Guide contents
You are preparing a 20% discount for your restaurant’s customers. Your list includes someone who visits every week, someone whose visits have become less frequent, and someone you have not seen for several weeks.
Sending everyone the same offer is easy. But what decision do you want each person to make?
The first customer already returns. The second may have changed their routine. The third may need a reason to return, even if price is not the issue.
Choose the behaviour you want to encourage before choosing a discount. A loyalty programme becomes more useful when rewards are considered alongside recorded visits.
The essentials in 30 seconds
- A points balance shows available rewards; it does not explain whether someone keeps returning.
- Review recorded visits, cumulative spending and the last recorded visit together.
- A regular customer, someone visiting less often and someone with no recent visit call for different responses.
- An offer can be a thoughtful gesture, something new or a specific benefit. A discount is one option.
- Measure returns and contribution after the offer, rather than coupon use alone.
01 · Why one discount can cost money without changing habits
A broad promotion can have a clear purpose: announcing an opening, introducing an offer or supporting a quiet service period. The difficulty arises when the same promotion becomes your response to every customer situation.
For someone who would have visited at the usual price, a discount reduces revenue without necessarily creating an additional visit. For someone disappointed by waiting times or food quality, it may leave the real problem unresolved.
Before launching, ask: are we trying to thank the customer, understand a change in visits or encourage a return? The answer changes the audience, message and acceptable cost.
02 · Points tell only part of the story
Two people can have the same points balance while having very different relationships with your restaurant. One earns points regularly; the other accumulated them months ago.
Start with three pieces of information:
| Information | Useful question | Limit to remember |
|---|---|---|
| Recorded visits | How many visits have been associated with this customer? | A lifetime total alone does not show a recent slowdown. |
| Cumulative spending | How much activity has been recorded on the profile? | High spending does not prove recent attendance or strong profitability. |
| Last recorded visit | How long has it been since a visit was added to the profile? | The person may have returned without being identified at checkout. |
To recognise a slowdown, you also need comparable periods or reliable knowledge of the usual visiting pattern. Three visits this month mean something different for a daily customer and a monthly visitor.
Checkout records describe what was recorded. They do not explain, on their own, why someone is visiting less often.
03 · Three situations, three possible responses
Use these situations as a working framework. They are management concepts, rather than permanent labels or categories automatically assigned by Kiwi.
The regular customer: recognise the relationship
A regular may appreciate a thoughtful gesture, an established loyalty reward or an invitation to try something new. Start with service too: recognising the person, keeping your promises and making the next visit enjoyable.
A discount can be part of that recognition. Decide what it rewards and what it costs rather than automatically reducing every bill.
The customer visiting less often: understand the change
Check the information before offering a reduction. Have your hours, menu or service changed? Could a closure, a trip or an unidentified checkout explain the gap?
Where appropriate, a friendly check-in can be more useful than a coupon. If you learn about a service problem, address it first.
The customer with no recent visit: give a relevant reason to return
A new lunch menu, a seasonal dish or a limited benefit can provide a concrete reason to come back. Choose something your team can deliver reliably.
The time interval should fit your business. Thirty days without a visit means something different for a daily café and a restaurant used for special occasions.
An absence in the records is a signal to examine. It does not prove the customer has left you.
04 · Before offering 20% off, calculate the remaining contribution
Contribution margin is the amount collected from a sale minus its variable costs. It helps cover fixed costs and then profit. The financial definition is explained in OpenStax’s contribution-margin method.
Consider a fictional meal priced at 100 MAD, with 40 MAD in variable costs per meal. All amounts use the same tax basis. Assume variable costs remain unchanged after the discount.
| Per meal | Usual price | With 20% off |
|---|---|---|
| Amount collected | 100 MAD | 80 MAD |
| Variable costs | 40 MAD | 40 MAD |
| Contribution margin | 60 MAD | 40 MAD |
The 20 MAD discount reduces contribution per meal by 33.3%: 20 ÷ 60.
To recover the 600 MAD contribution from 10 full-price meals, you would need 15 discounted meals: 15 × 40 = 600 MAD. That requires 50% more meals, assuming unchanged costs.
This is an educational scenario, rather than a demand forecast or pricing recommendation. Replace the figures with your ingredients, packaging, commissions and other costs that vary with sales. Revisit the model if extra volume requires more staff or exceeds kitchen capacity.
A targeted campaign can have a different economic effect: a genuinely additional visit brings contribution, while a discount on an already-planned visit reduces it. That difference is what you should try to measure.
Use the food-cost guide to work through ingredient costs and the break-even guide to connect contribution with fixed expenses.
05 · Write an offer the customer and cashier both understand
A useful proposal answers four questions: who is it for, why are you contacting them, what is the benefit and when does it end?
Here is a fictional example for someone without a recent recorded visit:
Hello [first name], our new lunch menu is available this week. If you would like to try it, we will include a tea with the menu from Tuesday to Thursday, until [date]. We would be happy to welcome you.
The benefit is specific and limited to a menu and time slot. Calculate its cost: a complimentary item still uses ingredients and service time.
You can also test an invitation without a financial incentive. Something new may be enough to give someone a reason to return.
Prepare the rules for staff before sending: eligible items, dates, service times, any minimum purchase, number of uses and whether other benefits can be combined. The customer and cashier need to understand the same promise.
Use available contact preferences, respect refusals and avoid repeated unanswered follow-ups.
06 · Use Kiwi to prepare the decision
Kiwi brings recorded customer activity and loyalty together. Visits, cumulative spending and the last recorded visit provide context for your next action.
In Kiwi, you can ask Kiwi AI which customers have no recorded visit in more than 30 days. The lookup shows the relevant customers, their cumulative spending and the time since the last recorded visit.
It covers customers identified at checkout and present in the customer book. The lookup runs at your request. You then review the results, choose the proposal and send the message manually.
The 30-day threshold produces a list to review. Your knowledge of the restaurant is still needed to decide whether an absence is unusual and which response fits.
The workflow: consult records → check the situation → choose the offer → review → send manually → measure.
07 · Measure returns and their value
Define the observation period before the test. Record the recipients, proposal, dates, cost of benefits and visits recorded after contact.
Track two separate measures:
- Observed return rate = recipients with at least one recorded visit during the period ÷ recipients contacted.
- Contribution from sales associated with the offer = sales net of discounts − variable costs − benefit costs not already included − direct campaign expenses.
These measures do not prove that every return was caused by the message. Some people would have returned anyway.
If your audience is large enough, retain a comparable group that does not receive the offer during the test. Compare returns over the same period. Treat differences in small samples as indications to confirm.
Also check whether people return after the benefit ends. A redeemed coupon describes one transaction; later visits tell you more about the relationship.
08 · Your first test: one situation, one offer, one review
- Check that visits are correctly associated with customer profiles.
- Choose a precise situation and objective: thank, understand or encourage a return.
- Review the audience and the limits of the records.
- Prepare one proposal and calculate its cost before contact.
- Set the conditions, budget and observation period; brief the team.
- Review the message and contact appropriate recipients.
- Review observed returns, contribution, later visits and lessons for the next test.
Start with a campaign you can explain and track. Improve the next proposal using your own results.
Frequently asked questions
Should we remove loyalty points?
No. Points can provide a clear reward. Combine them with visit information to understand when and how to act.
Does every inactive customer need a discount?
Examine the situation first. A new menu, an invitation or resolving a problem may better match the need. If you choose a reduction, calculate its cost and define the conditions.
How many days make a customer inactive?
There is no interval that suits every restaurant. Compare the absence with usual visiting patterns and the quality of the records. The Kiwi AI lookup described here uses more than 30 days without a recorded visit; interpret that threshold in context.
Is a high coupon redemption rate enough?
You also need to examine contribution and later returns. A campaign can distribute many benefits on visits that would have happened anyway.
The next offer starts with a question
Before choosing 10% or 20% off, ask: who do we want to bring back, and why do we think this proposal will be useful to them?
That is the starting point for loyalty that considers both the relationship and the economics of your restaurant.
Prepare your Kiwi demo: bring an example offer, its cost and the customer situation you want to address. Discover Kiwi and request a demo.
Sources and method
OpenStax: contribution margin and variable costs provides the financial definition. Amounts, customer situations and messages in this guide are fictional Kiwi teaching examples. Replace the assumptions with your own figures before making a decision.