15 Loyalty Metrics to Track in Your Shop

The figures and scenarios in this article are illustrative: adapt them to your margins and your own data. They are not EasyFid customer results. See how to measure your results.

The most useful loyalty statistics are the ones you can reproduce in your own shop. This page lays out fifteen metrics and how to calculate them. It does not present a 2026 industry survey or EasyFid customer results.

15 statistics to calculate for your shop

1. Sign-up rate

Number of customers signed up / number of customers offered the program × 100. Count every offer made so you have the right denominator.

2. 30-day return rate

Among customers who made their first purchase during a given period, what share made another purchase within 30 days? Wait a full 30 days for each customer before counting.

3. 90-day return rate

Same calculation with a 90-day window, useful for purchases spaced further apart. Don't compare a cohort observed for three months with one that started yesterday.

4. Visit frequency

Number of purchases / number of active customers over the period. Keep the same definition of a purchase and an active customer throughout.

5. Average basket

Revenue after discounts / number of receipts. State whether you use pre-tax or after-tax figures, and use the same basis for every period.

6. Spend per active customer

Revenue / active customers. It combines frequency and basket size, but can shift if the makeup of your customer base changes.

7. Time between purchases

Calculate the number of days between visits for customers who came back. The median can represent your regulars better than an average skewed by a few widely spaced purchases.

8. Share of inactive customers

Customers with no purchase since your chosen threshold / customers tracked. Adapt the threshold to your business: a monthly visit shouldn't be read the same way as a daily one.

9. Reward redemption rate

Rewards redeemed / rewards actually available. Exclude rewards that haven't been reached yet.

10. Time to first reward

Time between sign-up and the first threshold reached. Also look at customers who never reach it, rather than dropping them from the analysis.

11. Cost of rewards

Add up the cost price of any products given away and the discounts granted. A gift's cost isn't necessarily its sale price.

12. Program cost

Subscription + rewards + materials + management time + any paid messaging. Calculate it over the same period as the benefits you're measuring.

13. Acquisition cost

Campaign spend / new customers attributable to that campaign. Flag the limits of your attribution method.

14. Estimated additional margin

Margin observed with the program minus the margin you'd expect without it. This second figure is an estimate — price, weather, season and opening hours can all explain part of the gap.

15. Return on investment

(Estimated additional margin − program cost) / program cost × 100. The calculation is undefined if the cost is zero, and a revenue increase alone doesn't prove a positive ROI.

A reproducible worked example

Simulation, not a customer case. Take a cohort of 200 first-time buyers: 50 come back within 30 days, so the return rate is 50 / 200 × 100 = 25%. If a comparable cohort has 60 returns out of 200, its rate is 30%. The gap is 5 percentage points — about one fifth higher in relative terms. This calculation does not establish that the program caused the gap.

For an illustrative ROI, say the estimated additional margin is €80 and the total program cost is €50 for the same month: (80 − 50) / 50 × 100 = 60%. The result depends entirely on these assumptions — replace them with your own costs and margin before making any decision.

Interpreting results without overselling them

Set your definitions before you start. Compare periods of equal length and, where possible, comparable seasons. Note any changes in prices, hours, staffing or promotions. Don't mix up revenue and margin, signed-up customers and active customers, or relative change and percentage points.

Program members are sometimes already your most regular customers. A higher basket among them doesn't, on its own, prove the tool is working. A relevant comparison group or a controlled test can help you understand the actual effect; small cohorts remain sensitive to chance.

Sources and method

The fifteen metrics above are defined explicitly so your own calculations are reproducible — they are not averages drawn from a study. For an official definition of revenue, see France’s national statistics institute, Insee. The features available in EasyFid are described on the pricing page: this measurement guide doesn’t mean every metric is calculated automatically inside the app.

Also see our ROI calculation guide and our guide to setting up a loyalty program.

Frequently asked questions