How to choose rewards and set point values

Progress (stamp) cards are simple and instantly understood: nine coffees, the tenth is free. They work where purchases are frequent and roughly the same size - cafés, barbershops, car washes.

Point collection gives a point per manat. It is fairer when basket sizes vary: a 5 ₼ purchase should not earn the same reward as a 50 ₼ one. This suits shops and grocers.

Tiers - Bronze, Silver, Gold - work on basket size, because a customer one purchase from the next level tends to spend a little more. They only make sense once your customer base is large enough; four tiers across thirty customers is theatre.

Here is the mistake almost everyone makes: costing the reward at its selling price. What it actually costs you is its cost price.

An example, with illustrative figures: a coffee sells for 5 ₼ and costs 1.5 ₼ to make. If the tenth coffee is free:

• The customer spends 9 × 5 = 45 ₼ to reach the reward
• It costs you the cost price of one coffee - 1.5 ₼
• Your real discount rate: 1.5 / 45 = 3.3%

The customer experiences it as 10%, because one coffee in ten is free. Perceived value 10%, real cost 3.3% - and a good loyalty programme lives in exactly that gap.

The same arithmetic for points: if 1 ₼ earns 1 point and 100 points buys a 10 ₼ reward, you are giving away 10% of revenue. If your margin is 30%, that is a third of your margin. This number should be known before the decision, not after.

Set the threshold too low and the programme becomes a discount: the customer is rewarded for a purchase they were making anyway and nothing about their behaviour changes. Set it too high and it becomes invisible - nobody finishes, and the card is forgotten.

A working rule: set the threshold 1.5 to 2 times the customer's natural repeat rate. If your customer comes four times a month and you want the reward earned within a month, six to eight visits is right. Three is too easy; twenty is out of reach.

You can change the threshold later - one of the real advantages of a digital card - but never retroactively devalue points already earned. That resets the programme's credibility in a single move.

The best reward is one with a high margin and an even higher perceived value.

Things that work: drinks and desserts (low cost price, visible price tag), an added service (a wash at a salon, an interior clean at a car wash), skipping the queue or early access, a small branded item.

Things that do not: cash discounts - they come straight off margin and generate no feeling at all; something already cheap, which does not register as a gift; and a reward that requires a further purchase to claim ("on orders over 10 ₼"), which the customer reads as a condition rather than a reward.

One rule: the reward should be the thing the customer already buys, one step up. Giving the usual-coffee customer a speciality coffee is both cheaper than it looks and more memorable than a discount.

This section usually goes unwritten because it does not sell. It is still better than deciding blind.

A programme runs at a loss in three situations.

The reward goes to customers who were returning anyway. If your regulars already come three times a week and you now reward them, you are paying for an existing habit rather than a change in behaviour. That is pure cost.

The threshold is too low. The gap between 3.3% and 10% is entirely a function of the threshold. Make every fifth coffee free and the real rate rises to 1.5 / 20 = 7.5%.

The reward is a low-margin item. Giving away something whose cost price is 70% of its selling price is very nearly a cash discount.

The simplest check: divide the total monthly cost price of rewards issued by the monthly revenue from programme members. That is your real discount rate. Knowing that number is the difference between running the programme and hoping.

The standing reward rule is the skeleton of a programme; limited-time offers are what make it move.

The form that works most reliably is a double-points window: points earned on chosen days or hours count twice. Its advantage is that the cost rises only during that window, and you can aim it at whatever you want to fix - a quiet Wednesday, the morning hours, the dead week after a holiday.

Two rules. Keep them short - a special offer running longer than a week becomes the new normal and stops working. And measure the result: compare sales during the window with the same days before it, or you have simply given a discount to people who were coming anyway.

A café. Average order 8 ₼; coffee sells at 5 ₼, costs 1.5 ₼. Progress card, 9 + 1. The customer visits six times a month and reaches the reward in about six weeks. Real discount 3.3%, perceived 10%. With 200 active customers earning roughly a quarter of a reward each month: 50 rewards × 1.5 ₼ = 75 ₼ a month.

A shop. Average basket 25 ₼, margin 30%. Points: 1 ₼ = 1 point, 250 points buys a 15 ₼ item that costs you 10.5 ₼. The customer spends 250 ₼ to earn it, so the real rate is 10.5 / 250 = 4.2% - about 14% of margin. That is acceptable. Drop the threshold to 150 points and the rate rises to 7%, taking a quarter of margin with it.

In both cases the arithmetic is simple. Not doing it is what is expensive.