tuxedo_26q3_home_outcome-led| Arm | Visitors | consult booked | Rate |
|---|---|---|---|
| Control | 4,523 | 124 | 2.74% |
| Variant Better | 4,428 | 157 | 3.55% |
Call: Planned sample reached. 99% chance the variant beats control (threshold 90%).
Computed by Website OS from raw counts. PostHog's own significance flag is stored but never used for the call.
| Arm | Leads | Qualified | Qualified share | Qualified per 1,000 visitors |
|---|---|---|---|---|
| Control Better | 119 | 80 | 67% | 17.7 |
| Variant | 151 | 45 | 30% | 10.2 |
Quality gate: More form fills, but fewer qualified leads per visitor: 10.2 vs 17.7 per 1,000. Shipping needs an override. Under 1% chance the variant brings more qualified leads per visitor.
Qualified uses the person's label where there is one, else the AI suggestion. Spam and duplicates are left out.
Headline changes from 'Windows and doors installed right' to 'Cut your heating bill with windows built for Calgary winters'.
Rewrite the main headline around lower heating bills
Replace a service-name headline with the outcome the customer wants. On Tuxedo Windows & Doors, this should lift consult booked.
The outcome framing matched the search intent of paid traffic.
Schematic preview. Real desktop and mobile screenshots attach when the variant is built.
No evidence attached.
Replays are stored as session IDs and timestamps only. Recordings stay in PostHog.
No comments yet.