Guide 53 · Strategy
Does your business need a loyalty program?
A loyalty program is worth building when customers buy often enough to reach a reward, your margins can fund a reward they notice, you can link purchases to a person, and customers still spend part of their budget with competitors. If any of these fails, the program is unlikely to pay for itself. A simpler incentive, such as a stamp card or member prices, is often the better first step.
Are loyalty programs worth it? For some businesses clearly yes, and for others they are a steady cost that changes little. A program can look like an obvious next step: competitors have one, customers ask about it, and vendors make it sound simple. But it is a long-term commitment with real costs, and whether loyalty programs work depends on your customers, not on the idea. This guide sets out the tests that decide whether a program can pay for itself, the problems it cannot fix, and the cheaper options worth trying first.
Four conditions
What has to be true for a loyalty program to pay off?
A program pays off when the extra margin from behavior it actually changes is larger than what the rewards and the operation cost. Four conditions make that possible:
| Condition | The question to ask | Warning sign |
|---|---|---|
| Frequency | Do customers buy often enough to reach a reward? | Most customers buy once a year or less |
| Margin | Can your margins fund a reward customers notice? | The reward would consume a large share of gross margin |
| Identity | Can you link each purchase to a person? | Sales are anonymous with no natural sign-in point |
| Share left to win | Do customers also buy from competitors? | You already hold almost all of their spending in the category |
The last one is easy to overlook. A program redirects spending customers already make somewhere; it does not grow the category. If you already capture nearly all of a customer’s share of wallet, rewards subsidize purchases that were coming anyway.
Be careful with headline figures from large programs. McDonald’s CEO said loyalty members drive about 30% of its sales and eat at its restaurants about 2.5 times more often. That shows how much spending can run through a program, but the most frequent customers are also the most likely to join, so it does not show how much the program caused.
Sources for this section
McDonald's to create tiered loyalty program to lure repeat visitors (Reuters) F-MCD-01
Frequency test
How do you test purchase frequency?
Calculate the time to first reward for a typical customer: the reward threshold, divided by points earned per visit, divided by visits per month. In a hypothetical coffee shop where the median customer spends $6 twice a week and earns 1 point per dollar, a member earns about 48 points a month. A first reward at 500 points is more than ten months away, long enough for most members to forget the program.
If the wait is too long, do not simply raise the earn rate, because that raises the cost too. Lower the threshold and the reward together (250 points for a $2.50 reward instead of 500 points for $5), which halves the wait at the same cost, or give new members a welcome bonus. If customers buy so rarely that no reasonable design gets them to a reward, see loyalty for low-frequency purchases.
Margin test
How do you test margin?
Divide the give-back rate, the share of spend returned as rewards, by your gross margin. A 1% give-back at a 65% gross margin uses about 1.5% of margin, a modest experiment. The same 1% at a 10% gross margin uses a tenth of margin before any software or staff costs are counted.
Run this test assuming every point is redeemed. A design that only works if a third of points are never used is assuming members are engaged enough to change their behavior and disengaged enough to abandon their balances. Breakage that appears later is a cushion; breakage written into the business case is a weak foundation. Thin-margin businesses do run programs, but usually with partner funding or with rewards that cost them much less than their face value.
Captive customers
When is loyalty the wrong answer?
When customers return because they have to, not because they choose to. If a contract, a regulation or the practical cost of switching already brings them back, a program has no decision to influence. The points become a thank-you for purchases that were going to happen anyway, and later changes such as expiry feel like a broken promise.
Test it by asking what share of customers would buy from a competitor at the same price and effort if the program did not exist. If the answer is very few, the program would subsidize captive customers. Then try to name one specific action the program would cause that would not happen without it. If you cannot, do not build it. The exception is a benefit customers value apart from price, such as priority access or a service level that is genuinely limited.
What it cannot fix
What problems can a loyalty program not fix?
A program can make an existing reason to return stronger. It cannot create one. Rewards are usually small relative to the purchase, so they work at the margin of a decision the customer was already close to making.
- A price gap. If a competitor is 10% cheaper and your program returns 1% to 2% of spend, the reward covers a small part of the difference.
- Service failures. A customer who left after two damaged deliveries will not be won back by points. Fix the service.
- Availability. Members cannot buy what is out of stock or out of reach. If lapsed customers’ final months show stockouts and substitutions, the money belongs in inventory.
Rewards can also hide these problems. Repeat purchases driven by promotions keep the retention numbers healthy while customer preference declines underneath. If retention drops whenever promotions pause, the program is renting it.
Cheaper options
What are cheaper alternatives to a full program?
Several simple incentives test whether rewards change behavior without a points balance or a platform:
- A paper or digital stamp card: a free item after a set number of purchases.
- A one-time discount code after a customer’s third or fifth order.
- Member pricing on selected items. Tesco Clubcard Prices work this way: members scan the app or card at checkout to get exclusive prices.
- A voucher printed on the receipt for the next visit.
Count the real cost of each reward: a free coffee costs you the ingredients and labor, not the menu price. Set an expiry date so the offer does not become an open-ended obligation. Run it for a quarter and compare repeat purchases with the same period before.
Decide
How do you make the decision?
- Write down your median time between purchases and your gross margin.
- Run the frequency and margin tests above.
- Check that you can link purchases to customers, or plan how you will.
- Estimate how much of your customers’ category spending goes to competitors.
- Rule out the problems a program cannot fix.
- If the answer is still yes, pilot a simple version, such as a stamp card, before building a full program. Our guide on how to start a loyalty program covers the next steps.
A “no” is a useful result. It saves the cost of a program that would not have paid off and points the budget at the problem that actually needs solving.
More answers
Frequently asked questions
Is a loyalty program worth it for a small business?
It can be, if customers come back often and margins allow a reward they notice. Many small businesses start with a stamp card or a simple discount after a set number of visits, which costs little to run. Move to a full program only when that simple version shows it changes behavior and manual tracking becomes a burden.
What is the minimum purchase frequency for a points program?
There is no fixed rule, but a typical customer should reach a first reward they want within a few months. If it takes most of a year, many members forget the program before they get there. Calculate the time to first reward for your median customer before choosing the design.
How much of my margin should rewards cost?
Work out what share of gross margin the planned rewards consume. A give-back that is small relative to a wide margin is a safe experiment; the same give-back at a thin margin needs a large change in behavior to break even. Model it assuming every point is redeemed, and treat breakage as a cushion, not as part of the plan.
What are the pros and cons of a loyalty program?
The main advantages are more repeat visits, a larger share of customers' spending and data that links purchases to people. The disadvantages are an ongoing reward cost, software and staff time, a points liability on the books, and the risk of rewarding purchases that would have happened anyway. The four tests on this page decide which side wins for your business.
What are the alternatives to a loyalty program?
A stamp or punch card, a discount code after a set number of orders, member prices on selected items, or a voucher printed on the receipt. These avoid a points balance and most software costs, and they test whether rewards change behavior before you invest in a full program.
Do loyalty members really spend more?
Usually yes, but much of that is because your most loyal customers are the ones who join. McDonald's CEO, for example, said loyalty members eat at its restaurants about 2.5 times more often. The program's real effect is the difference between members and similar customers who did not join.