Four metros competing against each other, all under one brand.
A replacement contractor operating across four metros in the Southeast, with 96 near-identical service-area pages that ranked for none of them.
Nine months after launch.
Three figures from the engagement, each traced to the client's own reporting rather than ours.
Pages that earn traffic
Ninety-six thin service-area pages were consolidated into 34 that each answer a distinct query. Organic sessions rose while page count fell by two thirds.
Leads from the two weaker metros
The two metros the company had almost written off produced 3.4 times the qualified form fills they had in the comparable period.
Cost per booked job
Blended acquisition cost across all four metros, comparing the quarter before the rebuild with the third quarter after it.
FIGURES APPROVED IN WRITING BY THE CLIENT. COMPANY NAME WITHHELD UNDER A CONFIDENTIALITY AGREEMENT.
The situation
Growth by acquisition had left the company with four metros served from one domain and a service-area template that swapped the city name into otherwise identical copy. Google treated the set as duplication. The two acquired metros were producing almost nothing and the owner was close to shutting one down.
What the audit found
Of 96 service-area pages, 71 had no impressions at all in the previous six months. Every one carried the same three paragraphs. There was no entity connection between a metro, the services offered there, and the crews who worked it, so nothing established that this was a real local operation rather than a national skin.
What we did about it
The matrix ranked every service against every suburb across all four metros by revenue potential. Thirty-four suburbs justified a page. Each one got genuinely local content, including permit rules, common roof stock, and storm history, with schema tying the metro to the services and the service area.
Where it landed
The weaker metros recovered first, because nobody there had published a real answer to anything. The strongest metro moved slowest, which is what you would expect where three competitors already publish well.
Ninety-six pages, then thirty-four.
Fewer pages, more traffic. That is usually how it goes.
Templated across four metros
Thirty-four real local answers
FIGURES SUPPLIED BY THE CLIENT AND APPROVED IN WRITING. COMPANY NAME WITHHELD UNDER A CONFIDENTIALITY AGREEMENT.
Consolidate first, then build.
Paid Discovery
Crawl of all 96 pages, a 62-query set across four metros, and a matrix ranking every service and suburb by revenue.
Redirect and merge
Sixty-two pages retired into 34 with a full redirect map, so nothing that had earned authority lost it in the move.
Rebuild and launch
Metro hubs, per-suburb pages with local specifics, entity schema across the set, and post-launch crawl verification.
Read another one.
Three more engagements, each with the figures approved in writing by the client.
Invisible in his own city
Twenty-two years in business, cited in no AI answer about roofing in a metro of 420,000. Twelve months after launch the site is the quoted source and the ad budget is half what it was.
The season won before it started
Out-of-state chasers arrive within days of every event. Twenty-four claims pages published four months ahead meant the first enquiry landed in fourteen hours, not eight days.
Six-figure work sold from a homeowner site
Facility managers research materials, systems and lifecycle cost long before contractors. A specification answer set took commercial enquiries from four a quarter to seventeen.
Questions about this engagement.
Does cutting pages not cut traffic?+
Not when the pages were duplicates. Sixty-two of them had no impressions in six months. Consolidating concentrated the authority instead of splitting it.
How do you write local content at that scale?+
Research per suburb on permits, roof stock, and storm history, written by people who work in this trade. It is slower than a template, which is why the matrix decides where it is worth doing.
What happened to the metro they nearly closed?+
It is now the second-strongest of the four by booked revenue.