Loyalty features that move repeat purchases: patterns from work on restaurant and coffee apps

Short answer. Two large US chains publish their own loyalty results, and our engineers worked inside both app teams. Their public results suggest loyalty brings guests back when it sits inside ordering and paying and uses a simple earn rule. It works best when it talks to each member about their own balance. Before you build, run the maths. Take our calculator's conservative planning assumptions. There, a £12,500 loyalty module repays itself within a year only if the orders it covers bring in about £780,000 a year. Try your own numbers in the app cost calculator and the ROI calculator.

Our role, and whose numbers these are

Our engineers worked inside each client's own team, as the two case studies record. Each engagement ran for about a year, on the client's full US working day, and we were contracted through an intermediary rather than by the brand. We did not design or run either loyalty programme.

• Red Robin, a US restaurant chain (case study). A developer from our team joined the client's product team. The work covered app areas connected to ordering and the Red Robin Royalty dashboard. It also supported segmentation, personalised messages and member offers.
• Dutch Bros, a US drive-thru coffee chain (case study). Our Android developer joined the client's internal team and delivery process, working on Dutch Rewards loyalty features in the Android app.

Every figure below that describes either brand is that brand's own public figure. It comes from the brand's filings, its earnings calls, or trade-press reports of those calls. None of them is our result, and each one says whose it is.

Why repeat visits are worth building for

Our founder's rule for restaurant apps: "Don't build an app to find new customers. Build it to keep the ones who actually pay you." (LinkedIn, 5 Feb 2026)

The research usually quoted here is a 1990 Harvard Business Review article by Frederick Reichheld and W. Earl Sasser, "Zero Defections". Its summary reports that cutting the rate at which customers leave by 5% raised profits by 85% in one bank's branch system. In the same study, the cut raised profits by 50% in an insurance brokerage and 30% in an auto-service chain. Retellings online often say "up to 95%", but that figure is not in the article's own summary.

Two caveats before you lean on it. These are three service businesses from 1990, not restaurants, and the summary itself says the size of the effect "varies by company and industry". It is also about keeping customers, not about points. A loyalty programme is one way to try. The section on telling whether loyalty caused the lift shows how to check that yours works.

Six patterns

Each pattern starts with what the brands' apps and reports show, with the source linked. It ends with What to build: our recommendation, not a description of either client's code.

1. Put the balance inside ordering and paying

In both apps, as their case studies list, loyalty lives where the guest orders. Red Robin's Royalty dashboard sits in the same app as ordering for pickup, curbside and delivery. Dutch Bros combines Order Ahead, payment with Dutch Pass and Dutch Rewards in one app.

Dutch Bros' management links the two directly. On the company's Q4 2025 earnings call, CEO Christine Barone said Order Ahead had "proven to be a powerful catalyst for our loyalty program". She said it pushed the share of transactions from Dutch Rewards members to "70% plus each full quarter since launch". By Dutch Bros' own count, Order Ahead was about 14% of its transactions in the fourth quarter of 2025.

What to build: show the balance and the next reward on the order screen, and let the guest redeem in the same checkout. A separate loyalty app, or a card the guest must remember to show, adds a step at the moment you want none.

2. Use an earn rule people can work out in their head

Red Robin relaunched Royalty in May 2024 with a plain rule: one point per dollar, and a $10 reward for every 100 points. Under Red Robin's rules, rewards must be used within 90 days. On the next earnings call, CEO G.J. Hart gave the company's own early results (reported by Nation's Restaurant News). New members were coming back sooner: the average time to a second visit fell from 51 days to 39. Loyalty guests' checks were about $4.40 higher than non-members', up from about $2.90. Hart also said the 90-day window to use a reward gave guests a reason to return.

What to build: a progress bar toward the next reward that the guest understands without a help page. If you need a table to explain your tiers, simplify the tiers first. If rewards expire, print the date on the reward itself.

3. Send offers to groups, not to everyone

Both apps offer members more than points. Red Robin lists personal offers and member-only rewards. In February 2025, FSR Magazine reported that Red Robin had added guest data capabilities for personalised messages and member promotions. That is the area our developer supported in the app: segmentation (splitting members into groups), personalised messages and member offers.

Dutch Bros lists offers and challenges for different member groups. On the same Q4 2025 call, its CEO set out the plan: "expanding our customer targeting capabilities, reaching the right customer at the right moment".

What to build: offers that marketing can target and change without waiting for an app release, and a way to show each member only the offers meant for them.

4. Remind members about their own balance

Both apps send notifications and reminders about rewards and personal offers. Our own evidence for keeping every message personal comes from a very different app. In a legal client app we built, every message is about the client's own case, and 77% of users turned on notifications. That is our own figure, from the case study (how we got there).

What to build: reminders tied to a real event for that member: a reward they can now claim, a reward about to expire, an offer that ends today. Put a cap on how many a member gets each week.

5. Treat a wrong balance like an outage

At Dutch Bros, the brief was that loyalty features had to stay stable. In our view, a wrong balance at the counter costs more trust than a slow screen does, because the guest feels cheated in front of staff.

What to build: one source of truth for the balance on the server, and a clear "pending" state for points that haven't landed yet. Release each change to a small share of members first.

6. Add engineers to the team that owns the app

Both case studies record the same brief: more engineering capacity for loyalty without growing the client's own headcount. So both chains added outside engineers to their existing product teams. At Red Robin, the programme was being rebuilt around guest data, so new features had to move quickly. Our engineers worked in each client's own processes.

If you already have a product team and a loyalty roadmap, adding engineers to it keeps the people who own ordering and payments in the loop. Handing loyalty to a separate agency splits that ownership. Our mobile team augmentation page explains how our engineers join your team, including on loyalty and ordering apps.

Feature checklist

Ticks show what each case study lists for the brand's public app, not what we built. The "why" and "build note" columns are our own reasoning.

What a loyalty module costs

Our cost calculator adds loyalty as a fixed module of 275 hours, about £12,400 at its blended planning rate of £45 an hour. That is a planning figure: our actual rates are set per engineer, at about $30–90 an hour depending on skill and experience. The module is flat: design level and platform choice do not scale it. Our ROI calculator uses £12,500 for the same module. A referral programme is a smaller module: 115 hours, about £5,200.

Before you compare quotes, ask each team to list what its loyalty estimate includes and what it leaves out. The parts most worth asking about are the points engine, integration with your tills, fraud controls, and the tool marketing uses to build member groups.

When it pays back

The ROI calculator's loyalty add-on rests on two numbers per scenario. One is how many customers join. The other is how much more revenue each member brings, after the cost of rewards. These are our planning assumptions. We have not published a source for them, and they are not results from either brand. The calculator applies them to all the order revenue you enter, not only to orders placed in the app. That revenue is orders a month × average order value × 12.

Here is how much yearly order revenue the module needs to repay £12,500 within 12 months:

Add a year of maintenance (the calculator uses 17.5% of the build a year), and the conservative line rises to about £918,000.

Plan on the conservative line until your own data says otherwise. Replace our planning assumptions with your own sign-up and spend data as soon as you have it.

How to tell whether loyalty caused the lift

A member-versus-non-member gap, like Red Robin's own $4.40 figure, shows that members spend more. It can't show how much of that the programme caused, because your most frequent guests join first.

Totals mislead too. Red Robin reported a 13% rise in loyalty transactions after the relaunch. FSR's report describes it as more frequent visits plus new members joining. Member transactions rise every time a regular signs up, even if nobody visits more often.

Three better checks:

1. Compare before and after, not members against non-members. Red Robin's own figure is this kind of measure: after the relaunch, the average time to a second visit fell from 51 days to 39.
2. Hold a small random group out of each offer (a hold-out group), and compare their visits with members who got it. That is the cleanest test of whether an offer adds visits or just discounts visits that would have happened anyway.
3. Count the reward cost. An offer that brings a visit forward by two days, at the price of a free drink, may lose money.

What to track from day one:

• the share of customers who join;
• the share of transactions from members (Dutch Bros reports about 72% of its own transactions in 2025);
• where member visits come from: new, lapsed or regular guests. Red Robin reported that 25% of its loyalty visits came from new members and 20% from lapsed ones;
• time to second visit, and the redemption rate;
• each offer's result against its hold-out group.

When not to build one

• One site and no customer database. Our founder puts it plainly: "If you're a small coffee shop, a custom app is usually a waste of money". In his view, the case changes once a chain already has a customer database. Below the break-even line above, use the loyalty features in your till system or an off-the-shelf app.
• Ordering in your app doesn't work well yet. Loyalty works inside ordering (pattern 1). Fix the ordering flow first.
• You sell one-off services. Points don't pay back when clients rarely return. Our ROI calculator switches the loyalty add-on off for its legal template and offers a referral module instead.

Next step

• Already have a product team? See how our engineers join it: mobile team augmentation, including our work on loyalty and ordering apps. For the monthly cost against an agency or freelancers, read Agency, freelancers or dedicated developers?
• Price it: add the loyalty module in the app cost calculator, then check payback in the ROI calculator with the e-commerce template.
• Book a 30-min ROI review (below) and bring your monthly orders, your average order value, and what your loyalty programme looks like today.
• Read the cases: Red Robin · Dutch Bros.
• Building a new app around ordering and loyalty? Start with MVP app development.
• Native or shared code for your ordering app? How we choose.