What response time measures
Time to first byte depends on several steps: network activity, request processing, document generation and possible redirects. It does not represent the time needed to display the whole page.
Cache what can be cached
Versioned static resources can be kept longer by the browser. Dynamic pages need suitable rules. An account page or cart must never accidentally be shared between visitors. Check the rules against how your application works.
Reduce unnecessary redirects
Each redirect can add a step before the correct page loads. Use internal links pointing directly to the final address, and keep HTTP, HTTPS and your primary domain configuration consistent.
Observe before migrating
An extension, database query or external API can slow down the server. Measure processing and inspect available logs before changing plans. More powerful hosting does not necessarily fix an application problem.
PUT IT INTO PRACTICE
Questions for your technical team
- Is the document generated for every request?
- Which pages can be cached without mixing user data?
- Are there multiple consecutive redirects?
- Is an external call slowing down processing?
- Does response time vary by time of day or load?
Bring several reports and their timestamps to your hosting provider. They can help relate observations to logs and server load. Browser metrics alone cannot identify a specific SQL query or determine the right hosting capacity.
Sources and scope
This guide suggests a working method. No numerical improvement is guaranteed. To check definitions and explore the audits: official documentation.