Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Customer Experience, Retention & Loyalty › CX Foundations

Designing a Loyalty Programme

How to design a loyalty programme: points, tiers, paid membership and community models, what to reward, the economics of liability and breakage, and launch.
Guideintermediate10 min readUpdated

What a loyalty programme is for

A programme is a tool, and the first design question is which job it is for. Four jobs are common, and they lead to different designs.

The first is changing purchase behaviour: getting customers to buy more often, to consolidate spending that is currently split across competitors, to try a new category or to raise their AOV. The second is recognition: making the best customers feel noticed, which matters most in categories where service and status are part of the product. The third is data: a reason for customers to identify themselves at every purchase, which joins up transactions that would otherwise be anonymous, especially in physical retail. The fourth is switching cost: accumulated points or status that a customer would lose by leaving. For retailers with stores the data job is often the main one, and the retail industry marketing guide covers the consent rules and store measurement that decide how that data can be used.

The evidence is a useful corrective to enthusiasm. Work associated with the Ehrenberg-Bass Institute on how brands grow has argued for years that loyalty to a brand is largely a by-product of its size and of how easy it is to buy, and that programmes mostly attract the brand's existing heavy buyers. The page on the double jeopardy pattern shows how to check a brand's own loyalty figures against its size before setting a programme's targets. The practical lesson is not that programmes never work but that the design must target incremental behaviour, and that the data job is often the strongest case. The retention strategy guide covers retention outside a programme; a programme should sit on top of an experience that already gives customers a reason to stay, not substitute for one.

Programme models compared

The four main models differ in what customers earn, how quickly they feel it and what the business commits to.

ModelHow it worksBest forStrengthsRisks
PointsCustomers earn points per purchase or action and redeem them for rewardsFrequent, lower-value purchases; broad customer baseSimple to understand; flexible rewards; generates dataFeels transactional; easy to copy; liability builds up
TiersStatus levels unlocked by spending or activity, each with better benefitsCategories with a clear top group of customers; service businessesRecognition; aspiration to the next tier; retains top customersCustomers resent losing status; benefits must be real
Paid membershipCustomers pay a fee for ongoing benefits such as free delivery or member pricesFrequent buyers who value convenienceFee commits the customer; benefits are immediate; funds the programmeMust deliver clear value from the first purchase; churn at renewal
CommunityMembers get access, content, events or a voice in the brand rather than discountsBrands with enthusiasts; hobby and lifestyle categoriesHard to copy; builds advocacy; low reward costNeeds sustained effort; does not suit low-interest categories

Hybrids are common: points with tiers on top, or a paid tier above a free points scheme. The temptation is to build the hybrid first. Starting with one model, learning how members respond, and adding layers later is usually cheaper and clearer for customers. A paid model deserves particular care because the customer's judgement is immediate: if the benefits do not obviously exceed the fee within the first few orders, renewal will expose it.

What to reward and what not to

Reward behaviour the business wants more of and would not otherwise get. That rule rules out a surprising amount of what programmes typically reward.

Worth rewarding:

  • Frequency where customers have choice. Rewards that make the next purchase with you rather than a competitor more attractive, in categories where customers genuinely shop around.
  • Trying something new. A first purchase in a new category, or a first visit through a new channel, is a behaviour that often sticks once tried.
  • Actions with value beyond the purchase. Reviews, referrals, completed profiles and app downloads help the business in ways that are hard to buy elsewhere. Rewarding referrals well takes a scheme of its own, and the tutorial on designing a referral programme covers the reward, the sharing flow, fraud checks and disclosure.
  • Tenure and recognition. Small, unexpected gestures to long-standing customers build goodwill out of proportion to their cost.

Not worth rewarding, or worth rewarding cautiously:

  • Purchases that would happen anyway. Rewards on a customer's habitual weekly shop mostly return margin to people who were not going anywhere.
  • Spend alone at the top end. Pure spend thresholds can reward a few customers who buy at low margin, for example only during sales.
  • Behaviour that hurts the business. Points on items frequently returned, or bonuses that encourage bulk buying followed by long gaps.

The reward itself matters as much as the trigger. Experiences, access and service often feel more valuable to members than their cost to the business, while discounts feel exactly like what they are and train customers to wait for them. The repeat purchase and loyalty guide in the e-commerce root covers the tactics for online retailers that do not need a formal programme at all. Where customers reorder the same consumable on a steady cycle, the comparison of loyalty programmes and subscribe-and-save weighs rewarding each reorder after it happens against agreeing the next one in advance.

The economics: liability, breakage and margin

A programme's economics come down to three ideas. Liability is the value of rewards issued but not yet redeemed: an obligation the business owes members, which accountants usually require to be recognised and which grows as balances build. Breakage is the share of issued rewards that will never be redeemed, which reduces the real cost. Incremental margin is the extra profit from behaviour the programme actually changed, as opposed to behaviour it merely rewarded. Incremental margin is measured against a comparable group left out of the programme, before any of the programme's own costs are taken off.

The programme pays when incremental margin exceeds the cost of rewards actually redeemed plus the cost of running it: technology, staff, communication and the margin given back to customers who would have bought anyway. That last cost is the one most business cases omit. The break-even calculator helps frame the question: how much extra behaviour does the programme need to create to cover what it gives away? The LTV and CAC guide covers how changes in purchase frequency feed LTV, and therefore how much a programme can be worth.

Illustrative example — not a real result

A coffee chain tests the cost of its stamp card

Ember and Oak, a chain of twelve coffee shops, plans a digital points scheme: one point per £1 spent, with a £3 reward at 50 points, giving back 6% of spend. Members are expected to spend £400,000 a year. If every point were redeemed, the reward cost would be £24,000; the team assumes 30% breakage, so expected redemptions cost about £16,800, plus £9,000 a year for the app. A three-month pilot in four shops compares members with similar non-members and finds members' visits rise by about 8%, of which roughly half looks incremental once existing regulars are accounted for. At a 60% gross margin, 4% incremental sales on £400,000 is £9,600 of margin, well short of £25,800 in costs. The team redesigns: rewards move to off-peak hours and new menu items, cutting expected redemptions and steering them to where spare capacity exists.

The example shows the pattern most programmes follow on paper: the headline uplift looks healthy, but once behaviour that would have happened anyway is removed, the programme has to be redesigned to target what is genuinely incremental. Measuring that requires a comparison group, which is why a pilot beats a full launch. The loyalty programme pilot test designs that pilot with eligible customers invited at random or held out, counting everyone assigned whether or not they join. Report the programme's ROI on incremental margin only, and watch liability as a balance-sheet item rather than a marketing metric. A programme that depends on high breakage to break even is fragile, because customers who cannot redeem feel cheated, and that feeling is the opposite of what the programme was meant to create. When a scheme has been live for a year without moving retention, the diagnosis for a loyalty programme that is not increasing retention separates rewards that went to habitual purchases from the missing comparison group, the wrong model and a reading taken on a blended rate.

Launching and communicating the programme

A launch has three audiences: existing customers, staff and new customers. Each needs something different.

Existing customers need to understand what they get, how to join and what changes. If the programme replaces an older scheme, the transition needs special care: balances should be honoured, the conversion explained plainly and any loss of benefit stated rather than hidden. Staff need to understand the programme well enough to explain it in a sentence, and they need to believe it is good for customers, because a cashier or agent who apologises for the scheme kills it. New customers need the programme to be part of the first purchase, not a separate sign-up that interrupts it.

Communication continues after launch. Members should be able to see their balance, what it is worth and how close they are to the next reward without logging in to find out. The email automation hub covers the triggered messages that do this work: welcome, milestone, balance reminders and expiry warnings. The guide to email for loyalty and advocacy sets out how those messages recognise members without a discount and when they can fairly ask for a review or a referral.

Clarity in the rules is also a compliance question. UK consumer protection law, which the Competition and Markets Authority enforces, expects terms to be fair and transparent and marketing not to mislead; that applies to how rewards, expiry and changes to the scheme are described. This is general direction, not legal advice, and a programme with paid membership, expiry or significant changes to existing balances should be reviewed by someone qualified to advise on the relevant law.

Signs a programme is the wrong answer

A programme is often proposed as the answer to a retention problem that it cannot solve. These signs suggest a different answer is needed:

  • Customers leave because of the experience. If exits follow late deliveries, poor support or a product that disappoints, rewards add cost without removing the reason to leave. Fix the experience first.
  • The category is low-involvement and customers do not shop around. If customers already buy out of habit, a programme mostly returns margin to them.
  • Margin is thin. A programme gives back a share of every rewarded purchase; in thin-margin categories there may be nothing left to give.
  • Nobody will use the data. If the case rests on data but there is no plan or team to act on it, the data job will not be done.
  • Competitors all run near-identical schemes. Matching them may be necessary for parity, but it will not differentiate, and the business should be honest that it is paying to stand still.
  • Acquisition is the real constraint. If the business is small and growing, budget spent on rewarding existing customers may do more good spent on reaching new ones at a sustainable CAC.

The flywheel framework is the right lens for the decision: a programme belongs in the plan if it adds momentum to a loop that already turns, and not if it is meant to start one. The foundations hub places loyalty after journey, onboarding and retention for exactly this reason.

Before deciding who a programme should reward, the guide to segmenting the existing base with STP for retention shows how to sort customers into tiers with a distinct reason to stay.

To test whether each reward actually changes behaviour, the guide to the flywheel applied to customer loyalty treats points liability as friction and member advocacy as force.

To put a figure on that extra behaviour, the break-even calculator worked for loyalty programmes enters the scheme's full cost as spend and reads sales needed as extra member orders.

Common mistakes

  • Launching to fix churn. A programme rewards an experience; it cannot repair one.
  • Rewarding the habitual. Most cost goes to customers who would have bought anyway, with little change in behaviour.
  • Building the hybrid first. Points, tiers and paid membership at once confuse customers and make it impossible to tell what works.
  • Counting on breakage. A programme that only breaks even if customers fail to redeem is designed to disappoint them.
  • No comparison group. Without one, every purchase by a member looks like the programme's success.
  • Hiding the rules. Unclear expiry, quiet devaluations and buried conditions erode trust and create regulatory risk.
  • Treating it as a marketing project. Finance owns the liability, operations runs redemption and staff explain it; all of them need to be involved from the start.

Frequently asked questions

Do loyalty programmes actually make customers more loyal?

Less than their popularity suggests. Research from the Ehrenberg-Bass Institute and others has long pointed out that programmes tend to recruit customers who already buy the brand often, so much of the reward goes to behaviour that would have happened anyway. Programmes can still pay when they change a specific behaviour, such as consolidating spending or trying a new category, and when the data they collect is put to use.

What is breakage in a loyalty programme?

Breakage is the share of issued rewards that customers never redeem, because points expire, balances are too small to use, or members simply forget. It reduces the real cost of the programme, which is why finance teams watch it closely. Designing a programme that depends on high breakage is a mistake, though, because unredeemed rewards are also rewards that changed nobody's behaviour, and customers notice when rewards feel out of reach.

Should a small business run a loyalty programme?

Only if it can name the behaviour it wants to change and afford the rewards and administration without hurting margin. Many small businesses get more from a simple, generous gesture to repeat customers, such as a personal thank-you or early access, than from a points scheme. If a programme is chosen, a simple model with clear rules and an easy redemption is usually better than an elaborate one.

Terms used on this page

Tools to try

Frameworks

Where this page is used

Part of these learning paths

Also related

Sources

  1. Ehrenberg-Bass Institute for Marketing Science — Ehrenberg-Bass Institute
  2. Competition and Markets Authority — GOV.UK

Related