We Tested 265 NJ & CT Contractor Websites. 96.6% Miss Google's 2.5-Second Mobile Load Target.

Correction, October 16, 2026: an earlier version of this article said 94% of sites fail Google’s speed test. 94.0% is the share of sites slower than 3 seconds. The share slower than Google’s 2.5-second LCP threshold is 96.6% (256 of 265). We also clarified that this is a lab test of LCP only, not a full Core Web Vitals assessment, and removed claims we did not measure.

We measured the mobile load speed of every contractor website we could find across New Jersey and Connecticut, 278 businesses, 265 successful measurements, all through Google’s own PageSpeed testing infrastructure. The result: only 9 of 265 sites (3.4%) loaded their main content within Google’s 2.5-second LCP threshold in a lab mobile test. The median contractor website takes 7.8 seconds to show a homeowner its content. One in five takes more than 15 seconds.

Google’s 2016 mobile research reported that 53% of mobile visits are abandoned when a page takes longer than 3 seconds to load, and 94.0% of the sites we tested took longer than 3 seconds in the lab. We did not measure visitor behavior on these sites, so this is a risk, not a measured loss: slow mobile pages can cost contractors visitors before the page finishes loading.

Here’s the full data, how we measured it, and what the businesses at the top are doing differently.

The headline numbers

• 265 contractor websites measured (HVAC, plumbing, roofing, landscaping, electrical, and 7 more trades)
• Median load time: 7.8 seconds, more than 3× Google’s 2.5-second LCP threshold
• 96.6% (256 of 265) had a lab mobile LCP over Google’s 2.5-second threshold
• 94.0% (249 of 265) took longer than 3 seconds
• Only 3.4% (9 of 265) were at or under 2.5 seconds
• 21.1% took 15 seconds or longer in the lab test
• Median Google PageSpeed mobile score: 60 out of 100

How the 265 sites break down

Nearly half the market (49.4%) sits at 8 seconds or worse. These aren’t broken websites, and many look fine on a fast office connection. The lab test simulates a phone on a mobile connection, where these pages took 8 seconds or more to show their main content.

The trade-by-trade league table

Median mobile load time by trade (trades with 10+ sites measured):

Two things stand out. First, no trade passes. The best median in the study, garage door companies at 4.7 seconds, is still nearly double Google’s standard. Second, the trades with the most visual work to show off (painting, landscaping, roofing) are the slowest, possibly because their sites carry large photo galleries. We did not measure the cause of slowness, so treat this as a likely explanation, not a finding.

Why slow pages can cost jobs

Google’s mobile research found that 53% of visits are abandoned when a page takes longer than 3 seconds to load. Our median site takes 7.8.

Here is a hypothetical example of what that can mean for one business. A homeowner’s AC dies on a July afternoon. She searches, taps the first result, white screen. Three seconds. Four. She’s already back on the results page tapping the next one. The first contractor paid for a website, maybe paid for the click too, and never got seen. The second contractor got the call because his site simply showed up.

Speed isn’t a technical vanity metric. When a homeowner compares several contractors, the page that shows up first is the first one they can act on.

What the 9 passing sites have in common

We did not formally audit why the 9 fastest sites were fast. In general, lab LCP improves with images sized and compressed for phones, lighter themes, fewer plugins and better hosting. Whether a given slow site can get under 2.5 seconds with cleanup alone, or needs a rebuild, depends on the site.

Methodology

We started from a list of 278 licensed home-service businesses across New Jersey and Connecticut (12 trades) with live websites. In July 2026 we measured each homepage through Google’s PageSpeed Insights API v5 using the mobile strategy, recording Largest Contentful Paint (LCP), the time until the page’s main content is visible, and the overall mobile performance score. 265 sites returned valid measurements; 13 failed (unreachable or measurement errors) and are excluded. “Passing” here means a lab mobile LCP of 2.5 seconds or less, Google’s published “good” threshold for LCP. This is a lab test (Lighthouse run through the PageSpeed Insights API on a simulated mobile device), not Chrome UX Report field data from real visitors, and it measures LCP only: the other Core Web Vitals (INP and CLS) were not assessed, so this study does not say whether a site passes Core Web Vitals overall. The denominator is the 265 sites with valid measurements. Medians are used throughout because a handful of extremely slow sites (the worst measured over two minutes) distort averages.

The full anonymized dataset (all 265 measurements with trade, city, LCP, and PageSpeed score) is free to download: study-data-public.csv. Journalists and researchers: cite freely with a link. We can also cut the data by trade or county on request.

Check your own site

Two minutes, no tools needed: open pagespeed.web.dev, enter your website, and read the mobile LCP number. Under 2.5 seconds, you would be among the 3.4% of sites in our sample that met the threshold. Over 8 seconds puts you in the slowest half of our sample.

If you’d rather have us do it: we run a free breakdown that checks your speed, your reviews, your visibility, and where your calls go, using your real numbers. No pitch, you keep the report. Get your free breakdown.