How to Start a Customer Loyalty Program for a Small Business
The short answer
To start a customer loyalty program for a small business, first check that customers are happy with the basics, then choose a type that fits how often they buy (stamp card, points, tiers, referral or club). Set an earn rate, a first reward reachable in a few visits and an expiry rule. Calculate the break-even uplift, keep marketing consent separate at sign-up, and measure results against a comparison group.
Want the tools, not just the method? The The Customer Loyalty Programme Kit kit turns this guide into ready-to-use templates, spreadsheets and printables, with an instant download.
- A loyalty program rewards a good experience; it cannot fix a poor one.
- Rewards are paid on all member sales, so the program must create extra sales to pay for itself.
- The break-even uplift tells you how much extra members must buy before you launch.
- Make the first reward reachable in a few typical visits and never design for breakage.
- Keep joining and marketing consent as separate choices.
- Members out-spend non-members because your best customers join first: measure change, not levels.
What are the steps to start a customer loyalty program?
Most small-business loyalty programs that disappoint fail in the planning, not the launch. These eight steps keep the expensive decisions in the right order. Each one is quick on its own, and the order matters more than the speed.
- 1. Write down the one customer behaviour you want to change, for example 'a second visit within 45 days'.
- 2. Check the basics: product, stock, service and price. Fix weak spots first.
- 3. Choose a program type that matches your purchase cycle and your tills.
- 4. Set the earn rate, the burn rate, the first reward and an expiry rule.
- 5. Calculate the real reward cost and the break-even uplift.
- 6. Map the data you need, write consent wording and get your terms checked by a lawyer.
- 7. Brief staff with a short, consistent sign-up ask and pilot in one store.
- 8. Measure active members, repeat rate and uplift against a comparison group, then adjust.
Is a loyalty program worth it for a small business?
The case for keeping customers is well established. Frederick Reichheld of Bain & Company argued for decades that customers who stay become more valuable over time, and reported that in financial services a 5% increase in customer retention produced more than a 25% increase in profit. A neighbourhood shop is not a bank, but the direction is the same: a regular who stays another year is usually worth more than a one-off visitor.
That does not mean every business needs a program. Customers already carry plenty: the 2025 Bond Loyalty Report found consumers belong to about 17.4 programs on average and see only about a third of them as delivering real value. A program earns its place when three things are true: customers already like the basics, they buy often enough to reach a reward within a few visits, and the numbers work. If customers come to you for convenience and friendly staff, a better product range or a second till may do more than any card.
A useful test: imagine you closed for a month. Would your best customers miss you, or simply walk to the next shop? A program can add a small reason to stay, but it cannot create the main one.
Which type of loyalty program should you choose?
Choose by purchase cycle, basket and the data you need. Digital formats such as a phone wallet pass, an app or POS-linked loyalty (points recorded at the till from a phone number) can carry any of these types.
| Program type | How it works | Usually fits |
|---|---|---|
| Stamp or punch card | Buy a set number, get one free | One main product bought often: coffee, lunch, car wash |
| Points | Earn per unit spent, swap for rewards | Varied baskets and regular visits: grocery, pharmacy front-of-store, salons |
| Tiers | Levels based on 12-month spend or visits | A clear group of top customers |
| Paid membership | A fee for ongoing benefits | Frequent buyers and spare capacity |
| Referral scheme | Members earn when friends they refer buy | Services people talk about |
| VIP or community club | Events, early access, recognition | Enthusiast categories such as books or specialty food |
How do you design rewards customers actually want?
Five numbers define the rewards. The earn rate is how members collect (one point per unit of currency, one stamp per coffee). The burn rate is how many points a reward costs. Multiply the earn rate by the value of a point and you get the give-back: reward value as a share of spend. Breakage is the share of points never redeemed. The real cost share is what a reward costs you: a money-off voucher costs its face value, but a free pastry costs only its ingredients.
Make the first reward reachable within a few typical visits. A hardware store that needs twelve visits for a small voucher will find that nobody talks about the program. Prefer rewards that are worth more to customers than they cost you: early access, priority booking, a free service upgrade at quiet times, a member evening. The Bond Loyalty Report 2025 found special access and personal experiences had become the top driver of perceived program value.
Never design for breakage. A program that only works because people fail to use it will feel mean, and members notice. Assume breakage cautiously, and remind members before points expire.
How much does a loyalty program cost? A worked example
Here is the trap most owners miss: rewards are paid on everything members buy, including purchases they would have made anyway. The only sales that pay for the program are the extra ones it causes. So the key number is the break-even uplift: the extra member sales, as a share of what members would have spent anyway, needed for the margin on those extra sales to cover rewards on all member sales plus running costs.
In plain words: break-even uplift = (reward cost as a share of sales + running costs divided by baseline sales) divided by (gross margin minus reward cost as a share of sales).
Take a two-store bakery (an illustrative example, not a forecast). It has 9,800 transactions a month at 6.40 each, a 62% gross margin, and expects 30% of transactions from members. It gives one point per unit spent and a free item worth 3 at 100 points.
| Line | Value | How it is worked out |
|---|---|---|
| Member sales per month | 18,816 | 9,800 x 30% x 6.40 |
| Reward face value | 3% of spend | 1 point per unit x 3 / 100 points |
| Real reward cost | 0.8% of member sales | 3% x cost share 0.38 x 70% redeemed |
| Reward cost per month | about 150 | 18,816 x 0.8% |
| Running costs per month | about 253 | Platform, printing, team incentive, messages, admin, set-up spread |
| Break-even uplift | about 3.5% | (0.008 + 253 / 18,816) / (0.62 - 0.008) |
| Extra member sales needed | about 660 a month | 3.5% x 18,816 |
| Extra visits needed | about 103 a month | 660 / 6.40 |
Reading helps; measuring tells you what to work on. These AI-graded assessments on AssessAll pair with this topic:
- Retail Counter Selling - Situational Judgement (AssessAll)
- Personal Data Judgment (AssessAll)
- Commercial Judgment & Market Sense (AssessAll)
How do you sign customers up and handle their data?
Sign-ups depend on the moment and the words. Ask at payment, after the customer has been served, using a short two-line ask: the benefit, then the question, then accept the answer. Ask every customer the same way; deciding who to ask from how someone looks is unfair and also loses you members. Keep incentives small and team-based, count only real consented sign-ups, and coach privately rather than posting league tables.
Collect the minimum: a phone number or email to find the account, a first name if offered, a birth month rather than a full date of birth for a birthday treat. Keep joining and marketing consent as separate choices, with marketing boxes unticked and each channel clear. Marketing rules differ by country. In the UK, the ICO describes a 'soft opt-in' for existing customers under strict conditions. In the US, the FTC's CAN-SPAM guidance requires an opt-out in every commercial email, honoured within 10 business days. In Canada, CRTC guidance explains when a purchase creates implied consent. In Australia, the ACMA expects unsubscribe requests to be actioned within 5 working days. Check your data protection law and ask an adviser before launch.
How do you know if your loyalty program is working?
The obvious comparison misleads. Members almost always out-spend non-members, because your best customers join first. Comparing the two groups after launch mostly measures who joined.
A fairer test compares change. Take members who joined in one window and bought before joining. Measure how much their spending changed from three months before joining to three months after. Do the same for a comparison group of customers who did not join, over the same months and in the same stores. The difference between the two changes is your measured uplift. Compare it with the break-even uplift.
Alongside that, track a few numbers monthly: active member percentage, repeat purchase rate, visits per active member, sign-ups per 100 transactions, reward cost as a share of member sales and the value of outstanding points. For campaigns, hold back a random control group and measure lift: response rate minus control rate. If you want a structured way to measure your team's starting point, the AssessAll Personal Data Judgment assessment covers the data-handling side of the job.
How do you run a loyalty program across several stores or a franchise?
Members expect the same program in every store: the same earn rate, the same rewards and the same welcome. That needs a short rulebook agreed before launch. Decide what is standard across the network (earn and burn rules, brand, core benefits, sign-up scripts) and what stores can do locally (events and partnerships within an approved budget).
Then decide who pays. A common model uses a program fund. Each store pays in the expected cost of the points it issues, after breakage, plus a small contribution to running costs, and the fund reimburses each store for rewards redeemed at its tills, wherever the points were earned. A store near offices that issues many points pays in; a weekend store where members redeem gets paid. Send every store a one-page settlement statement each month.
Finally, agree who controls member data, what each store can see and what happens if a franchisee leaves. Franchise rules differ by country, so these points belong in the franchise agreement or operations manual, checked by a franchise lawyer.
Common mistakes to avoid
For a guided set of tools, The Customer Loyalty Programme Kit from Bodhih Training includes a break-even calculator, a programme workbook, till scripts and a dated country rules sheet. If you want to turn your launch into a personal goal you will follow through on, Jobulary's guide to the WOOP method for professional goals is a practical place to start.
- Launching a program to cover for poor service, stock or prices.
- Choosing software before the design.
- Counting reward costs only on extra sales.
- Rewards that take a dozen visits to reach.
- Bundling marketing consent with joining, or letting staff tick the box.
- Judging success on member versus non-member spend.
- Running campaigns without a control group.
- Cutting reward value overnight without notice or protection for points already earned.
- Fraud checks that treat some customers differently from others.

Design, cost and measure your programme in one kit
The Customer Loyalty Programme Kit gives you the workbook, break-even calculator, till scripts, member messages and forms to run every step in this guide.
More from the Bodhih family
Questions people ask next
What is a good reward value for a small business loyalty program?
There is no single right number. Work out the give-back (earn rate times point value), the real cost share of the reward and your margin, then calculate the break-even uplift. Lower-cost rewards such as free items or experiences let you offer a higher face value for the same real cost.
What is breakage in a loyalty program?
Breakage is the share of points or rewards that are never redeemed. It lowers the program's cost, but designing for it makes a program feel mean. Assume it cautiously until you have a year of your own data.
Is a stamp card or points program better?
A stamp card is simple and cheap to run but gives away more per unit spent, is easy to fake and gives you no customer data. Points need a POS link, app or wallet pass but give you data and control. Compare both on your own numbers.
How long does it take to set up a loyalty program?
A small business can design, cost, brief staff and pilot a program in about eight weeks: two weeks of design and costing, two of data, terms and set-up, one of training and testing, then a pilot in one store before roll-out.
Do I need customer consent for a loyalty program?
Running the program needs only the data to keep points. Sending marketing messages usually needs a separate choice from the customer, and the rules differ by country and channel. Keep marketing boxes unticked, record consent and check your local law.
How do franchise loyalty programs share reward costs?
A common model is a programme fund: each store pays the expected cost of the points it issues plus a small running-cost contribution, and is reimbursed for rewards redeemed at its tills. The franchise agreement and a franchise lawyer decide the real model.
How do I measure the ROI of a loyalty program?
Compare the measured uplift, from a before-and-after comparison against a comparison group, with the break-even uplift. If the measured uplift is higher, the program is paying back on those assumptions; if not, change the rewards or costs.