Why you need to know your site’s performance plateau (and how to find it) [PerformanceObserver 2026]

Have you ever asked yourself these questions?

"I made my pages faster, but my business and user engagement metrics didn't change. WHY???"
"How do I know how fast my site should be?"
"How can I demonstrate the business value of page speed to people in my organization?"
The answers might lie with identifying and understanding the performance plateau for your site.

Transcript

1. The Performance Plateau Why it matters + how to find it + what do with it
2. Tammy Everts CXO, SpeedCurve Sr. Director of Community, Embrace
3. Recurring Questions…
4. Recurring Questions… “I made my pages faster, but my business and user engagement metrics didn't change. WHY???”
5. Recurring Questions… “I made my pages faster, but my business and user engagement metrics didn't change. WHY???” “How fast should my pages be?”
6. Recurring Questions… “I made my pages faster, but my business and user engagement metrics didn't change. WHY???” “How fast should my pages be?” “How can I demonstrate the business value of page speed to people in my organization?”
7. The Performance Plateau Can Help Answer All Those Questions
8. What Is the Performance Plateau? The point where improvements in performance metrics no longer translate into measurable business gains
9. Correlation Charts: A Short History speedcurve.com/blog/site-speed-business-correlation/
10. Correlation Charts: A Short History speedcurve.com/blog/site-speed-business-correlation/
11. We Can Test Different KPIs speedcurve.com/blog/site-speed-business-correlation/
12. We Can Validate Different Metrics speedcurve.com/blog/long-animation-frames-user-behavior/
13. But Wait… ?
14. Why do fast pages “fail” KPIs? • Session bucket size is smaller and statistically insignificant • Unidentified bots/scrapers • Pages left open in tabs (session times out, then reloads) • Error pages
15. But Wait Again… ?
16. But Wait Again… performance plateau
17. We Started to Wonder… • Does every site have a performance plateau? • How similar is the performance plateau across different sites? • Where do Google’s thresholds for Core Web Vitals land on the plateau?
18. So We Investigated • 4 sites with significant traffic • 30 days of RUM data • Correlated Start Render and Largest Contentful Paint to bounce rate speedcurve.com/blog/web-performance-plateau/
19. Site A
20. Site B
21. Site C
22. Site D
23. Key Observations
24. Key Observations ✔ Every site was different speedcurve.com/blog/web-performance-plateau/
25. Key Observations ✔ Every site was different ✔ A clear plateau emerged for every site – for both Start Render and LCP speedcurve.com/blog/web-performance-plateau/
26. Key Observations ✔ Every site was different ✔ A clear plateau emerged for every site – for both Start Render and LCP ✔ Plateaus can start quickly (e.g., 400 ms) speedcurve.com/blog/web-performance-plateau/
27. Key Observations ✔ Every site was different ✔ A clear plateau emerged for every site – for both Start Render and LCP ✔ Plateaus can start quickly (e.g., 400 ms) ✔ Differences between optimal bounce rate and plateau points can be huge speedcurve.com/blog/web-performance-plateau/
28. Key Observations ✔ Every site was different ✔ A clear plateau emerged for every site – for both Start Render and LCP ✔ Plateaus can start quickly (e.g., 400 ms) ✔ Differences between optimal bounce rate and plateau points can be huge ✔ Median performance often lies within the plateau speedcurve.com/blog/web-performance-plateau/
29. Key Observations ✔ Every site was different ✔ A clear plateau emerged for every site – for both Start Render and LCP ✔ Plateaus can start quickly (e.g., 400 ms) ✔ Differences between optimal bounce rate and plateau points can be huge ✔ Median performance often lies within the plateau ✔ Google’s thresholds are too optimistic speedcurve.com/blog/web-performance-plateau/
30. How to Find Your Plateau
31. How to Find Your Plateau 1. Pick performance + business metrics (e.g., LCP vs bounce rate) You may need to experiment with different combinations
32. How to Find Your Plateau 1. Pick performance + business metrics (e.g., LCP vs bounce rate) You may need to experiment with different combinations 2. Gather significant RUM data 5-10M pageviews is typically a good bucket size
33. How to Find Your Plateau 1. Pick performance + business metrics (e.g., LCP vs bounce rate) You may need to experiment with different combinations 2. Gather significant RUM data 5-10M pageviews is typically a good bucket size 3. Create correlation charts Other RUM tools are starting to include this feature
34. What to Do with Your Findings
35. What to Do with Your Findings ✔ Share with stakeholders – help them see opportunity!
36. What to Do with Your Findings ✔ Share with stakeholders – help them see opportunity! ✔ Explain why improvements haven’t moved metrics
37. What to Do with Your Findings ✔ Share with stakeholders – help them see opportunity! ✔ Explain why improvements haven’t moved metrics ✔ Reset performance goals – think beyond Core Web Vitals!
38. Goal Setting
39. Baselines & Tracking Over Time
40. Baselines & Tracking Over Time ✔ Repeat plateau analysis monthly or after major changes
41. Baselines & Tracking Over Time ✔ Repeat plateau analysis monthly or after major changes ✔ Compare distributions over time
42. Baselines & Tracking Over Time ✔ Repeat plateau analysis monthly or after major changes ✔ Compare distributions over time ✔ Update performance goals as needed
43. Takeaways ✔ You need to connect performance to business
44. Takeaways ✔ You need to connect performance to business ✔ Fast isn’t fast enough unless it affects user behaviour
45. Takeaways ✔ You need to connect performance to business ✔ Fast isn’t fast enough unless it affects user behaviour ✔ Real user data is essential for guiding decisions
46. Let’s talk about your performance plateau! tammy@embrace.io tammyeverts tammyeverts.com