Is a loyalty program worth it for your business?

In 2022 six researchers pooled loyalty program studies published or available between 1990 and 2020: 429 measured effects in all. They report "strong evidence" that programs enhance customer loyalty, but mostly of one kind. Programs "particularly enhance behavioral loyalty", meaning visits and purchases, while "shifting consumers' attitudinal loyalty is more challenging".

A card can give someone a reason to choose your counter this Thursday. It is a weak tool for making them fond of you. A sceptical paper from 1997 put it from the other side: programs "must enhance the overall value of the product or service". A program adds to something people already want to come back for. It cannot replace it.

At a convenience store franchise, customers who were heavy buyers when the program began "were most likely to claim their qualified rewards, but the program did not prompt them to change their purchase behavior". Those who started as light or moderate buyers "gradually purchased more and became more loyal to the firm".

A second study, in the account published by one author's business school, followed more than 5,500 clients of a men's hair salon chain for 30 months. Ten months in, the chain began giving members $5 off for every $100 spent. The authors estimate that this raised the value of a customer by "almost 30% over a five-year horizon". More than 80% of the lift came from fewer clients leaving and less than 20% from more frequent visits. The effect on spend per visit was "negligible".

The reward was small, 5% back, so the discount alone is unlikely to explain it. The school's account suggests "some psychological benefit, just from being part of the program". The clients who gained most were those who had shown "either great engagement with the firm or almost no engagement at all", because both groups became less likely to leave. A program can help you keep a regular. Neither study reports a regular buying more.

Occasional customers have more room to change. That is our reasoning, not a finding we can quote. A regular already buys from you nearly every time the need arises. The customer you see once a month buys the same thing elsewhere in between, and those are the visits a card can win.

The convenience store finding has a cost. The customers who change least claim their rewards most readily, so your program has to be able to afford its regulars.

A study of 358 grocery store brands in 27 European countries asked when a program helps a retailer's sales productivity. It found a positive effect for the more basic kind of program, the one "that offers direct and immediate rewards". The effect "disappears when retailers operate a more complex progressive-reward system", or when the program is shared between several companies. It was lower for discounters, and higher where fewer competitors ran a program of their own. The published summary does not describe the designs, so take the direction and not the detail.

The evidence is not all one way. A study of 322 listed companies reports "significant increases in sales and gross profits" for programs "with tiers or earning mechanisms". Neither sample is a kopitiam.

For a small business, we read this as: a rule your staff can say in one sentence, and a reward the customer gets soon. There are no tiers, levels or multipliers in native.cards. Members earn Points at a rate you set, Stamps towards a card, or a count of Visits.

An example with made-up numbers. A bakery serves 500 customers in a month. Fifty are regulars who come eight times. The other 450 come twice. It launches a card. In the first month 40 of the regulars join, being at the counter twice a week, and 60 of the others.

Members: 40 × 8 + 60 × 2 = 440 visits by 100 people, or 4.4 each. Everyone else: 10 × 8 + 390 × 2 = 860 visits by 400 people, about 2.2 each.

Members visit twice as often as non-members, and nobody has changed what they do.

The gap only shows who joined. A study of Dutch supermarket shoppers, covering all seven grocery loyalty programs and the 20 major supermarket chains, corrected for it. Membership had "a small positive yet significant effect" on the share of their grocery spending that a chain received. That effect was "seven times smaller than is suggested by a naïve model", one that ignores who chooses to join.

Never judge a program by comparing members with non-members. Compare the same customers before and after: How to test whether your program makes a difference shows how.

These three are often quoted to sell loyalty programs. None is a finding you can apply to your own business.

"Keeping a customer is five times cheaper than winning one." The authors of the book Loyalty Myths looked for its origin and found only an attribution to a research project in Washington in the late 1980s. By 2014 Harvard Business Review had it as "anywhere from five to 25 times", "depending on which study you believe", and named no study.

"A 5% rise in retention lifts profit by 25% to 95%." The source is a 1990 Harvard Business Review article. Byron Sharp of the Ehrenberg-Bass Institute describes it as "merely a 'thought experiment'". In it, defection falls from 10% to 5%, "a halving", at no cost. The profit that rises is the profit from one customer, not the company's.

"80% of your sales come from 20% of your customers." The pattern is named after Vilfredo Pareto, an economist describing the distribution of wealth. Measured on the buyers of consumer brands, the share depends on how long you count. The Ehrenberg-Bass Institute puts a brand's heaviest fifth of buyers at about 40% of its sales in a quarter and over 50% in a year. A six-year study of packaged-goods purchases found an average of 73%. Count your own before you quote it.

The dashboard counts only your members, and only what is recorded for them. Home says of its spend figure: "This is the qualifying amount recorded for members, not your total takings." Four of its figures show the behaviour the research says to watch.

None of the four compares members with non-members, so read each against your own earlier months. A figure shows a dash until there is enough history: Came back says "Needs an earlier period to compare", and Return window waits for 30 return visits. The figures describe what members did. They cannot show that the program caused it.

The first month costs money and flatters you. The first to join are mostly your most frequent customers. They collect rewards on visits they were already making, and they make your members look excellent.

Occasional customers change slowly: "gradually" was the convenience store study's word. Two weeks after your first members join, Second visit begins to count them. It is the figure to watch through the first quarter: see How to get new customers to come back a second time.

Profit comes later than sales. The 322 listed companies introduced their programs between 2000 and 2015. Sales and gross profit rose within the first year, and the gains held for at least three. But the effects on gross profit did "not become significant until the second quarter" after launch and lagged "substantially behind sales". Those are large firms, so borrow the timing only: rewards go out from the first week, and a customer who stays shows up only as time passes.

Fix the day you will judge it. At the end of the second quarter, switch Home from 30 days to 90 days and read Came back: of the members who visited in your first quarter, the share who visited again in your second. That is the share of customers you kept. Set your gross profit for those six months beside the same months a year earlier, if you were trading then.