What separates page-one local-service sites — measured, not guessed
Every figure here is measured on real sites — never estimated. These are correlations in our own measured sample, not claims about the whole market, and never statements of cause. A group is shown only when we have measured at least 5 distinct sites; below that we say so plainly. The sample grows as we measure more.
Measured on 163 distinct real sites (75 top-ranked, 89 bottom-ranked) across 19 markets and 8 industries. Grows as we measure more.
The cross-industry board
Across every industry we have measured, these are the signals the top-ranked pages share more often than the bottom-ranked ones — the same handful, regardless of vertical.
| Signal (measured) | Top-ranked | Bottom-ranked | Gap | Leans |
|---|---|---|---|---|
| Has an embedded video | 70% | 38% | +32 pts | ▲ top-ranked |
| Terms / legal link | 55% | 30% | +25 pts | ▲ top-ranked |
| Privacy policy link | 83% | 59% | +24 pts | ▲ top-ranked |
| Site icon present | 96% | 78% | +17 pts | ▲ top-ranked |
| Testimonials / reviews on page | 52% | 35% | +16 pts | ▲ top-ranked |
| Mentions the target city/market | 75% | 59% | +16 pts | ▲ top-ranked |
| Title length in the effective range | 59% | 75% | -15 pts | ▼ bottom-ranked |
| Visible phone number | 84% | 69% | +15 pts | ▲ top-ranked |
| Visible contact email | 18% | 33% | -15 pts | ▼ bottom-ranked |
| Visible street address | 76% | 62% | +15 pts | ▲ top-ranked |
| Has JSON-LD structured data | 87% | 73% | +14 pts | ▲ top-ranked |
| Navigation landmark region | 92% | 79% | +13 pts | ▲ top-ranked |
| Review/AggregateRating structured data | 27% | 14% | +13 pts | ▲ top-ranked |
| BreadcrumbList structured data | 37% | 25% | +12 pts | ▲ top-ranked |
| Breadcrumb navigation | 47% | 35% | +12 pts | ▲ top-ranked |
| Organization structured data | 50% | 38% | +12 pts | ▲ top-ranked |
Presence among the 75 top-ranked and 89 bottom-ranked distinct sites we have measured. Higher on top-ranked = a signal the winners share more often. Correlation, not cause.
The basics no longer differentiate
These are present on ~everyone — top and bottom alike: Served over a secure connection (HTTPS), Character encoding declared, Open to search indexing, Has a title tag, Share of descriptive (non-generic) link text, Forms submit securely. In our sample they do not separate page-one sites from the rest. The gap is in the completeness signals above, not the basics.
By market
A market shows a benchmark only once we have measured ≥5 top-ranked sites in it; the rest are honestly marked insufficient and fill in over time.
| Market | Measured (distinct sites) | Status |
|---|---|---|
| Personal Injury Lawyer — Phoenix | 9 top / 6 bottom | benchmark |
| Plumbing — Austin | 8 top / 16 bottom | benchmark |
| Interior Designer — San Diego | 7 top / 7 bottom | benchmark |
| HVAC Repair — Broadview | 6 top / 8 bottom | benchmark |
| Med Spa — Denver | 4 top / 3 bottom | insufficient (N<5) |
| Med Spa — Scottsdale | 4 top / 3 bottom | insufficient (N<5) |
| Personal Injury Lawyer — Houston | 4 top / 3 bottom | insufficient (N<5) |
| Plastic Surgery — Houston | 4 top / 7 bottom | insufficient (N<5) |
| Plumbing — Houston | 4 top / 3 bottom | insufficient (N<5) |
| Plumbing — Dallas | 4 top / 6 bottom | insufficient (N<5) |
| Cosmetic Dentistry — Austin | 3 top / 3 bottom | insufficient (N<5) |
| Cosmetic Dentistry — Raleigh | 3 top / 3 bottom | insufficient (N<5) |
| HVAC Repair — Phoenix | 3 top / 3 bottom | insufficient (N<5) |
| Interior Designer — Austin | 3 top / 3 bottom | insufficient (N<5) |
| Med Spa — Miami | 3 top / 4 bottom | insufficient (N<5) |
| Pdf Editor — Houston | 3 top / 3 bottom | insufficient (N<5) |
| Personal Injury Lawyer — Atlanta | 3 top / 3 bottom | insufficient (N<5) |
| Plastic Surgery — Dallas | 3 top / 3 bottom | insufficient (N<5) |
| Plastic Surgery — Miami | 3 top / 3 bottom | insufficient (N<5) |
How we measure: every site is read live and scored on the same 58-signal catalog; only confirmed / observed measurements enter these aggregates (a site that refused a read is excluded, never counted as zero). We never publish an individual site's numbers.