Quick answer
A useful performance baseline records what was tested and under which conditions. It helps separate server delays, large assets, slow scripts, and interaction problems instead of treating one score as the entire diagnosis.
Before you begin
Choose representative URLs and important visitor tasks. Include a landing page and a page with the features people use, such as search, a product selector, or a form.
What to do
- Record device type, browser, connection conditions, test location, and whether caches were warm or cold. Use comparable conditions for later measurements.
- Run several controlled tests and retain the reports. Note large variations rather than selecting only the fastest result.
- Review available real-visitor data separately from laboratory tests. Field reports may summarize a wider period or similar pages rather than the exact current release.
- Inspect the actual loading and interaction behavior to identify likely causes. A low score does not by itself identify the correct fix.
- Prioritize changes with a clear expected benefit and retest the same journeys after implementation. Keep usability and functionality in the acceptance checks.
How to check the result
Compare like-for-like reports and describe both improvement and limitations. A single faster run is weaker evidence than repeatable results supported by an identified cause.
If you need help
Send report links, URLs, dates, and test settings when requesting help. Explain what visitors experience, such as slow first content or a delayed button response. Do not promise a fixed score before the underlying site and third-party dependencies are reviewed.
Related help
- Understand LCP, INP, and CLS in a website review
- Improve a slow-loading homepage hero image
- Find which third-party tools slow a website
Official references
Reviewed October 31, 2026. Interfaces, vendor availability, and account features can change. Check your current account before making changes.
