Start with the behaviour, not the points
Most loyalty programmes begin with a mechanic: points per rupee spent, a tier ladder, a card with stamps. The better starting point is a behaviour. What do you want customers to do more of that they would not otherwise do? Buy more often, try a second category, shop online instead of through a marketplace, refer a friend, leave a review?
Once the behaviour is clear, the mechanic follows. If you cannot name the behaviour, the programme will default to rewarding all spending equally, which mostly rewards your best customers for what they were already doing. That is a cost, not a strategy.
The main types of programme
| Type | How it works | Best suited to | Watch out for |
|---|---|---|---|
| Points | Customers earn points per purchase or action and redeem them for rewards | Frequent, moderate-value purchases | Points becoming a disguised discount; growing liability |
| Tiered | Status levels earned through spend or activity, with rising benefits | Businesses with a wide spread of customer value | Tiers that feel unattainable or benefits that feel trivial |
| Paid membership | Customers pay a fee for ongoing benefits such as free delivery or member pricing | High-frequency categories with clear benefits | Members who feel they are not getting value for the fee |
| Value-based or community | Rewards are access, content, events or causes rather than discounts | Brands with a strong point of view | Being vague about what members actually receive |
| Referral | Rewards customers for bringing in new ones | Products people naturally talk about | Rewarding low-quality or fraudulent referrals |
Why members value a programme
Fig. 01 · Overlap
Tap to explore
The three sources of programme value
CentreA programme members would miss
Economic value is the easiest to offer and the easiest to copy. Emotional value, such as being recognised, thanked or given status, is harder to copy and often cheaper to deliver. Practical value, such as free delivery, priority service, easy returns or early access, can change the experience itself. A test of a good programme: would members notice and mind if it disappeared?
The economics of rewards
Every reward has a cost, and every points programme creates a liability: points issued but not yet redeemed represent value you owe. Finance should be involved from the start, both to model costs and to agree how the liability is accounted for. The calculator below shows the effective discount a reward structure represents.
Calculator
Effective reward rate
How much of each purchase you are giving back, once redemption behaviour is considered. Defaults are an illustration only.
Headline reward rate
3%
What the programme appears to give back.
= reward / spend
Expected cost as a share of spend
1.8%
Allowing for unredeemed rewards; do not plan on breakage you cannot justify.
= reward / spend * redeem / 100
Reward cost as a share of gross margin
4%
The share of profit the programme consumes before any incremental behaviour.
= (reward / spend * redeem / 100) / (margin / 100)
Defaults are illustrations. Use your own numbers. Nothing you enter leaves this page.
The final number is the sobering one. A programme only pays for itself if the incremental behaviour it creates, such as extra purchases, larger baskets or reduced churn, generates more margin than the rewards consume. That has to be measured, not assumed.
Launching a programme
Fig. 02 · Timeline
Tap to explore
An illustrative launch sequence
Withdrawing or devaluing a programme angers exactly the customers you most want to keep, so it pays to get the structure right before launch. Pilot with a subset of customers or one region, and compare their behaviour with a similar group outside the programme.
Data, consent and communication
Loyalty programmes generate rich first-party data: who buys what, how often and where. That data is valuable for personalisation and segmentation, but only if members understand and agree to how it is used. Under laws such as India's DPDP Act, be clear about purpose, collect only what you need and make withdrawal easy. See first-party data.
Programme communication should be useful, not relentless. Members want to know their balance, what they can do with it and when something is about to expire. A monthly statement that tells them something worth knowing beats weekly reminders that they are members.
Designing tiers that motivate
Tiers work when the next level feels reachable and worth reaching. If the top tier requires spend that only a handful of customers will ever approach, most members see a ladder they cannot climb and stop paying attention. If every tier offers a slightly larger discount, the ladder offers nothing to aspire to.
Set thresholds from your own customer distribution, so each tier contains a meaningful group and the step to the next is a stretch rather than a leap. Make higher tiers different in kind, not just in degree: priority service, early access, invitations, a named contact. These benefits often cost less than larger discounts and are harder for competitors to match.
Tell members where they stand. Progress towards the next tier, shown in account pages and in occasional emails, is one of the few loyalty messages members consistently find useful. Decide in advance what happens when a member falls below a tier's threshold, and communicate it kindly; demotion handled badly can turn your best customers against the programme.
Coalition, marketplace and partner programmes
Some businesses join programmes run by others: card schemes, marketplace memberships, coalition points shared across brands. These can bring reach and convenience, but the relationship and the data often belong to the programme operator rather than to you. Before joining, ask what customer data you receive, whether members become your customers or remain the operator's, and what happens if you leave. A programme that grows someone else's first-party data at your expense is a media cost, not a loyalty strategy.
Measuring whether it works
The trap is comparing members with non-members. Members spend more, of course; your best customers joined first. That comparison proves nothing about the programme. Instead, compare behaviour before and after joining against a matched group, or run controlled tests of specific rewards. Measure incremental margin, not just incremental revenue.
Self-diagnostic
0/5Is your loyalty programme earning its cost?
Answer for a current or planned programme.
01Can you name the specific behaviours the programme is designed to change?
If yes: Check each reward rule ties back to one of them. If no: Define them before changing any mechanics.02Has finance modelled reward costs and agreed the liability treatment?
If yes: Review the model against actual redemption regularly. If no: Involve finance now. Unmodelled liability is a balance sheet surprise.03Do you measure incrementality against a matched group, not members versus non-members?
If yes: You know what the programme actually adds. If no: Set up a matched comparison or controlled tests.04Does the programme offer emotional or practical value, not just discounts?
If yes: It is harder for competitors to copy. If no: Add recognition, access or convenience benefits.05Would members notice and mind if the programme disappeared?
If yes: You have built something of value. If no: Ask members what they value; the programme may be a cost with little loyalty attached.
Loyalty is the top layer of retention marketing. It works best when the layers below it, product, service and relevant communication, are already strong.
Reviewing and evolving the programme
Treat the programme as a product with a roadmap. Review it at least annually against the behaviours it was designed to change, the margin it consumes and what members say about it. Add benefits that members use and value; retire those nobody redeems. When changes reduce value for members, announce them early, explain the reason plainly and, where possible, offer something in exchange. Loyalty is easily lost through a programme change handled carelessly.
A loyalty programme should reward the behaviour you want, not rent the loyalty you already had.
Key takeaways
- 01Design a loyalty programme around specific behaviours you want to change, not around a points mechanic.
- 02Programmes offering emotional and practical value are harder to copy than discount-led ones.
- 03Model reward costs and liability with finance before launch, and pilot before rolling out.
- 04Measure incrementality against a matched group; member versus non-member comparisons prove nothing.
- 05Loyalty strengthens good retention foundations but cannot replace them.
Frequently asked
- What is a customer loyalty programme?
- A loyalty programme is a structured scheme that rewards customers for behaviours a business wants to encourage, such as repeat purchases, higher spend, referrals or reviews. Rewards can be points, discounts, tiers with rising benefits, paid memberships with perks, or access and recognition.
- Do loyalty programmes actually work?
- They can, when designed around specific behaviours, funded sustainably and measured for incremental effect. Many programmes mainly reward customers who would have bought anyway, which makes them an expensive discount. Measure behaviour change against a matched group to know whether yours works.
- What is the best type of loyalty programme?
- It depends on your purchase frequency, margins and brand. Points suit frequent, moderate purchases; tiers suit businesses with a wide spread of customer value; paid memberships suit high-frequency categories with strong practical benefits; value-based programmes suit brands with a clear point of view.
- How much should a loyalty programme cost?
- There is no standard figure. Calculate the effective reward rate as a share of spend and of gross margin, allowing for realistic redemption, and compare it with the incremental margin the programme generates. If the extra behaviour does not cover the rewards, redesign the programme.
- What data should a loyalty programme collect?
- Collect what you need to run the programme and personalise it meaningfully, typically identity, contact details, purchases and stated preferences, with clear consent for each use. Explain purposes plainly and make withdrawal easy, in line with data protection laws such as India's DPDP Act.
Published by Fabulous.Media, a network of specialist marketing agencies. Updated 9 October 2026. Platform features change often; check current official documentation before acting on platform-specific detail.
