How soon should customers earn their first reward?

In 2006 three researchers published a study of a café with several outlets on an American university campus. The café ran the card most of us know: buy ten coffees, get one free, one stamp a visit, with no expiry date stated. The researchers collected 949 completed cards covering about 10,000 purchases, each stamp secretly numbered so that the purchase could be traced, and looked at how the gap between visits changed as a card filled up.

It shrank. Early in a card, customers left about 3.3 days between coffees. By the last stamps the gap was about 0.7 of a day shorter, which the authors describe as "an average acceleration of 20%". A card took 24.6 days to finish, against the 29.4 days it would have taken if the customer had kept to the pace set between the first two stamps.

Two details matter as much as the headline.

The speed-up belonged to people who finished. The researchers also bought back 73 unfinished cards from customers. Those with at least three stamps were analysed, and they showed no acceleration at all. The bought-back cards had been in use for an average of 65 days, against about 25 for the finished ones, which the authors took to mean they would never have been completed. For a customer who is not going to reach the reward, the card changes nothing.

The effect ended at the reward. Among 110 customers who completed two cards, the first gaps on the second card were back to about three days, where the gaps at the end of their own first card had been about two. The pull comes from being close to something, and a customer who has just collected a free coffee is ten purchases away again.

This is one café on one campus, selling a cheap item that people buy several times a week. It is an example, not an average for all retail. The pattern itself has a long history in psychology, where it is called the goal gradient: effort rises as a goal gets nearer. The café is the clearest case of it measured at a counter.

The usual objection to a near reward is cost. A card of ten with one free gives back one item in eleven, about 9% of what the customer is served. A card of five gives back one in six, nearly 17%. On paper the longer card is cheaper.

The café study points the other way, for two reasons.

First, the faster visits are sales the café would not otherwise have made that month. By the authors' estimate, a member working towards the free coffee bought about two more coffees a month than the starting pace implied. Against those two sales stands one free coffee, which costs the café its ingredients and a little labour, not the menu price.

Second, a card that is never finished costs nothing in rewards and earns nothing either. The unfinished cards the researchers bought back showed no speed-up at all. The further away you put the reward, the more of your customers end up in that group. That last step is reasoning, not something the study tested: every card in it took ten purchases to fill.

So judge the distance by whether members finish, and price the reward separately.

Distance is measured in time, and time depends on how often your customers already come. Ten stamps is a month at a café where regulars come every three days. At a salon where a regular comes every six weeks, ten stamps is more than a year.

Start from how often a regular visits, and choose the number of purchases that fits inside the window you want.

Where visits are a fortnight or more apart, a month is the wrong window. Use the number of visits instead: a first reward after three or four visits is within sight; one after ten is not. A barber whose regulars come every three weeks can run a card of four and reward the fifth cut. For shops where the basket varies and visits are occasional, a stamp card may be the wrong tool altogether: see Points, stamps or visits: which should you use?

With points, do the same sum in ringgit. Divide the cost of the first reward by what a typical visit earns. As an example with made-up numbers: at 1 point for every RM1 and a typical bill of RM14, a reward that costs 100 points is seven or eight visits away. If your regulars come weekly, that is two months, and a cheaper first reward at 50 points would halve it.

Ten is the number most people know, and it was the campus café's. Ten suits a business whose customers pass an outlet several times a week. Copy the principle, a reward a regular can reach in about a month, and choose your own number.

The same study ran a second test. It gave 108 customers one of two cards at random: a plain card needing ten stamps, or a card needing twelve with two stamps already on it. Both groups had exactly ten purchases to make. The customers with the two bonus stamps finished in 12.7 days on average; the others took 15.6. In a separate questionnaire, the researchers described a pizza chain's stamp card to 65 train travellers, to some of them with two bonus stamps already on it, and asked how likely they were to join. The head start made no difference. A head start helps people finish. It does not recruit them.

A separate study published the same year named this the endowed progress effect: people given a task that already looks partly done were more likely to complete it, and completed it sooner, than people given the same task from zero.

native.cards does not start a new member's card with stamps on it, and staff cannot add stamps by hand: a stamp is given only when a purchase is recorded. You can still make sure nobody walks away with an empty card. Most members join at the counter while they are paying. If your team records that purchase as soon as the customer has added the card, the member leaves with a stamp on it. On a card of five, that is a fifth of the way there on day one.

Make it part of how the counter works: join, then record the purchase they are making, every time.

You set it in two places, depending on what members earn.

Stamps. When you create the card program, the Stamp card rule has Stamps to fill a card, which takes any number from 2 to 20 and starts at 10. The form adds: "The last stamp is the reward." The card draws one circle for each stamp and marks the last as the reward, so a card of ten takes ten stamps, and the tenth completes it. Stamps for one purchase sets how many a purchase gives, normally one. When a member claims the reward it uses one full card, and any extra stamps carry over to the next, so the new card begins at once.

Points. Each reward has its own cost under It costs. Nothing stops you offering two rewards: a small one a new member can reach in a few visits, and a larger one for the regulars.

Decide the stamp card's size before you submit the program for review. Once a program is live its earning rules are fixed; the Rules view says so in as many words: "Earning rules on a live program can't be changed here." Rewards are fixed too once added, but you can always add another. For a points program that is the remedy when the first reward turns out to be too far: add a closer one.

After launch the dashboard measures the distance for you. Open the program from Programs and choose Rewards. For each reward, Time to earn it shows how long a typical returning member takes, once ten returning members give it something to measure. A chart shows where members are on the card now, including those who have not visited in 60 days. If a reward takes a typical member longer than eight weeks, the program's Overview says so in a line such as "Free coffee takes about 10 weeks to earn", and the Rewards view highlights Time to earn it. Home may show the same line and suggest Add a closer reward.