Customer loyalty program for small business: what works

A customer loyalty program for a small business has one job. It gives a happy customer a reason to come back sooner than they otherwise would, in words they can repeat to a friend in one breath. Most schemes fail that test on day one. They ask for an app download, a sign-up form and eleven purchases before anything good happens. Here's what each type costs to run, how to size a reward against your own margin, and the check worth doing before you print a single card.

What a customer loyalty program can and can't do

A loyalty program doesn't make people like you. It gives people who already like you a reason to pick you this week instead of next month, and to walk past the shop two streets away when both would have been fine. That effect is real and worth having. It's also smaller than the software marketing suggests.

The economics are simple. Selling again to someone who already bought from you costs a fraction of finding a stranger. Harvard Business Review has put that gap at five to twenty-five times depending on the industry. A modest discount aimed at people who already buy from you is usually cheaper than the advertising it replaces.

Here's the other half, which the loyalty-app websites skip. A customer loyalty program can't rescue a product people don't rate. It can't undo a message that sat unanswered for three days. And it won't win back someone who has quietly moved to a supplier who answers faster. Those are the problems worth fixing first, because a reward makes them more expensive rather than less.

Five types of customer loyalty program for a small business

Nearly every small business loyalty program you'll come across is one of these five, or two of them stitched together.

The stamp card

Buy nine coffees, the tenth is free. It's the oldest version and still the best starting point for anything people buy often. Anybody understands it instantly, it costs a print run, and there's nothing to set up. The weaknesses are lost cards and the customer who photocopies theirs. Both go away if you record stamps against a phone number at the till instead of on paper.

Points

A point per dollar, five dollars off at a hundred points. Points suit shops where basket sizes vary a lot, because the reward scales with what someone actually spends. Keep the exchange rate round enough to do in your head. If your customers buy from you twice a year, skip points entirely. Nobody remembers a balance they last touched in March.

Tiers

Silver, gold and a bit of status. Tiers earn their keep when a small slice of your customers spends far more than everyone else, and you want to hold on to that slice specifically. They're fiddly to run, and demoting someone at the end of a year is a genuinely unpleasant email to send. Below a few hundred active customers the complication rarely pays.

Paid membership

An annual fee that buys free delivery, member pricing or first pick of a limited run. The paying part is what makes it work. Once someone has handed over money for access, they use it to justify the spend. It needs frequent buying and real confidence in what you're offering, so it's the hardest of the five to start cold.

Referral rewards

Reward both sides when a customer brings you someone new. For most small businesses this is the cheapest way to get a customer that exists, and it stacks neatly on top of any of the others. We've written a longer piece on how to build a customer referral program that people actually use.

What a customer loyalty program actually costs

Cost the reward at what it costs you, never at the price on the menu.

Say a coffee sells for $5 and the beans, milk and cup cost you $1.50. The free tenth coffee costs you $1.50, and it sits on top of $45 of sales from the nine paid ones. That's a bit over 3% off the whole run. A flat 10% discount would have bought you the same goodwill for three times the money.

Then run the sum again on your worst case. Every regular who was already coming back joins in the first week, and not one of them buys any more often than before. The reward is now a pure gift to people you already had. If that number still fits inside your margin, you can afford to be wrong about the scheme. If it doesn't, shrink the reward or lengthen the goal until it does.

Software is the second line. Dedicated loyalty apps start free at small volumes and climb quickly once you want tiers, integrations or branded emails, with plenty of them sitting in the $30 to $50 a month band and the heavier ones well into three figures. Prices in that category move constantly, so treat any figure you read, including that one, as something to confirm on the vendor's own pricing page.

The third cost is the one nobody budgets for, and it's the reason the next section exists. Every loyalty program makes more questions.

What if you sell to other businesses?

Stamps and points sit awkwardly in a business-to-business setting. The person placing the order often isn't the person paying, so a personal reward can read as awkward at best and improper at worst. Plenty of company purchasing policies say so in writing.

What works instead is a customer loyalty program for a small business built around terms rather than treats:

• Volume pricing that steps down automatically. Spend crosses a threshold, the unit price drops, no negotiation required.
• Priority answers. A named contact and a faster reply promise is worth more to a business customer than a free coffee, and it costs you nothing if your inbox is already in order.
• Committed stock or booked slots. Holding capacity for a regular is a genuine favour they can't get from a stranger, and it locks in your own forecast too.
• Free training or a yearly review. Cheap for you, useful for them, and it puts you in the room once a year.

Same principle as the stamp card underneath. Give something that costs you less than it's worth to the person receiving it, and make it easy to explain.

How to design a reward people finish

The difference between a card in a wallet and a card in a bin usually comes down to six choices.

• Make it reachable in one normal buying cycle. If a typical customer visits twice a month, a reward at ten visits is five months away and might as well be never. Six is better.
• Say it in one sentence. If you need a diagram, customers won't explain it to their friends, and word of mouth is half the point.
• Sign people up at the counter with a phone number. No app, no password, no email verification. Every extra step loses a chunk of the people who were willing.
• Give a head start. Cards handed over with two stamps already on them get finished more often than empty ones. People hate abandoning something they've started.
• Match the reward to what you sell. A free version of your own product costs you less than a gift card and keeps the customer inside your shop to claim it.
• Brief whoever is at the till. A scheme nobody mentions gets nobody signed up. One sentence at the point of payment does more than any poster.

Set a generous expiry on unclaimed rewards, tell people about it when they join, and remind them before it bites. Silent expiry is how a goodwill scheme turns into a complaint.

Fix your reply times before you launch anything

Before the print run, check one thing: how long does a customer currently wait for an answer when they email or message you? If the honest answer is a day or more, that's where your repeat business is leaking, and a stamp card won't plug it. We've made the same argument at more length in our guide to getting repeat customers, because it keeps turning out to be the same story.

A loyalty program also creates support work of its own. "I lost my card." "How many points do I have?" "Does this work online as well as in the shop?" "Can my partner use mine?" Four or five questions will cover most of what anyone ever asks you about it.

Write those answers once and put them somewhere customers can reach them without asking. That's most of what a help centre is for. Our own support bot runs off the same articles you publish there: hand-written rules search what you've written, then a model replies in plain sentences using only those articles. If nothing you've published covers the question, it says so and hands the conversation to you rather than inventing a rule you never made. When it drafts a new article from a question you've solved, that draft waits for your approval before any customer sees it.

All of that is included on every SupportifyGPT plan, including the $15 one, under flat monthly pricing with no fee per answer and no ticket overages. There's a 14-day free trial if you'd rather test it against your own inbox than take our word for it.

How to tell whether it's working

Check four numbers, once a month, on one sheet of paper.

• Sign-up rate. What share of paying customers joined? Under a quarter usually means the offer is confusing or nobody at the till is mentioning it.
• Redemption rate. What share of members claimed a reward? A low number isn't free money, it's evidence the goal is too far away.
• Purchase frequency, before and after. For members only, compare how often they bought in the six months before joining with the six months after.
• Average spend per visit. Watch this one for the opposite reason. If it drops as fast as frequency rises, you've bought the same revenue at a discount.

That third number is the one people get wrong. Comparing members against non-members flatters every scheme ever run, because your keenest customers join first and would have come back anyway. Comparing each member against their own earlier behaviour is harder to fake.

Give it a full quarter before you judge it, then change one thing at a time. And keep watching the customers who stop showing up, since a reward scheme has a habit of hiding a churn problem behind a nice enrolment chart. Our notes on reducing customer churn pair well with this. The rest of our small business guides cover the neighbouring ground.

Start with the cheapest version you can run by hand. The simplest customer loyalty program for a small business is a stamp card and a column of phone numbers in a spreadsheet, and it will tell you inside a quarter whether your customers want this at all. That answer costs you a print run, which is a bargain compared with a year of monthly software fees spent finding out the same thing.