Chasing Core Web Vitals for Real Users, Not Lighthouse Scores
A perfect Lighthouse score on a fast office connection tells you almost nothing about the experience on a mid-range phone on real-world mobile data.
Tanjil Ahmed
Lead Software Engineer · Notionhive
Lighthouse is a lab tool, run on a clean simulated connection, on hardware nothing like most of your actual users' devices. Optimizing purely for the number it reports is a common and completely understandable mistake — it's also how teams end up with a 98 score and real users still complaining about a sluggish site.
Field data beats lab data every time
The Chrome UX Report and your own Real User Monitoring tell you what actually happened on actual devices on actual networks. That's the number that correlates with conversion and bounce rate — not the lab score. I treat Lighthouse as a pre-flight check and RUM as the actual scoreboard.
- Test on a throttled mid-range Android device, not your development laptop, before calling anything fast.
- LCP is usually an image or font-loading problem; INP is usually a JavaScript execution problem — diagnose accordingly.
- Ship RUM (web-vitals library reporting to your analytics) from day one, not after a complaint.
- A 'fast enough in the office' site can be genuinely broken for the actual customer base — check the field data.
Your Lighthouse score in the office means nothing. Your Time-to-Interactive on a customer's actual phone is the real product.