Restaurant loyalty program: fill your quiet hours

A restaurant loyalty program earns its keep in one situation and one situation only: when it brings you a visit you would not otherwise have had. Most schemes fail that test. They hand a discount to the regulars who were already coming on Saturday night, book it as marketing, and quietly lower the margin on your best shift. This guide is about the version that works instead, store credit with rules you control, pointed at the hours your dining room is empty. It is written for owners and GMs in Thailand, so every point comes back to covers, baht and the shifts you cannot sell.

What a loyalty program actually is, once you strip the branding

At its core you are offering a guest something in the future in exchange for a visit now. There are only a few real shapes it can take.

A punch card is the cheapest and the weakest. It is impossible to measure, you learn nothing about who came back, and the card lives in a wallet you never see again.

A points scheme converts spend into a currency the guest has to learn. Points work at scale, where there is a big catalogue to redeem against. In a single venue they usually stall, because the guest cannot tell what a point is worth and the reward sits too far away to change behaviour tonight.

Store credit is a baht balance the guest can see. Spend ฿400, get ฿32 back into a wallet, come back and spend it. No conversion, no threshold, nothing to explain. This is the shape that changes behaviour fastest, because the guest is holding money that only works in your restaurant.

A membership charges up front for guaranteed perks, which suits a venue with a strong regular base and a reason to prepay, like a lunch club near an office tower.

Papaya Loyalty is built around store credit for exactly this reason. Guests earn a percentage of each order as credit, and you set the rules for how it comes back.

Why your quiet hours are the only place worth pointing a reward

Here is the arithmetic that decides whether a loyalty program makes or loses money.

At 8pm on Saturday your room is full. Every baht you hand out that night comes off a sale you already had, because there is no empty table left to fill. Rewarding that hour is the most expensive thing you can do.

At 3pm on Tuesday the kitchen is already on, the staff are already rostered, the rent is already paid. Nothing in your cost base changes if one more table sits down, so the only real cost of that cover is the food on the plate. At 30 percent food cost, a ฿400 bill in a dead hour leaves roughly ฿280 of contribution, and it costs you nothing you were not already paying.

Work the example through. Set the earn rate at 8 percent. A guest spends ฿400 and banks ฿32 in credit. They come back on a Wednesday afternoon to spend it, run a bill of ฿450, redeem the ฿32, and you take ฿418 on a visit that would not have happened. Food cost on that visit is around ฿135, and everything else was already covered.

Now the honest part, the bit most loyalty pitches skip. Some of that credit will be redeemed by guests who would have walked in anyway, and no rule set eliminates that. What the redemption window does is shrink it. If credit only works from 2 to 5pm on weekdays, the guest who already comes at 8pm on Saturday has to move their visit to use it, and a shifted visit into a dead hour is worth more to you than the same visit in a full one.

The rules that decide whether it makes money

A loyalty program is not an on switch, it is four numbers. Get them right and the maths above holds. Get them wrong and you are running a permanent discount.

1. The earn rate. Start at 5 to 8 percent of the bill. It is easy to raise later and painful to cut, so start low. Higher rates rarely buy more visits, they raise the cost of the visits you were getting anyway.
2. The redemption window. Your deadest stretch, and only that. Monday to Friday, 2 to 5pm is the classic in Bangkok. If you are a bar, it is the first hour after opening. This single rule does more for your margin than the other three combined.
3. The cap per bill. Never let credit cover the whole ticket. Capping redemption at 30 to 40 percent of the bill keeps every redeemed visit a paying visit, and stops one heavy saver from eating a whole afternoon of covers.
4. Expiry, or the lack of it. Long-lived credit is a real liability sitting on your books, but credit that expires quickly makes the guest feel tricked. Somewhere around three to six months is a fair line, and say it plainly on the wallet.

On Papaya those four live in the dashboard rather than in a staff member's head. You set the earn rate per outlet, define redemption windows by time of day, day of week or specific dates, cap the share of a bill credit can absorb, and scope a rule to one venue if you only want it running at the quiet branch. One balance follows the guest across your outlets and redeems at the till, on QR ordering and online, so there is nothing to reconcile by hand.

Membership without making anyone download an app

The fastest way to kill a loyalty program in Thailand is to require an install. Every download you demand is a guest who nods politely and never joins, especially a tourist watching their data roaming.

Handle the sign-up at the two moments the guest is already looking at their phone or your counter: checkout, where opting in is one tap on the payment screen, and the scan, when they are ordering from the table anyway. Both keep the flow going instead of interrupting it.

Then keep the wallet visible. A guest who does not know they have ฿32 sitting there will never come back for it. The balance belongs on the ordering screen and the receipt, with the window written next to it in plain words: spend it any weekday afternoon before 5pm.

Use a promotions engine for the offers that are not credit at all, a coded rainy-afternoon deal, an automatic lunch combo, a happy hour that turns itself on and off, each with a minimum spend, a usage cap and an end date. Scheduling them means they stop when you said they would, without anyone remembering to switch them off.

What loyalty data gives you beyond the credit itself

Once guests opt in, two things become possible that have nothing to do with rewards.

Reviews, caught at the right moment. Asking while the guest is still at the table beats emailing them tomorrow. Route the happy ones to your Google listing while the meal is fresh, and send anything under four stars to a private message to you instead, so you hear about the cold soup before the internet does. Papaya does this per outlet, so each venue points at its own listing.

Win-back, aimed at people who already know your food. Most ad spend chases strangers. The guests who visited twice and then vanished for two months are a better audience, and they are sitting in your own data. Papaya builds segments from visit time, spend and recency and syncs them to Meta and the Facebook Pixel, though that sits on the higher tiers, so check the pricing before planning a campaign around it. Pair it with an off-peak credit offer and you are pitching people who already liked the food once. Papaya Copilot then tells you which offers moved covers rather than just moved margin.

Where a loyalty program is the wrong fix

Honesty is worth more than the pitch here, because loyalty is often sold as a cure for problems it cannot touch.

If guests are not coming back because the food was inconsistent or service was slow, credit will not fix it. You will pay people a few percent to return to the same disappointment, and they will not come a third time. Fix the room first.

If your margin is already thin because your plate costs have drifted, start with food cost, not rewards. Giving away 8 percent of a bill you were barely making money on is how a loyalty program becomes a slow leak, and cutting your food cost is the better first month of work.

If you run a single till at a counter with a queue out the door at lunch, a loyalty program is not your first move either. Your constraint is capacity at peak, not demand at 3pm, and a lean POS is probably all you need for now.

And be clear-eyed about where credit works. It redeems at your till, on your QR and web ordering, and online with you. It does not follow a guest into a third-party delivery app, and any vendor who tells you otherwise is stretching the truth.

What to measure after you switch it on

Four numbers, one month, compared against the month before. That is enough to know.

• Covers inside the redemption window. The headline. If your Tuesday afternoons are not busier, nothing else matters.
• Repeat rate within 30 days. The share of guests who came back once inside a month. This is the number the program exists to move.
• Redemption in the window versus outside it. If most credit is being spent during your busy service, your window is too wide and you are funding a discount.
• Average bill on redeeming visits. It should sit at or above your normal check. If it is well below, tighten the cap per bill.

Run those for a month before you change anything. If the quiet window fills, widen it or open a second one. If it does not, you have spent a few percent of a small number of bills to learn something useful, which is the cheapest research in this business. You can turn the whole thing on for one outlet and start free.

Frequently asked questions

Start with one quiet window

A loyalty program is not a marketing badge, it is a way of buying covers in the hours you cannot sell. Set a modest earn rate, point redemption at your deadest window, cap what one bill can absorb, and let the guests who want the credit move their visit to get it. That is the whole play, and it works because your fixed costs at 3pm are already spent.

Pick your single quietest window, run it for a month, and watch covers in that window against the month before. If the room fills, widen the window or add a second one. If it does not, you have lost a few percent of a small number of bills and learned something cheap.