§ RankingsMethodology
Run the numbers yourself.
Every edition of the rankings is built to be recomputed. This page publishes the whole rubric — the eight scored dimensions, the objective checks, the exact tools, and the rules that keep the scores honest.
01The 8 dimensions
Eight scores, one to ten, evidence attached.
Each website is scored 1–10 on the eight dimensions below. Every score is recorded with written, cited evidence at grade time — a URL, an observed fact, a measured number. The composite is the unweighted mean of the eight, rounded to one decimal. No dimension weighting, no judgment calls hidden in the math.
| Dimension | What it measures | How it's checked |
|---|---|---|
| Structure & depth | Does the site have real pages for real intents — services broken out, service areas, about — or is it a one-page brochure? | Sitemap + crawl of the nav; approximate page count and service breakout recorded as evidence. |
| Technical SEO | Titles, meta descriptions, heading hierarchy, Open Graph tags, sitemap.xml and robots.txt status, JSON-LD structured data, HTTPS. | Direct fetch of the homepage with a browser user-agent; HTTP status probes on /sitemap.xml and /robots.txt; JSON-LD @type inventory extracted from the source. |
| Google Business Profile signals | Does the website reinforce the business's map-pack presence — NAP consistency, LocalBusiness schema with geo and hours, embedded map, live review surfacing? | Schema inventory + on-page NAP/map/review-widget observation. |
| Content cadence | Is anything shipped over time — a dated blog, a projects feed — or has the site been static since launch? | Blog/articles discovery via nav and sitemap; cadence judged from dated entries. |
| Conversion paths | How hard is it to become a lead: tap-to-call links, forms, CTAs above the fold, multiple contact modes. | tel: link detection, form count, CTA inventory on the homepage. |
| Mobile experience | Real-phone usability: speed, layout stability, tap targets. Most local-service traffic is mobile. | PageSpeed Insights mobile run (Performance, LCP, CLS, TBT) plus tap-to-call presence. |
| Brand consistency | Is there a recognizable, coherent identity — or interchangeable template copy? Named owners, real proof, consistent design system. | Qualitative read of the homepage and about page, scored with written evidence. |
| Programmatic local depth | Differentiated per-city pages with genuinely local content — the difference between claiming a service area and covering it. | Per-city URL discovery + a read of whether the copy actually differs by city. |
02Objective extras
The checks that are pure measurement.
- PageSpeed Insights, mobile, homepage.Every site gets one PSI mobile run on its homepage via Google's free PageSpeed Insights API — Performance, Accessibility, Best-Practices, and SEO scores plus LCP, FCP, TBT, CLS, and Speed Index. Lab data (these are low-traffic sites without CrUX field data), which is exactly why anyone can reproduce it.
- JSON-LD inventory. The full list of schema.org @type values found in the homepage source. LocalBusiness or a trade-specific subtype counts for more than generic Organization/WebSite markup.
- Review counts, as-seen-dated. Star rating and count recorded from the named source (Google panel, Birdeye, Chamber, BBB) with the date they were read. Review counts drift — the date is part of the claim, and each figure carries a confidence label when the source is thin or ambiguous.
- SSL / HTTPS validity. A live certificate handshake. A site that throws a browser security warning before it loads fails the baseline, and that gets stated factually.
03Reproducibility
If you can't recompute it, it isn't a score.
Run these tools yourself. Every number in every edition points to a URL plus an observed fact:
- PSI numbers — rerun any site at pagespeed.web.dev (mobile strategy, homepage).
- Technical signals — fetch the homepage with a browser user-agent and read the source; probe
/sitemap.xmland/robots.txtfor their HTTP status. - Schema — paste any URL into Google's Rich Results Test or read the
application/ld+jsonblocks directly. - Review counts — visit the cited aggregator page. Values drift over time, which is why every count in the rankings carries its as-seen date.
Lab numbers vary run to run — a Performance score a few points off from the published one is normal; an entirely different grade is not. If you find a real discrepancy, tell me and the edition gets corrected and re-dated.
04The rules
What keeps the scores honest.
- Scores are real, always. Machine-checkable metrics only. The raw metric table ships public with every ranking — anyone can recompute. A retainer never moves a number.
- Disclosure.CDS-built sites are included and badged “Built by Campbell Digital Studio” — never hidden, never excluded. When a studio client appears in an edition, the badge is on the row; when none do, the edition says so explicitly.
- Category awards, not rigged ties. Best Overall, Fastest Site, Best Booking Experience, Best Local Coverage — multiple honest winners per industry, each with a stated, evidence-backed basis. Awards follow the numbers; when a category needs a caveat, the caveat is printed with the award.
- No worst-lists.Best-lists and aggregate stats only (“85% of sites fail the mobile LCP threshold”). Every graded business appears in the raw table because the table is the data — but businesses below the top of the field get no individual negative commentary. Sharp about craft, never cruel about a named business.
- Refresh cadence. Editions are quarterly-refresh assets, not one-shots: each graded industry gets re-scored on the refresh cycle (first refresh: Q1 2027), movers get written up, and stale numbers get replaced rather than left to rot. Every figure keeps its as-seen date so staleness is self-declaring.