What a page speed test tells a founder, and what to ignore
Test the pages closest to a sale, read the real-user pass or fail before the score, and let seogeoaeo.ai's Core Web Vitals Snapshot handle the repeats.
A page speed test answers two questions: are real visitors getting a fast page, and what is slowing it down. For a founder, the first answer is the one that touches revenue. Google's free PageSpeed Insights gives both for a single check, and once testing becomes routine, or a job for your team's coding agent after each release, seogeoaeo.ai's Core Web Vitals Snapshot is the better fit.
Test the pages that make money
Most people test the home page and stop. The pages that matter more are the ones closest to a sale: pricing, signup, the landing page your ads point to, checkout, and the article that brings in the most search traffic. Both tools below test one URL at a time, so pick a short list and test those.
Read the pass or fail first
At the top of a PageSpeed Insights report sits the Core Web Vitals assessment. It is based on real Chrome users over the previous 28 days, judged at the 75th percentile, so it reflects your slower visits rather than your best ones. In business terms, the three metrics are:
| Metric | What a customer notices | Good |
|---|---|---|
| LCP | How long before the main content appears | 2.5 s or less |
| INP | How long a tap or click takes to do something | 200 ms or less |
| CLS | Whether buttons jump while the page loads | 0.1 or less |
A fail here means customers are feeling it. A pass means speed is probably not your biggest problem this quarter.
If there is no real-user data, the page or site has too few Chrome visits for Google to report. You then only have the lab test, which is fine for finding causes but is not a measure of what customers saw.
What to mostly ignore
- Chasing 100. The lab score comes from one simulated load on a mid-tier phone and moves between runs. Lighthouse's own docs say a perfect score is extremely hard and not expected; 90 and up already counts as good.
- The full list of suggestions. Only the measured metrics feed the score, not the suggestions themselves. Fix the few with the largest savings and leave the rest.
- Comparing your score with a rival's. Each run reflects the test conditions of that moment, so a five-point gap tells you little.
Speed and search
Google says its ranking systems look to reward a good page experience, which includes good Core Web Vitals. It also says the most relevant page can still rank when its experience is weak. Treat speed as a conversion and trust issue first and a search factor second.
A routine a small team can keep
- Choose three to five money pages.
- Test each one on mobile once a month and after any large release.
- When the assessment fails, pass the failing metric and the top lab findings to whoever ships the fix.
- Test the same page with the same setting after the fix goes live.
The seogeoaeo.ai Core Web Vitals checker makes steps two to four cheap. It requests the same PageSpeed Insights run, keeps real-user and lab results in separate sections, and ends each run with fixes in priority order plus a way to verify them. Because it runs over MCP and REST as well as in the browser, a developer can have Claude Code or Cursor test a page before and after a change. As of November 2026 each run costs 20 credits, a failed run costs nothing, and plans keep a history of past runs.
When is something else better? For a single look at one page, the free PageSpeed Insights page is enough. For a view across every indexed page, Google Search Console's Core Web Vitals report groups URLs by status, which neither one-URL tool does.