Opening night in Hangzhou looked like a product launch and a store opening at the same time. Bright TV panels covered the facade. A crowd filled the sidewalk. Inside, customers collected orders at a pickup counter beside a kitchen the size of a closet. A pizza order started with a WeChat QR code on the wall or table. That scan was the product.
I led development of the mini-app behind that scan. Our translated Russian layout failed with local testers. Five iterations later, the ordering flow matched the habits Hangzhou customers already brought into the store. Broader China GTM lives on China market entry.
China’s licensing stack added ICP, PLF, government review, and the blue-poster filing step to the same launch. That bureaucracy has its own story. This case follows the customer screens and the store floor.
Related field noteICP, PLF, and the 2×2-Metre Blue PosterOur first China web filing took four months. The process moved through Chinese paperwork, Beijing corrections, a Shanghai trip, and the required blue-poster photo.Read the filing story →
Why orders moved into WeChat
Dodo’s first China years used a familiar Western model: a secondary location, delivery-heavy operations, a company website, and company couriers. The website produced single-digit order volume. Chinese customers already lived inside WeChat and Alipay, while Meituan and Ele.me owned food delivery traffic.
The Hangzhou pivot put a compact “Pizzeria of the Future” on a high-traffic line. The kitchen shrank to roughly seven or eight square metres. Pre-baked rectangular pinza bases brought cooking time down to a few minutes. Customer ordering moved into a WeChat Mini-Program, with the counter focused on pickup.
Every scan needed to reach payment faster than a conventional cashier conversation. That became the bar for the interface.
Launch night was part of the UX
I have photos from that night: the first storefront under large LED panels, the launch crowd, and the team placing QR codes on the facade. On the floor I wore a gray cap and green sneakers, handing out tasting slices so guests could try the food before scanning a foreign brand’s code.
QR codes on the facade
The team placing WeChat self-service ordering codes before customers arrived.
The storefront served as onboarding. QR placement, scan reliability, menu speed, and the pickup screen formed one customer journey. My lead-developer role continued from the merged pull request onto that launch floor.
What the QR actually opened
We placed codes around the pizzeria. A WeChat scan opened the menu, promotions, and coupons. Every product had a detail view with composition, price, and photos large enough to support a decision directly from the screen.
Promotions shape Chinese QSR ordering habits. We gave campaigns their own page, combining general deals with coupons customers could save to WeChat Wallet and apply with one tap. That page sat inside the primary ordering flow.
WeChat Pay handled the standard payment path directly inside the mini-app. Staff used a store phone for tourists and elderly guests who needed help with WeChat.
After payment, status lived on the phone and the in-store TV. When an order reached pickup, the panel flashed and a voice announced it. About fifteen minutes later, an approved subscribe message could ask for a short review.
WeChat subscribe-message permissions
WeChat Mini-Programs use subscribe messages for order updates. Each message type requires a customer permission, which makes permission timing part of checkout UX.
Checkout asked for the permissions tied to pickup status, cancellation, and the later review prompt. A permanent opt-out closed future prompts for that customer.
The system modal shaped the order summary and the timing of each request. A poorly timed prompt spends customer attention and weakens the operational message that follows.
Five UI iterations
The fifth design reached launch.
Iteration 1: translate the Russian app. We kept the original structure and replaced the copy. Local feedback rejected the visual hierarchy and ordering flow.
Iteration 2: feature pile. Profiles, gifting, ratings, sauce automation, a review camera, circular tracking, and size toggles filled the roadmap. At the counter, those choices slowed checkout. Every extra tap created another guest waiting behind the current order.
Iteration 3: cut hard. A floating cart, larger product cards, and cleaner photos shortened the path. Rectangular pinzas replaced round pizzas while R&D changed the food format, giving the interface its own product shape.
Iteration 4: Hangzhou custdev. The team tested the flow with people near the store. We put a phone in each participant’s hand and watched the checkout. Store managers translated the Chinese-language sessions. Five minutes of live use exposed friction that passed every Figma review.
Iteration 5: food photography as product. Chinese QSR mini-apps treat photos as the main CTA. We studied Luckin, HeyTea, McDonald’s, Pizza Hut, KFC, and other local references. Their shared pattern combined category navigation, large food images, a short payment path, and wallet-aware promotions. Our launch version used vertical categories, pinza cards, orange add actions, and clear prices.
The rule that survived all five: build around the experience customers already know inside WeChat. A translated home-market interface carries the wrong assumptions into every screen.
The same cashless format later moved into a second Hangzhou location.
Second Hangzhou pizzeria
The cashless format at its second Hangzhou location.
Beacons inside the mall
Indoor GPS drifts across floors in a shopping mall. Bluetooth and Wi-Fi beacons near the unit gave the mini-app the local signal it needed to select the correct store menu and promotions.
That infrastructure choice protected the customer experience. Correct store context meant correct coupons, correct prices, and a clean path to the first order.
What we can stand behind
On launch day, customers scanned the codes and ordered through the mini-app. The cashierless store processed real orders through the journey we had tested.
Corporate China strategy later went its own way. The UX lesson remains: local research beats a translated interface, and a focused checkout beats a feature-heavy roadmap. Wallet-aware promotions and WeChat’s permission model define the product.
The companion piece, What founder-led development actually means, explains why launch-floor ownership matters.