01 2024–now
Rewriting a customer ordering platform
From nothing to launch, including the product and client-facing role.
- Client
- One of Speedlo’s largest customers, a Czech fast-food chain
- Period
- 2024–now
- Role
- Lead frontend developer → Product owner
- Stack
- Next.jsReactTypeScriptGraphQL + ApolloTailwind CSS
The client is kept anonymous and unbranded. Only ratio metrics are published here, with no revenue figures.
Context
The original site was built and maintained by a sister company that does not operate in hospitality and was not delivering what the client actually needed.
The site had to be rewritten into our stack and connected to our backend. Otherwise the product could not be developed any further.
Problem
The challenge was scope, not complexity: a large number of features had to ship at once.
And this was the storefront of an operating restaurant chain. Launch day could not mean a drop in sales or confused regulars.
Decisions
- 01
Rewrite rather than patch
Connecting to our own backend and our own stack was the precondition for the product going anywhere. Patching someone else’s solution would not have moved that ceiling.
- 02
Mobile-first as a requirement, not a bonus
85% of visitors arrive on mobile. Design and testing therefore started on mobile rather than desktop.
- 03
Build for extension
It is not just an e-shop, because it also carries a credit system. The rewrite finally made it extensible with the client’s own ideas: express delivery, a loyalty scheme with rewards.
- 04
Fix how the collaboration runs
I introduced regular meetings and a communication rhythm with the client so the product actually moved instead of being discussed.
- 05
Ratings as a business feature
I reworked the rating template: five stars redirect to the branch’s public Google review, fewer stars to an internal form. The goal was to raise the number of positive public branch reviews. Leadership called it a feature with a fundamental impact on the client’s business.
My role
- I was on the frontend from the start through to launch, and implemented a large part of the application myself.
- In the final phase I took over as Product Owner while the PO was away, and ran the last stretch before launch directly with the client.
- The client stayed with me afterwards, and we go on developing the site together. What began as covering a holiday became a permanent role.
- I ran the backlog with type, status, teams, priority and expected impact, including the things we rejected.
Result
- ~1,100 active users on day one according to Google Analytics. In reality there were more, as roughly a fifth of traffic never reaches GA.
- 50% of all visitors completed a purchase.
- 82% of visitors who reached the order form went on to complete it. That is comparable with the large e-shops we used as a benchmark.
- On launch day sales performance held at the level of the previous site. Regulars were not thrown by the change.
- No errors surfaced after launch.