Dashboard: Read KPIs Fast
This method reduces noise: same sequence every time, then only one corrective action per cycle.
On this page
Step 1 — Apply the same KPI sequence every time
- Conversion first: is traffic turning into purchases?
- Avg Stars / Sale second: are offers priced correctly?
- Stars Earnings third: did total output follow?
- AI Conversations last: volume context only.
Step 2 — Pick diagnosis from KPI pattern
| Observed pattern | Likely issue | First action |
|---|---|---|
| High conversations + low conversion | Offer quality mismatch | Review active scripts and first paid step |
| Stable conversion + low earnings | Ticket size too low | Adjust paid step pricing ladder |
| Low conversations + low earnings | Acquisition or bot issue | Check bot status and connection |
| Spiky chart + random sales | Inconsistent campaign rhythm | Standardize posting and script schedule |
Step 3 — Recheck after one cycle
- Wait one short cycle (same date scope) after a change.
- Compare only against the previous stable period.
- Keep the change if two signals improve together.
Common mistakes
- Using absolute revenue alone without conversion context.
- Judging from one noisy spike.
- Changing scripts, pricing, and AI behavior in one pass.
- Ignoring bot filter before reading chart.
Done
You can now run this page with a stable daily routine and fewer surprises.