Website Growth Advisor — limits and best practices

Best practices

• Start with one primary objective.
• Use a period that matches your business cycle.
• Enable only sources that answer the question.
• Pick a primary KPI your team actually follows.
• Use medium or high evidence for important decisions.
• Keep hypotheses visible and verify them before publishing a change.
• Use targeted pages instead of a broad scope for your first analysis.
• Start manually before scheduling recurring analyses.

A safe review sequence

1. Read the summary.
2. Check source status.
3. Separate facts, correlations and hypotheses.
4. Select one or two priorities.
5. Check the relevant page or journey.
6. Launch a measurable, reversible change.

Avoid these mistakes

Comparing unlike periods

A comparison is misleading when one period contains an exceptional campaign or a different season. Use comparison only when the context is similar.

Treating a missing source as zero activity

An unavailable, empty or partial source does not prove that there were no visitors or conversations. Always read its status.

Publishing without review

Website Growth Advisor suggests options. The person responsible for content, marketing or the website should approve the location, change and measurement.

Scheduling too early

A daily or weekly run does not replace a stable configuration. Test manually first.

When to run it again

Run another analysis after a major website change, a campaign, a form or CTA change, an offer change, a new source connection or a comparable reporting period.

Confirmed limits

• The agent explains and recommends; it does not change the website automatically.
• An unavailable source is not evidence.
• A correlation does not prove a cause by itself.
• Results depend on the period, sources and settings selected.

Information to complete

The project does not define a universal threshold for applying a recommendation. Use your business context, the displayed evidence and your approval process.