5 Metrics Every AI Lab Dashboard Should Track
Before you ship a fancy analytics suite, capture the five signals that actually guide product decisions: runs, share rate, credit burn, tool mix, and creator lift.
1. Runs per device
Pipe `/api/generate` logs into a database with `client_id`, type, and timestamp. A simple time-series chart shows whether daily active labs are growing. We aggregate by UTC day to match the credit reset.
2. Share completion rate
Every share modal call hits `/api/share`. Log the platform (TikTok, Instagram, etc.) and added credits. This lets you see which labs inspire the most shares and whether copy tweaks improved the conversion.
3. Credit burn and refunds
Track how many credits get spent versus refunded due to rate limits. Large spikes in refunds usually mean the anti-bot challenge fires too early. We visualize this by overlaying spend and refund bars.
4. Tool mix
What percentage of runs come from Toxic, Crush Match, Villain Story, etc.? This informs what to highlight on the homepage and which labs deserve new prompt pools.
5. Creator lift
When running collaborations (see our collab guide) assign unique share IDs so you can report runs and credit purchases driven by each partner.
Further reading