution per order, and the share that returns within a chosen period. Then compare program participants with a reasonable non-participant group or a staged rollout. Participants may already be more engaged, so a simple comparison can overstate the program’s effect.
Where practical, use a holdout group that continues to receive the normal service without the new reward. Keep the treatment fair and clear; do not deny an existing entitlement. If a holdout is not feasible, compare similar cohorts and document differences in season, offer exposure, and customer type.
A loyalty program may increase repeat visits while lowering contribution through rewards, discounts, or extra administration. Calculate net contribution after reward costs, returns, payment fees, and staff time. Review customer complaints and opt-outs alongside sales measures. A program that produces more transactions but creates confusion or poor service may need to change.
Start small and review the full economics
Pilot one straightforward reward with a limited, clearly eligible group. Check how staff explain the rules, how customers redeem, whether records match, and whether the reward fits the margin. Collect customer questions and adjust unclear wording before widening the program.
Review the economics at regular intervals. Stop or redesign a reward if it subsidizes purchases that would have happened anyway, creates unprofitable discount stacking, encourages returns abuse, or costs more to administer than the benefit it produces.
Common loyalty program questions
Does enrollment prove the program works?
No. Enrollment shows that people joined. Compare repeat purchases and contribution with a suitable baseline and include reward and administration costs.
Should every customer receive the same reward?
Not always, but eligibility and value should be clear and applied consistently. Avoid hidden conditions or personal-data targeting that customers would not reasonably expect.
Should rewards expire?
That depends on the economics and customer experience. If they do, state the date or rule before enrollment, send clear reminders where appropriate, and make the balance easy to check.
For channel choice and campaign measurement, read local marketing planning. For consent and opt-out considerations in direct campaigns, see promotional SMS guidance.