Paid Search for Roofers
Spend when it rains. Almost nobody does this.
The tactic your competitors aren’t using
Roofing demand isn’t steady. It’s made by weather. A storm rolls through and searches for roof repairs jump across an entire region for about seventy-two hours, then settle back.
Most roofers respond to that by being overwhelmed for a week. A few respond by turning ads on — usually two days late, by which point the work has gone.
The version that works is having the budget pre-configured and triggered by the forecast, so you’re visible the morning the wind drops rather than the following Thursday. That’s a genuine competitive advantage and it’s available to anyone willing to plan it.
Organic and map pack position can’t do this. They’re slow-moving and you can’t turn them up. Paid is the only channel that can respond to weather, which makes it unusually valuable in your trade specifically.
Between storms there’s a steadier case too — re-tiles, flat roofs and roofline at £1,790 to £5,090, where the job value easily justifies a paid click.
What actually works
A dormant storm campaign, ready to go. Ad copy written, landing page built, negatives in place, budget set. It sits at zero until you need it, then goes live in minutes.
Watching the forecast, not the news. Named storms are announced days ahead. That’s your signal.
Emergency copy that’s honest about capacity. “We’ll be out within 48 hours” beats “immediate response” if you can’t actually be there immediately. Overpromising in storm week produces reviews you’ll regret.
A separate always-on campaign for planned work. Re-tiles, flat roofs, roofline. Different intent, longer consideration, landing pages rather than call buttons.
The survey as an entry offer. £175 for a survey is a low-commitment click that converts into £5,090 re-tiles. Cheaper to advertise than the re-tile directly.
Negatives for DIY, materials and jobs. Roofing attracts a lot of research traffic that will never hire you.
What we’d actually do
Build the storm campaign first and leave it paused — that’s the asset, and it needs to exist before you need it.
Build the always-on planned campaign around surveys, re-tiles, flat roofs and roofline, with a modest steady budget.
Set up call tracking so we can see which campaign produced what, and which storm weeks actually paid.
Then agree a trigger process: when a named storm is forecast for your area, the campaign goes live at a pre-agreed budget, runs for the window, and switches off. No decisions needed in the moment, when you’re on a roof and not thinking about advertising.
Then review after each event, because the pattern will be specific to your area and it gets sharper each time.
The first 90 days
Weeks one to two. Build the storm campaign and pause it. Ad copy, landing page, negatives, geography and budget all agreed in advance. This is the asset and it has to exist before you need it — building it during a storm is the mistake that defines this trade.
Weeks two to four. Always-on planned campaign live at a modest budget, built around the £175 survey as the entry point rather than advertising re-tiles directly. Surveys are a much cheaper click and they progress into the big jobs.
Month two. Cost per survey should be clear, and we’d start tracking how many surveys become re-tiles. That progression is the number that matters, not the cost of the initial lead.
Month two onwards, whenever weather arrives. The trigger process runs: named storm forecast for your area, campaign goes live at the pre-agreed budget, runs the window, switches off. No decisions needed while you’re on a roof.
Month three. First review of the whole picture. If a storm has happened, we’ll know a lot. If it hasn’t, we’ll know the planned side and still be waiting on the interesting half.
Be prepared for this to look flat until it suddenly doesn’t. Your results arrive in bursts, and a quarter with no significant weather is not evidence that the approach is wrong.
What good looks like
Storm weeks producing three to five times normal enquiry volume. That’s the whole point of the exercise.
Cost per lead in storm week higher than normal, and worth it. Competition spikes too. Paying more per lead is fine when the alternative is being invisible.
Planned campaign at a cost per job under 10% of job value. At £1,790 average that’s generous room.
Survey ads producing re-tiles. Track the progression, not just the initial job.
The mistakes we see most
Building the campaign during the storm. Too late — it takes days to settle and the window is three days.
Not watching the forecast. The signal is free and public.
Promising immediate response you can’t deliver. Storm-week reviews last a lot longer than storm week.
Advertising re-tiles directly. Expensive clicks on a big-ticket decision. Sell the survey instead.
Leaving storm budget running after the spike. Costs stay high, volume doesn’t.
No negatives. You’ll fund a lot of people pricing up materials for a DIY repair.
What’s it worth?
Average job £1,790 · storm-week enquiries 20+ · normal month 6
A single well-caught storm can be worth more than a quarter of steady advertising.
Every enquiry lands in Gaffer
Storm-week enquiries land in one place instead of across three phones and a notepad, which is exactly when things get lost. Survey photos attach to the job and feed your before-and-after library. And everyone you’ve worked for is a list for next time.
Questions roofers ask
How far ahead can we set this up?
Now, and leave it paused. That’s the correct way to do it.
What budget for a storm week?
More than you’d normally spend, for a short window. We’d agree the number in advance so nobody’s deciding under pressure.
Is it worth advertising in summer?
Yes, for planned work. Re-tiles and flat roofs get booked in dry weather.
Ad spend separate?
Always, in your own account. You see every penny and control the tap.
Free review · Bigger outfit? Talk to us
See what we'd do for roofers
Put your website in and we'll tell you what's losing you work — speed, mobile, your Google profile, reviews and the pages you're missing. Two minutes, no call unless you want one.