Every card carries two labels: priority (how much it costs you) and confidence (how sure Smart Insights is). Read them together.
Priority #
| Label | Score | Extra conditions |
|---|---|---|
| Critical | 80–100 | At least 100 visits measured, High confidence, and at least 25 lost visits or 5% of your site’s visits |
| High | 60–79 | A card that fails the Critical conditions becomes High |
| Medium | 40–59 | |
| Low | 0–39 |
The score blends the visits lost relative to the other cards of the batch with the visits lost relative to your whole traffic, so a busy site gets a short ordered list instead of dozens of “high” items. Cards with the same numbers get the same label.
Confidence #
| Visits measured | Maximum confidence |
|---|---|
| Under 30 | 35 (Low) |
| 30 – 99 | 60 (Medium) |
| 100 – 299 | 80 (High) |
| 300 or more | 95 (High) |
The statistical margin of the measured rate is subtracted, so a borderline rate lowers confidence. A problem already seen in the previous period adds a small bonus. Labels: 70+ High, 40–69 Medium, under 40 Low.
Badges you will see #
- Provisional · N visits – built on few visits: watch it, do not act on it alone.
- Observation – too little traffic to size the impact; confirm before acting.
- WORSE / BETTER – shown only for a real change against the previous period.
How to use them #
| High confidence | Low confidence | |
|---|---|---|
| High priority | Act now | Check the evidence first |
| Low priority | Plan it | Ignore for now |
Expected result #
You can explain why a card is Critical or not, and how much to trust it.
FAQ #
Why is a big problem only High? #
It did not meet one Critical condition, usually the number of visits measured.
Why did priorities change after the update? #
1.9.2 made scoring stricter and more honest; old cards the rules no longer confirm were closed.
