Keeping a customer costs less than winning one

Most restaurants spend the bulk of their marketing budget on people who have never walked in, and almost nothing on people who have already come three times.

That is backwards. Someone who came once already knows where you are, what you cook and what you charge. Bringing them back costs a fraction of persuading a stranger to try.

The number that explains everything

lifetime value = average check × visits per year × years retained

A customer spending 180, visiting twice a month, staying three years:

180 × 24 × 3 = 12,960

That is what you actually lose when they stop coming because one order arrived cold. Seen that way, fixing their experience is cheaper than any advertisement.

Why most loyalty programmes fail

1. Too complex. “A point per 10, a point is worth 0.7, not valid on promotions, expires in 90 days.” Nobody will do that arithmetic.
2. The reward is too far away. A reward after 15 visits looks impossible, so the pursuit stops at visit two.
3. Nobody knows about it. A programme the staff never mention at payment does not exist in practice.
4. It rewards people who were coming anyway. A standing discount for your weekly regular is not loyalty, it is a price cut.
5. It is not measured. Without measurement you cannot tell whether you bought visits or built a habit.

Design rules

Make it explainable in one sentence. “Every 100 = 1 point, 10 points = 100 credit.” Done. If it needs a longer explanation, redesign it.

Make the first reward close. A small, fast reward proves the programme is real. A large, distant one proves the opposite.

Reward the behaviour you actually want. Do not reward spend alone:

• Visiting during quiet hours
• Trying a new item
• Referring a friend
• Returning packaging
• Rating an order

Do not make it a permanent discount. A discount re-prices your menu forever. Credit and rewards bring someone back a second time.

Segment before you send

Three simple axes (RFM):

Then treat each group differently:

Sending all four the same offer wastes money on people who were coming, and is not enough to bring back people who left.

Referral: your cheapest channel

A friend’s recommendation outperforms any advertisement. Make referrals reward both sides — the referrer and the newcomer. Rewarding only one side makes the referrer feel they are selling to a friend.

What to measure

A very low redemption rate is not a saving. It is evidence that customers do not believe they will ever reach the reward.

The foundation that makes this possible

All of the above needs one thing: knowing who bought what, and when. A restaurant recording sales with no customer identity cannot segment its base or measure its return rate.

In GoSufra, customer profiles carry order history and saved addresses, loyalty points run on earn and redeem rules, referrals reward both sides, promotions and coupon codes cover percentage, fixed and BOGO, and ratings can be answered from the dashboard — all of it surfacing as customer insight reports rather than impressions.

Start with four steps

1. Attach every order to an identified customer, even if only a phone number.
2. Launch one simple earn rule and one close first reward.
3. Train staff to mention the programme at payment.
4. Measure the 90-day return rate before launch and two months after.

Then do not add a second rule until customers understand the first.