Case study
A site that ranks. An ad budget that finally buys leads.
Alan Construction is a roofing and exteriors contractor in Pittsburgh. A year ago they were on a Webflow site that was not ranking, and paying every month for a Google Ads account that was not producing a single roofing job. We rebuilt both, website first. This is the year since.
- 0 qualified leads in 2025On a full ad budget, every month
- 2/day qualified ad leads, from zeroNone in 2025, two a day by September 2026
- 6.8x local search clicks2026 vs the last Webflow quarter
- 93% lower cost per leadRebuilt account vs the inherited one
What we did
Rebuild, do not patch. In that order.
The site was not ranking and the ads were not converting, and both problems started with the website. We started over there, then rebuilt the ad account on top of it.
How we market roofing companiesThe audit
A Webflow site that was not ranking, and a Google Ads account that had spent a full budget every month of 2025 without a qualified lead.
The website, rebuilt
Fast, mobile-first, built around the phone number and the quote form, with real service and area pages instead of a brochure. Built to rank from the first day.
Google Ads, rebuilt from zero
New account, built from zero, with every call and every form tracked from day one. Bing Ads followed.
Call tracking on every channel
A different number for every source, so a new customer from the ads is counted separately from a repeat customer calling the office.
SEO, done for real
Service pages for the work Alan actually sells, area pages for where he sells it, and articles answering what homeowners search before they call a roofer.
Google Ads, before and after
A year of ad spend, no roofing jobs. Then not one month without.
The inherited account spent a full budget every month of 2025 and did not produce a roofing job. The few leads it recorded were not customers. We started a new account in December rather than fix the old one, with every call and every form tracked. Each square is a month. Since February, the first full month, the account has averaged more than one qualified lead a day and passed two a day by September, at 93% less per lead than the old account paid for the few it recorded. Bing Ads followed and now brings one in five ad leads.
Phone calls and form submissions recorded by Google Ads. Shade shows each month against the best month; the scale is deliberately not printed.
Local search clicks vs the Webflow quarter Search Console
SEO
From no SEO to a site that ranks.
On the Webflow site, Search Console shows a trickle of clicks a month, nearly all people typing the company name. The rebuilt site started ranking within a quarter, for Pittsburgh roofing terms and for the questions homeowners search before they call a roofer.
Roofing search is seasonal: it climbs through spring, peaks in summer and eases into fall. Since February, even the slowest month runs nearly five times the last Webflow quarter, and across 2026 local search clicks run 6.8 times it.
New customers from the ads
Most ad calls are people who never called before.
Every ad has its own tracking number, so a new homeowner from the ads is counted separately from a repeat customer calling the office. On Google Ads, 77% of callers are calling for the first time. On Bing Ads, which brings one in five of the ad leads, 72%. Those are the leads that grow the business.
By the numbers
Numbers you can check, not claims.
Every figure on this page comes from Google Ads, Search Console, GA4 and CallRail, pulled on the same day. Alan Construction owns the ad account, the website and the data. We have access. They have the keys.
Why clients own everythingWhat we would tell another roofer
Three things this account proves.
The same three show up on every contractor account we take over. This one shows them in a single year.
Clicks without leads is not a budget problem
It is the setup, the landing page, or the tracking, usually all three. More money into that account would have bought more of the same.
Rebuild instead of patching
An account that has spent months without converting has taught Google the wrong customer. Starting over was faster than unteaching it.
Count new customers, not calls
Repeat customers calling the office are good business, but they are not what the marketing bought. Track first-time callers by source or the report flatters itself.