Routeless deduplicated audited business domains within the stated batch, keeping the newest completed audit per domain. Homepage signal rates use only sites with a successful homepage fetch. Performance metrics use only sites with a successful mobile PageSpeed run. Each rate includes its numerator and denominator; missing observations are not counted as failures.
The sample is an observational cohort of local business websites, not a census of every US business website. Results describe the measured sample and do not establish causation.
We fetched each live homepage and recorded which signals its markup and scripts contained: contact forms, tappable phone links, analytics tags, structured data, sitemaps, online-booking tools, HTTPS, the mobile viewport tag, and the site platform. A signal rate uses only sites whose homepage fetch succeeded.
Google PageSpeed Insights ran a mobile Lighthouse test on each homepage. Performance figures use only completed runs. Lab results show one controlled mobile profile. Real-user Core Web Vitals come from Chrome field data, available for a smaller subset, and are reported apart from lab results.
| Finding | Instrument | Valid n | Limit |
|---|---|---|---|
| 95.4% of tested sites missed the 2.5-second LCP benchmark | Mobile lab test | 11,184 | This is mobile Lighthouse lab LCP for completed PageSpeed runs, not field performance for every visitor. |
| 63.1% of audited homepages had no detectable contact form | Homepage scan | 11,432 | Detection is limited to the live homepage and known technology fingerprints; absence means not detected, not proven absent. |
| 29.6% had no detectable analytics tag | Homepage scan | 11,432 | Detection is limited to the live homepage and known technology fingerprints; absence means not detected, not proven absent. |
| 22.1% had no detectable structured data | Homepage scan | 11,432 | Detection is limited to the live homepage and known technology fingerprints; absence means not detected, not proven absent. |
| Only 8.4% used a detectable online-booking tool | Homepage scan | 11,432 | Detection is limited to the live homepage and known technology fingerprints; absence means not detected, not proven absent. |
A signal recorded as not detected means the scan did not find it on the homepage. It can still exist on an inner page or behind a script the detector does not recognize. Missing observations are never counted as failures.
The data is available under CC BY 4.0: credit Routeless, link to the source page, and keep the sample size and collection window with any figure you quote. The JSON snapshot holds distributions and cohort records; the CSV is a flat extract of national, platform, industry, regional, and state metrics.
Report a reproducible data, labeling, or methodology error through the company contact page with the affected figure and source. A material correction receives a new revision, a dated correction note, regenerated downloads, and an explanation of whether the reported finding changed.
Read the full report →
AI recommendation study methodology →