This is a composite scenario built from how we deliver this kind of work. The client is fictional and the figures are targets, not measured results. We will replace it with a named engagement when a client agrees to be quoted.
A two-person studio in Dhaka with a working puzzle prototype, no backend and savings for about four more months
- Playtesters liked the puzzle loop, but the prototype ran at 18 frames per second on a mid-range Android phone, crashed on resume, and had no analytics, no cloud save and no way to run events.
- The two founders had four months of savings left. They needed a co-development partner, not a vendor who would take the project away.
How we approached it.
Two-week review
We profiled the build, listed every frame-time cost and agreed a scope the founders could afford: performance, feel, backend, soft launch.
Performance pass
Object pooling, a single sprite atlas per scene, batched draw calls and a fix for the resume crash. The target was 55 frames per second on the founders' test phone.
Game-feel pass
Squash and stretch on tiles, hit-stop on a clear, particles and sound on combos, tuned in weekly playtests.
Backend and analytics
Cloud save, remote config for level tuning and events, and analytics for day-1 and day-7 retention and level drop-off.
Soft launch and iterate
Two countries, three builds, each changing what the drop-off data pointed at. Then the store launch, with the founders doing the releases.
Try it yourself.
A simplified illustration with made-up data. It plays once when it scrolls into view, and you can pause it or step through.
Hover or focus a point to read its value.
Step 1 / 4 Axes for day-1, day-7 and day-30 retention.
What we built.
- Performance fixes: pooling, atlasing, batching, resume-crash fix
- Game-feel layer: squash, hit-stop, particles, audio
- Backend: cloud save, remote config, events
- Analytics with retention and level drop-off dashboards
- Release pipeline the founders run themselves
Outcomes (targets).
These are the results we would aim for. They are not measured results from a named client.
- Target18 to 58 fpson a mid-range Android phonefounders' own test device
- Target31% to 42%day-1 retention across soft-launch buildsthree builds, two countries
- Target99.4%crash-free sessions at launchfrom the analytics
- Target14 weeksreview to store launchfounders own the repository
Timeline and stack.
14 weeks, including a six-week soft launch
What we learned.
- Performance is game feel. Nothing feels good at 18 frames per second.
- Soft-launch data beats everyone's opinion about which level is too hard.
- Co-development means the studio ships the next release without you.
How we would deliver it.
Want this
for your business?
Message an engineer on WhatsApp or send a brief. We agree scope and price before work begins and build in small working steps you can test.