Reward loyalty. grow retention.

Retailers run this experiment often. The program launches, staff are briefed, enrolment looks good for a month, and then a busy Saturday arrives and nobody asks anybody about points again.

Loyalty built into the sales system takes the remembering out of it. The customer is looked up on the same screen the cashier is already using, the balance is visible there, and redeeming happens as part of taking payment rather than as a separate procedure with its own login.

The mistake worth avoiding is copying the rate a large chain uses. Their margin, their buying power and how often their customers visit are not yours. A rate that turns out to be too generous is painful to reduce, because the people who notice are exactly the customers you were trying to hold on to.

Automatic tiering is the practical part. Programs with manual tiers decay quietly, because promoting and demoting customers is nobody's actual job and the list ages until it describes last year.

Whether tiers help at all depends on the business. Where customers buy often and compare openly, a visible tier gives them a reason to put all their spending in one place. Where purchases are rare and large, tiers do very little and a straightforward earn rate works better.

Multi branch loyalty is where home built programs fall over. Each till keeps its own list, a merge job runs at night, and the customer discovers the gap at the worst possible moment, standing at a counter with a queue behind them.

A shared balance protects you in the other direction as well. The same points cannot be spent twice in two locations, which is the failure nobody mentions when describing a system that syncs after the fact.

Self service matters more than it sounds for a program's credibility. A balance a customer can check is a balance they believe, and a customer who believes the balance is one who eventually uses it.

There is a staff cost to it too. Every points question answered at a counter is a minute of a queue on the busiest day of the week, and it is a minute spent on something the customer could have looked up in the car park.

The honest caveat: a loyalty program is a retention tool and nothing more. If customers are leaving because of price, range or how they were treated, points delay the decision rather than change it, and you will have paid for the delay.

It also carries a real cost. Points are margin you have promised to give back, and it is worth modelling before launch rather than after the first heavy redemption week. Take last year's sales, apply the rate you are considering, and look at the number.

Where it does work, the effect is usually concentrated in a minority of customers who care enough to change where they shop. That group is worth the program on its own. Measure them rather than the whole base, or the average will hide what is actually happening.