Do loyalty points actually bring customers back?

The uncomfortable truth about most loyalty schemes is that they pay your best customers to do what they were already doing.

Ten per cent off for the regular who comes every fortnight is not retention. It is a ten per cent price cut on your most reliable revenue, dressed as a growth initiative.

The two things a scheme can actually do

1. Change behaviour at the margin. Someone who would have come every six weeks comes every four. That is real, it is measurable, and it is worth paying for.

2. Buy a recommendation. A referral is worth more than a repeat purchase, because it arrives with trust attached that no advertising can buy.

Everything else is a discount with a nicer name.

The k-factor, and why the honest number is deflating

The k-factor is the average number of new customers each existing one brings. Above 1, the loop compounds on its own. Below 1, it decays and needs refilling from marketing spend.

Almost every real small-business scheme sits well below 1. That is not a failure — a k-factor of 0.3 is a third of your new customers arriving free, which is excellent. It just is not the self-sustaining machine the word "viral" implies.

Every projection here is a labelled estimate, because it is arithmetic performed on your assumptions. Change the assumption and the projection changes. That is a model, not a forecast, and anyone presenting it as a forecast is counting on you not asking which assumption moved.

What a reward is allowed to cost

The same arithmetic that decides what you can afford to pay an affiliate: revenue, minus variable cost, minus the margin you are protecting. What remains is what acquisition can spend.

A reward above that ceiling is not growth. It is buying revenue at a loss, and it is particularly dangerous in a loyalty scheme because the cost is spread across thousands of small redemptions that never appear as a line item anybody reviews.

The consent rule, which is not optional

A referral invite goes to somebody who has not asked to hear from you. So: the recipient's consent is required, and a hard cap of five touches in seven days applies.

Two reasons, and the second is the one most vendors skip. The first is legal — in the UK and EU, an unsolicited marketing message to an individual subscriber is a PECR matter, and "my customer gave me their friend's email" is not consent from the friend.

The second is shared infrastructure. Mail here leaves on a sending pool whose reputation is shared with every other business on it. One enthusiastic referral blast generating complaints degrades deliverability for businesses that did nothing wrong. That is why the cap is in the engine rather than in the terms.

And it cannot fix the offer

A referral scheme multiplies what your customers already think of you. If the honest answer is "they are satisfied", no incentive structure turns that into advocacy — and the thing that does is usually sitting in the feedback you already have.

The question to ask

Before launching any scheme: would this customer have come back anyway?

If the answer is yes for most of your redemptions, you have built a discount. Point the money at the margin instead — the customers who were nearly lost, and the ones who would recommend you if asked.