Quick answer
A well-built loyalty program is about getting existing customers to come back more often and spend more, not about attracting strangers. Choose a simple mechanic (stamps or points), calculate the reward's real cost on food cost (not on menu price), and keep the value returned between 3% and 6% of the spend that earned it. Measure two things: return frequency and the average check of your members.
What a loyalty program is actually for
The most common starting mistake is thinking a loyalty program acquires new customers. It doesn't. Its job is to raise the visit frequency and average check of people already walking through your door. Acquiring a new customer costs five to six times more than bringing back an existing one — and that's exactly where loyalty pays off.
Think in terms of two levers:
- Frequency: the customer who used to come once a week now comes twice, because they're "two stamps away from a free coffee".
- Average spend: the customer adds a pastry or a second coffee to hit the reward threshold.
If you move neither lever, the program is just a disguised discount that erodes your margin. Before you launch, set a numerical goal: "I want 20% of my regulars to go from 4 to 6 visits a month".
Choosing the mechanic: stamps, points or tiers
There are three families of mechanics. The choice depends on your type of venue and average check.
| Mechanic | How it works | Best for | Pros | Cons | |---|---|---|---|---| | Stamps | 10 purchases = 1 free | Bar, coffee shop, gelateria | Dead simple, zero tech | Doesn't reward bigger spenders | | Points | $1 spent = 1 point, reward threshold | Restaurant, pizzeria | Rewards real spend | Needs POS or app | | Tiers | Rising status with perks | Venues with loyal regulars | Builds belonging | Hard to communicate |
For a bar selling coffee and pastries, stamps are unbeatable: they cost pennies and customers grasp them in a second. For a restaurant with variable checks, spend-proportional points are fairer and more profitable. Tiers only work if you have a deeply loyal base and the ability to communicate the perks.
How much the reward should be worth: the calculation that matters
This is where it all happens. The reward's value must be calculated on your real cost, not the list price. A coffee you sell for $1.50 costs you about $0.30 in pod, milk, sugar and energy. Giving away one coffee in ten looks like a 10% discount, but your effective cost is far lower.
Real value-returned formula:
Reward cost = food cost of the reward
Spend generated = purchases needed for reward × average check
Real cost % = Reward cost ÷ Spend generated
Bar example:
- Reward: 1 free coffee every 10 coffees bought
- Food cost of the free coffee: $0.30
- Spend generated: 10 coffees × $1.50 = $15.00
- Real cost % = 0.30 ÷ 15.00 = 2.0%
Just 2.0% cost to retain a customer: an excellent marketing investment. To set thresholds and reward value precisely, use the bar loyalty points calculator. Rule of thumb: keep the reward's real cost between 3% and 6% of the spend it generates. Below 3% the reward doesn't motivate; above 6% you erode too much margin.
Setting the right threshold
The threshold (how many purchases or points the reward requires) decides whether the program motivates or frustrates. Too high and the customer quits; too low and you give away margin without driving extra behavior.
Start from your average check and current frequency. If you don't know these numbers well, work them out first with the average check calculator. The ideal threshold is reachable in 4 to 8 visits: enough to generate repeat returns, close enough to stay motivating.
| Venue type | Average check | Suggested threshold | Reward | |---|---|---|---| | Coffee shop | $2.00 | 10 stamps | 1 free coffee | | Pizzeria | $20 | 100 points ($1=1 pt) | $10 off | | Restaurant | $40 | 300 points | Free starter or dessert |
A behavioral psychology trick: give one or two "welcome" stamps at sign-up. A card with 2 stamps out of 10 feels "already started" and is more likely to be completed than an empty 0-out-of-8 card, even when the required number of purchases is identical.
Collecting data without scaring the customer
A loyalty program's hidden value isn't the stamps: it's the data. Knowing who your regulars are, how often they return and what they order is worth more than the cost of the free coffee. But data collection must be handled well.
If you ask for name, email or phone to send promotions, data-protection rules kick in: you need a privacy notice and explicit marketing consent, separate from program enrollment. You cannot make the reward conditional on consenting to emails.
Practical tip: only ask for the data you'll actually use. One email to send a monthly offer is worth more than ten fields filled in and never touched. Start with the anonymous paper card, and digitize only when you have a concrete reason to collect contacts.
Communicating the program to staff and customers
A loyalty program fails nine times out of ten not because of the mechanic, but because staff don't offer it. The cashier who forgets to ask "do you have a card?" wipes out the whole investment.
Three rules:
- Train the staff: every team member must explain the program in one sentence and offer it at every till.
- Make the benefit visible: "you're 2 coffees from a free one" is the sentence that brings the customer back tomorrow.
- Reward staff too: a small incentive for whoever signs up the most customers speeds adoption in the first weeks.
Measuring results: the KPIs that count
Without measurement, you don't know whether you're investing or burning money. Three indicators are enough:
- Member return rate: how many come back at least twice a month. Compare with non-members.
- Differential average check: do members spend more than non-members? By how much?
- Program ROI: revenue generated by members minus the cost of rewards issued and management costs.
Sample monthly ROI calculation:
- Incremental member revenue: $2,400
- Cost of rewards issued: $180
- Management cost (app/printing): $40
- ROI = (2,400 − 220) ÷ 220 = 9.9x
Review these numbers monthly for the first three months, then quarterly.
Common mistakes
- Reward too far away: a 20-purchase threshold discourages. Stay between 4 and 8 visits.
- Calculating the discount on the sale price instead of food cost: you scare yourself and offer rewards that are too thin.
- Not training staff: the card stays in the drawer if nobody offers it.
- Asking for too much data: long forms cut sign-ups. Ask only the essentials.
- Not measuring: without KPIs you don't know whether the program builds or erodes margin.
- Rewarding with high-food-cost items: give away what costs you little (coffee, house dessert), not the priciest dish.
- Forgetting point expiry: points without expiry become an endless liability. Set a 12-month validity.
Related resources
- Bar loyalty points calculator — set thresholds, reward value and the real cost of your points scheme.
- Average check calculator — measure customers' average spend, the starting point for defining program thresholds.