August 2026 · Retention marketing · 8 min read

Loyalty Programmes: Why Most of Them Buy Nothing

I worked on customer loyalty inside a large retail operation and I have built partner incentive structures in B2B. The failure mode is identical in both.

Rewarding the right behaviour is harder than rewarding any behaviour

Almost every loyalty programme I have seen up close does the same expensive thing: it hands a discount to the customers who were going to buy anyway.

It looks like it works. Members spend more than non-members, the dashboard is green, everyone is pleased. But that comparison is upside down. Your best customers joined the programme because they were your best customers. The programme did not create the behaviour. It attached itself to it and took the credit.

The question the dashboard will not answer

The only question worth asking is: what did this customer do that they would not have done otherwise?

That is uncomfortable, because it usually shrinks the reported win considerably. But you cannot design a good programme while measuring the wrong thing, and "members spend more" is the wrong thing.

The honest measures:

• Purchase frequency before and after joining, for the same customer, not member versus non-member.
• A held-out group. Even a small one. Some customers who do not get the offer, so you can see what happens without it.
• Margin, not revenue. A programme that lifts revenue and destroys margin is a slow way of buying your own customers back.
• Second purchase rate. In most businesses the gap between first and second purchase is the single biggest leak, and almost no programme is designed around it.

If the programme cannot show a behaviour change, it is not loyalty. It is a discount with a membership card.

Three design mistakes that repeat everywhere

Rewarding the purchase instead of the next one

Points on what someone already bought is a rebate. It has no forward pull. The reward has to sit slightly ahead of them, close enough to be worth reaching for.

Tiers that mean nothing

Bronze, Silver, Gold with marginally better discounts. Nobody organises their behaviour around 3% instead of 2%. Tiers work when the higher tier gives something that is not money: early access, a named contact, priority service, a decision made faster. Those cost little and feel like status, which is what tiers are actually for.

Designing for the top 5% only

Your best customers were never the problem. The value is in the second tier: people who buy occasionally and could buy regularly. They are also the group most likely to disappear without you noticing.

Baku, Mostar, Kathmandu: loyalty looks different in every market

What I would build instead

• Start from a single behaviour you want more of. One. Second purchase within 60 days, or a return visit in a slow month. A programme aimed at everything moves nothing.
• Make the value legible in one sentence. If a customer cannot explain what they get without reading terms, the programme will not spread by word of mouth, and word of mouth is most of the payoff.
• Give something in the first week. Not after eight purchases. The early reward is what buys the second visit, and the second visit is what you are actually paying for.
• Use non-monetary rewards wherever they will do. Access, speed, recognition, being remembered. These carry margin and are much harder for a competitor to copy than a discount.
• Run it out of the CRM, not a separate silo. If loyalty data does not sit next to purchase and service history, you cannot personalise anything and you cannot see churn coming.

The B2B version nobody calls loyalty

In B2B there is rarely a points scheme, but the same mechanics decide whether a partner keeps sending you business.

Building partner networks taught me that partners are loyal to the relationship that makes their life easier, not to the one with the best commission. The partners who stayed longest were not the ones with the highest margin. They were the ones who got answers fastest, who knew who to call, and who were never surprised by something they should have been told.

So the B2B loyalty programme is unglamorous:

• A named person who actually replies
• Response times that do not vary by how large the partner is
• Being told about changes before they affect them
• Recognition that costs nothing: being asked first, being mentioned, being consulted
• Tiering by reliability rather than volume, so a small consistent partner is not treated worse than a large erratic one

Where to start if you already have one

Do not redesign it. Measure it properly first, with the same-customer before and after comparison and a held-out group. That single piece of analysis usually reveals two things: a segment where the programme genuinely changes behaviour, and a much larger one where you are paying for nothing.

Then shrink it to the part that works and spend the difference on the second-purchase gap. That is normally the biggest available win and it is sitting there unclaimed in almost every business I have looked at.

Updated: August 2026