Local search for roofing companies.
Dedicated city pages, a working Business Profile, and seasonal content published before the season.
The pattern, stated directionally.
The consistent finding across sources is directional rather than numeric: one generic service page underperforms against dedicated city and neighbourhood pages, Google Business Profile optimisation drives enquiries without ad spend, and storm-damage and seasonal content performs strongly at a local level. No invented percentages appear anywhere in this cluster, because none are sourced. Where a claim is a best-practice pattern rather than a measured figure, it is written as one.
Why a generic page loses.
A single page covering twenty suburbs cannot answer any of them specifically. A dedicated page can name the roof types common in that suburb, the hail history, the permit authority, response times, and completed projects nearby. That specificity is what a local searcher and an answer engine are both looking for, and it is the only defensible reason for the page to exist.
The line between specific and thin.
A real city page says something true about that place. A thin page swaps the suburb name into the same paragraph twenty times. The second kind is worse than not having the page, because near-duplicates compete with each other and dilute the authority that should have concentrated in one place. Consolidating them into fewer strong pages is the most common recommendation in an audit of a multi-metro roofing site.
The three inputs to map-pack visibility.
Proximity to the searcher, which you cannot change without a real location. Relevance, driven by categories, services, and local content on the site. Prominence, driven by review volume and recency, citations, and how well known the business is. Two of the three respond to work, and both of those respond to the same work.
What actually moves the profile.
Six items, all clerical rather than clever.
Timing decides the seasonal pages.
Storm demand arrives in hours and indexes in months, which makes seasonal content an off-season build. Templates ready before the season, evergreen claims and inspection answers live early enough to settle, target suburbs drawn from your own hail history. After the event the pages stay up as a dated record of it, because that record is what the next one ranks on.
Multiple metros need structure, not more pages.
Volume is not the problem, hierarchy is. A metro hub with suburb pages under it, linked upward and sideways, holds together at any size, provided each hub carries something only it can say: crews, response times, local projects, permitting. Growth by page-adding produces the opposite, and the repair for that is structural rather than editorial.
Schema is the part most sites skip.
Five blocks cover nearly everything a roofing site needs: an organisation block site-wide, Service with an Offer on each service page, AreaServed on the organisation and each city page, FAQPage wherever a real question section exists, and BreadcrumbList on every page below the homepage. Write it as JSON-LD, keep it consistent with what is visible on the page, and validate before launch. Schema that contradicts the page is worse than none.
Where this cluster sits.
Local search is table stakes rather than an edge. Every competent competitor in your market is doing some version of it, which means doing it well is the price of being considered rather than a way to win. The guides below cover each piece concretely, and none of them claim a percentage that nobody has measured.
What a real city page contains.
Six things, none of which survive being templated.
Consolidation usually comes before expansion.
Most sites that grew a page at a time are now competing against themselves. Retiring the thin pages into fewer strong ones, with a redirect map so nothing that earned authority loses it, is unglamorous and routinely the highest-return line in an audit. It requires no new writing and it stops the site bidding against itself.
Reviews pay twice.
Review volume, recency, and genuine responses influence map-pack visibility and Local Services Ads ranking as well as convincing the person reading them. Answer every review, including the bad ones, with specifics rather than a template. A well-answered complaint reads better to a homeowner than an unbroken run of five stars, which most people now read as purchased.
Citations are clerical work that decides rankings.
Name, address and phone have to agree character for character across the site, the schema, the Business Profile and every listing that carries them. Nothing about that is hard, and it stays one of the most common reasons a listing underperforms. It is also the first job after a rebrand, a move or a new number, each of which breaks dozens of listings quietly.
What not to do, in order of risk.
Addresses you do not occupy carry the highest cost, because a suspension takes everything the listing was earning with it. Keyword-stuffed business names sit in the same bracket. Below those: a service area drawn around an entire state, which reads as non-local to Google and to homeowners alike, and stock photography on a city page, which removes the only reason the page existed.
How storm content actually gets published in time.
Build the event template in the off-season. Publish the evergreen claims, inspection, and process answers first so they have months to index. Identify the suburbs your own hail history says get hit. Then, when an event lands, publishing is filling in the event, the date, and the affected areas, which can happen the same day. Keep the page afterwards as a dated record. Deleting it discards the authority it earned.
Where local search stops being enough.
It gets you considered, not chosen. Every competent competitor in your market is doing some version of this, so doing it well is the price of entry rather than an advantage. The advantage sits in the questions nobody in your market has answered properly yet, which is what the AI search cluster is about and what an audit is built to find.
Every guide in this pillar.
The setup that drives enquiries without ad spend, and the clerical errors that quietly cap a listing.
Dedicated local pages outperform one generic service page. Thin copies of each other do not.
Seasonal pages have to be indexed before they are needed, not written during the event.
Proximity, prominence, and relevance, and which of the three you can influence.
How to structure a site covering several service areas without the pages cannibalising each other.
The types that matter, where they go, and how to check they are actually valid.
Questions about this pillar.
How many city pages is too many?+
The limit is content, not count. If a page cannot say something true and specific about that suburb, it should not exist. Thin duplicates actively hurt the pages you care about.
Does the Business Profile matter more than the site?+
For map-pack visibility it is decisive. For everything else the site carries the weight, and the profile has nothing relevant to point at without it. They are not substitutes.
When should storm content be published?+
Months before the season, not during it. Pages need time to be indexed and cited before they are needed, and the same-day publishing that works after an event only works when the template and surrounding content already exist.
Why are there no percentages in this cluster?+
Because none are sourced for these claims. The research behind this site records local-search findings as directional patterns, so that is how they are written here.
Can a virtual office win a nearby suburb?+
It is the fastest route to a suspended listing, and a suspension costs more than the ranking was worth. Win adjacent suburbs with content and reviews, not with addresses you do not occupy.
See what your market rewards.
$2,500, and you get the city-by-city order to build in.