Why Your Site Looks Fast To You And Slow To Everyone Else

Two Subjects, Same Pattern

Both audited in the same week. Both small CT brands. Both have a press story working in their favor — one was featured in a national magazine, the other is a marketplace running ticketed events. Both have repeat-customer bases.

Subject A: Boutique apparel, Shopify, on Cloudflare. Mobile Lighthouse Performance 28/100. Lab LCP 12.6 seconds. Field LCP, from real Chrome users in the last 28 days: 1.5 seconds.

Subject B: Event-ticketing marketplace, Next.js, on Cloudflare. Mobile Performance 30/100. Lab LCP 13.6 seconds. Field LCP, from real users: 1.2 seconds.

If you only looked at the Lighthouse scores, both sites are broken. If you only looked at the field data, both sites are great. Neither read is the full picture.

What Is Actually Happening

There is one site. There are two kinds of visitor.

The repeat visitor, the regular, the existing customer: they have warmed caches, Cloudflare has their assets fingerprinted, the service worker is doing its job, the browser already has the fonts and the framework runtime. They experience the field-data version. Site loads in 1.2-1.5 seconds. They convert.

The first-time visitor, the one who clicked a link from a press article, a friend’s text, an Instagram ad: cold cache, no CDN warm-up, has to fetch the entire payload from scratch. They experience the lab-data version. Site loads in 12-14 seconds on mobile. They bounce.

For Subject A, that is the visitor most affected by the recent press feature. The exact group of new customers the press was supposed to win is the group most punished by the slow first paint.

For Subject B, ticketing is by definition mostly first-time visitors for any given event drop. Cold-cache performance is the entire game. A 4-second TBT means the buy button is interactive seconds after the page renders, and on a sold-out drop, those seconds are the entire conversion difference.

What This Audit Looked At

Three things:

1. Lab data via Lighthouse. What does Google’s testing tool see when it loads the page from a clean Chrome instance, throttled to mobile-network speeds. This is what most agencies report.
2. Field data via Chrome User Experience Report (CrUX). What actual Chrome users have experienced visiting this page over the last 28 days. This is what Google’s search ranking uses.
3. The gap between them. Because the gap is the diagnostic. A small gap means the site is consistent. A large gap means the site behaves very differently for repeat vs. first-time visitors, and you need to fix the first-time experience specifically — not “the site.”

The Diagnosis For Both Subjects

For Subject A (Shopify boutique), the cause is third-party app overhead. Most boutique Shopify stores accumulate 4-8 apps over the years: reviews, popups, analytics, upsell, abandoned-cart. Each app injects JavaScript that runs on first paint. The fix is an app audit — identify which apps are earning their cost, defer or remove the rest. Estimated mobile-LCP improvement: 12.6s → 4-5s on a cold cache. Two-week engagement.

For Subject B (Next.js marketplace), the cause is bundle weight. Total Blocking Time on mobile is 4.07 seconds — the highest we measured in the audited cohort. Next.js gives you the tools to fix this; the dev team likely shipped features faster than they shipped perf. Recommendation: route-level code splitting + a real performance budget. Likely 6-8 weeks of engineering augmentation, not an agency rebuild.

What This Means For You

If your owner-eyes-only experience of your site is fine and your conversion rate from new traffic is bad, this is probably the diagnosis.

The Lighthouse score in isolation is misleading. The CrUX field score in isolation is misleading. The gap between them is the actual signal.

A real audit looks at both. A Fit Check from us takes about 90 minutes of analysis and produces a written report at the same level of detail as the above. No engagement required. No sales call attached.