The platform · in alpha
appanalytics.ai —
analytics that think.
Why a games studio built this
Nine years of live ops is a very long user test.
Operating a real-time PvP game teaches you that the interesting questions are never the ones you planned a dashboard for. Why did retention move on Thursday for players who joined a clan in their first session? Which of the eleven things we shipped last month is responsible?
Every one of those questions used to cost a week: define the events, write the query, wait, discover you asked it slightly wrong, repeat. By the time the answer landed, the cohort it described had already churned or converted.
So we stopped writing queries and started writing agents — systems that watch the event stream continuously, form their own hypotheses about how people move through the product, and surface what changed without being asked. That became appanalytics.ai.
Built on our own traffic
- Six live titles as the first production workload
- Real-time PvP events, not synthetic test data
- Unity and native iOS SDKs, written for our own builds
- Streaming architecture — events analysed as they arrive
- SOC 2 Type II, end-to-end encryption, granular access
How it works
Zero to insight in three moves.
01
Connect
A lightweight SDK — Unity or native iOS — or a direct connection to an existing warehouse. Nothing to model up front.
02
Agents analyse
Multi-agent systems read the raw event stream, reconstruct journeys, discover the funnels that matter and learn what normal looks like.
03
Insights arrive
Findings are pushed to Slack, email or the dashboard as they happen, in plain language, with the drop-off already attributed.
The difference
Agents that work, versus dashboards that wait.
| Feature | appanalytics.ai | Traditional tools |
|---|---|---|
| Setup time | Under 10 minutes | Days to weeks |
| Funnel creation | Discovered automatically | Manual SQL and config |
| Insight discovery | Agents come to you | You go looking |
| Query speed | Milliseconds | Seconds to minutes |
| Data team required | No | Usually |
| Monitoring | Built-in, continuous | Custom alert rules |
Who it is for
One source of truth.
Product
Journey analysis that shows where a feature actually loses people, not where you assumed it would.
Growth
Retention and monetisation levers surfaced automatically, with predictive signals on cohorts still forming.
Engineering
Behaviour tied to performance, so a regression shows up as a drop before it shows up as a review.
Marketing
Attribution from campaign to conversion, with the channels that pay for themselves ranked honestly.
Early access is open and free.
No credit card. The product site handles the waitlist and the live alpha.
Publishing, partnerships,
or a very specific data question.
We read everything that arrives at the studio inbox — user acquisition, publishing, licensing our question bank, or getting early access to the analytics platform.