Methodology: This check follows wSEO methodology 1.1. How the assessment works
A/B Test Setup
Test a disputed change as a hypothesis instead of presenting design preference as fact.
- Hypothesis and success criterion
- Control and changed variant
- Allocation, events, and quality checks
What to prepare
- the canonical URL and exact test journey;
- dated baseline evidence;
- a list of implemented changes;
- the same device, browser, region, and account where these affect the result.
Step-by-step verification
- Each user consistently receives one variant
- The primary metric is fixed before launch
- The test does not break indexing or the key journey
Repeat the check at least twice in a fresh window. For non-deterministic AI answers or experiments, use multiple repetitions and do not draw a conclusion from one response.
Acceptance criterion
One defined hypothesis, control and variant, primary metric, technical QA, and analysis plan.
Save the URL, date, screenshot, HTTP response, or analytics event—the evidence type that directly proves the promised outcome.
Does not guarantee a positive result. Duration depends on traffic and required sensitivity.
If verification fails
Do not hide the symptom with additional changes. Return to the baseline, check caching and configuration propagation, then compare the actual result with the agreed scope.
