The loyalty question a board actually wants answered is not how many members you have, it is how many of them show up in the channel that still produces most of your margin. For a food brand that is the physical restaurant. For a retailer it is the store. And for almost every operator I talk to, that number is embarrassing: a loyalty programme driving ninety percent of app orders and roughly a fifth of counter orders is not a loyalty programme, it is a digital coupon engine wearing a bigger hat.
Chipotle has been working on exactly this gap in public, quarter by quarter, in a way few other operators are willing to be this precise about. I have been watching it closely because the shape of the problem is not specific to burritos. The same mechanics apply to a fashion retailer with a strong site and a quiet counter, to a grocer with a slick app and an anonymous checkout line, and to any WooCommerce operator who runs a physical pickup location or a pop-up store on top of their online storefront.
What Chipotle reported for the third quarter of 2026 is the clearest run of numbers I have seen on this problem all year. I want to walk through it, say what I think it actually proves, and then be honest about what a mid-market operator can copy and what a mid-market operator cannot.
One Operator Who Tells You Exactly Where the Gap Is
Chipotle Mexican Grill is a 3,800-plus unit fast-casual chain with a disproportionately strong digital business. In its third-quarter 2026 earnings, reported 28 October 2026, digital orders accounted for 38.8% of food and beverage revenue, up from 36.7% a year earlier. Revenue rose 9.6% year on year to $3.3 billion, and comparable sales rose 2.1%. The reason this company is worth studying for loyalty is that the digital half of its business trained its loyalty programme first, and left a visible gap in the other half.
The July 2026 restaurant technology piece around the Q2 earnings call put the gap in one line: about 20% of in-restaurant transactions were tied to a rewards account, versus nearly 90% of digital ones. That is the before-state. A seventy-point spread between the channel that is already measured and the channel that is not.
The Fix Started With a Programme Rewrite, Not a Points Inflation
On the Q4 2025 earnings call, CEO Scott Boatwright told analysts the rewards programme would be relaunched in the spring to target in-restaurant customers, reduce friction at checkout, and use AI for more personalised offers. The February 2026 coverage noted the programme was at roughly 21 million active members and about 30% of sales at the time, and quoted Boatwright on the “significant runway for growth” he saw in the channel.
The relaunch, called Rewards on Repeat, shipped in April 2026. Curt Garner, Chipotle’s president and chief strategy and technology officer, described the design principle on an earnings call reported by Marketing Dive on 28 May 2026: add value, more choice, and more flexibility “without increasing the number of points it takes to get a reward.” That is the opposite of the usual rewards-programme reflex, which is to hand out more points, call it a relaunch, and watch redemption inflate the unit economics.
The perk list is less interesting than the design choice behind it: free guacamole and chips on signup, lower redemption thresholds, additional birthday rewards the member picks, monthly free-food offers, and a Rewards Exchange with more options. Each one of those is a specific lever to make the programme worth joining at the counter, where the barrier was never price, it was friction.
Then the Harder Fix: Make the Programme Invisible at the Till
The second move is the one I actually think will matter more over a five-year horizon. On the Q2 2026 earnings call Boatwright previewed an August pilot of what the company calls frictionless rewards: customers who pay with Apple Pay in the restaurant earn points automatically, no scanning, no app open, no card. He described it, per the Restaurant Technology News account, as “an important step to not only creating a more seamless in-restaurant loyalty experience.”
By the Q3 2026 call on 28 October 2026, Boatwright was able to say that nearly one-third of Chipotle’s in-store customers already pay with Apple Pay, and that the frictionless rewards technology is expected to roll out across the chain in 2027. His framing of why that mattered was characteristically understated: “It’s a simple change, but an important one because it reduces a key friction point while also supporting throughput.”
Translate that for a buyer and it reads differently. If the rewards programme already captures roughly 20% of in-restaurant transactions and one-third of in-restaurant customers pay with a method the brand can passively identify, there is a path to lifting the in-store attach rate by an order of magnitude without changing a single customer behaviour. That is a different kind of loyalty investment than giving away more free food.
What the Programme Is Actually Producing
Boatwright told the Q3 2026 call that active membership had grown to more than 24 million, up from 23 million at the Q2 call and 21 million at the time of the April relaunch. Members drove one-third of total sales in the quarter, the highest loyalty share the company has reported. Engagement metrics from the Summer of Extras programme, which ran June through August, included 600,000 challenges completed and 25 million badges earned.
For context on the gamified side, the 2025 Summer of Extras drove roughly 6.4 million activations and an estimated $12 million in incremental sales, per Chipotle’s own figures reported by Marketing Dive. CFO Adam Rymer summarised the thesis in one line on the same call: “deeper engagement builds loyalty and drives comps.” The numbers for 2026 are not fully disaggregated yet, but the share-of-sales move from 30% to 33% in nine months is the kind of shift a loyalty team can take to a CFO.
What Is Still Unsolved, Because Every Good Case Has One
Two things, honestly. The first is the in-restaurant attach rate itself. Chipotle did not update the 20% figure on the Q3 call, only noting that in-store enrolments remain elevated. If the frictionless rewards pilot was meant to crack that number, we do not yet have a reported after-state. The chainwide rollout is scheduled for 2027, and until the follow-up quarterly data lands, the lift is a promise, not a result.
The second is the definition of “active member.” The Restaurant Technology News piece flagged, politely, that Chipotle does not define active member the same way some competitors define 90-day active users. A 24 million figure sounds different against a 90-day cohort than against a twelve-month one. This is a known soft spot in loyalty reporting across the industry, not a Chipotle-specific sin, but it is the figure a sceptical buyer should press on first.
What a WooCommerce or VIP Operator Can Build This Quarter
Here is where I earn the invoice. A mid-market retailer with a strong online store and a physical point of sale, whether that is a shop, a counter, or a click-and-collect lane, can translate the Chipotle move into a specific build. None of this requires a 3,800-unit chain. The bones of what makes it work are a well-structured customer record, a payment channel the brand can read, and a rewards economics model that does not inflate itself on relaunch. We have shipped versions of each.
The build has six moving parts, sized for a team of two engineers plus a product owner on a WooCommerce or WordPress VIP estate running a real retail or food operation.
One, the unified customer record. A single WooCommerce customer object has to carry both the online order history and the in-store identifier. This is a mu-plugin plus a small schema addition in wp_usermeta: a stable external customer id (not an email, because customers change those), plus a map of payment-instrument tokens the brand is allowed to recognise. Two to three sprints, including the privacy review. Without this layer the rest of the stack is a coupon engine, not a programme.
Two, the in-store identification path. If the brand has a POS vendor that exposes a tokenised payment identifier, Apple Pay included, the integration is a webhook from the POS to a WooCommerce REST endpoint that looks the token up against the customer record and awards points. Where the POS does not expose it, the fallback is a QR flow at the till that resolves to the same endpoint. Two to four sprints, and almost all of the risk is in the POS vendor’s willingness to cooperate, not the WordPress side.
Three, the rewards economics model. Store the points economy as data, not as business logic scattered across plugins. A custom post type for reward definitions with a fields-based structure (point cost, perk value in currency, expiry rule, channel eligibility) makes it possible to change thresholds without a release, and to answer the question “what did we give away last quarter” in one query. One sprint of modelling plus one of admin UI. The alternative is the sprawl that kills most rewards plugins within eighteen months.
Four, the gamification surface. Challenges, badges and streaks are a template of their own: a Gutenberg block pattern that renders against the customer’s reward state and nothing else, so the same template drops onto the account page, the thank-you page and the app. The 2025 Summer of Extras produced 38 million earned social impressions for Chipotle on the strength of this surface being shareable. Make it shareable on day one, not on day one hundred. One sprint for the template, plus whatever it takes to agree on the first three challenge types with marketing.
Five, the honest measurement baseline. Before any of this ships, instrument two numbers and publish them internally weekly: in-store transactions attached to a member, and in-store transactions attached to a payment-instrument token that resolves to a member. Separate those from digital, cleanly. Two weeks of instrumentation plus sixty days of baseline reading. Without this the relaunch has no before-state to compare against, which is the mistake almost every loyalty refresh makes.
Six, the governance for not inflating the economy. Write down the rule Garner stated out loud: no more points per reward on relaunch. Put it in the project tracker, name it, make the product owner approve any exception. Half a day of writing, worth every minute the first time a marketing lead proposes a doubling in Q4. This is the single piece of the build with the shortest timeline and the longest half-life.
The total effort is roughly eight to twelve weeks of concentrated work plus a measurement quarter. That is a fraction of a Rewards on Repeat budget, and it buys a loyalty surface that can tell you honestly whether the next relaunch worked.
Why the Shape of This Case Matters More Than Its Figures
I started this piece with the complaint that loyalty programmes drive ninety percent of app orders and a fifth of counter orders. The Chipotle case is the one on the public record right now that treats that gap as the actual problem and spends money fixing it, rather than pointing at the ninety and calling the programme a success. Everything else in Rewards on Repeat, the extra birthday rewards and the lower redemption thresholds and the Rewards Exchange, is in service of a channel where the customer did not have to be a member in the first place.
If you are running a mid-market retail or food operation on WooCommerce or WordPress VIP and your rewards programme reports mostly digital share, the question to put to the next planning meeting is the one Chipotle put to its own: what is the in-store number, who owns it, and what is the smallest build that would move it next quarter. If you would like us to size that build against your estate honestly, including the awkward estimate for the POS integration, that is the conversation we are set up to have. Our WordPress VIP agency page is the route in; the first call is free and we will say so if the answer is “do not build this yet.”
Last modified: November 17, 2026