Meta sells what nobody searches for
Search advertising has a hard ceiling: it can only reach people already looking. For a lot of service work, that is a small fraction of the people who would happily buy.
Nobody types “backyard pergola” into Google before the idea has occurred to them. They see one, in a neighbourhood that looks like theirs, and then they want it. The ad is what created the demand.
That is what Meta is for, and it is why comparing it to Google on the same report is usually unfair to both.
Which trades it suits
Meta works when seeing the result is what produces the desire, and when the purchase can wait.
Remodeling and outdoor construction are the clearest cases — the work is visual, the value is obvious in a photograph, and the buyer was not searching. Detailing and ceramic coating work for the same reason. Pest and lawn plans work because the offer can be made attractive to someone who had not planned to act. Dealerships work because inventory is inherently visual and most people are quietly open to replacing a car.
It works badly for emergencies. Nobody scrolling Instagram has a burst pipe. For that moment search advertising is the right channel and Meta is the wrong one, and an agency that sells you Meta for emergency plumbing either does not understand the channel or is not being straight with you.
It is a creative problem, not a targeting problem
Meta’s targeting options have been progressively simplified for years, and the algorithm now does most of the audience work on its own. What still separates a good account from a bad one is the creative and the offer.
Two things matter more than production value.
Volume. Creative fatigues. The same ad shown to the same Austin audience stops working within weeks, and an account fed one set of images a quarter will decline predictably. A steady supply of ordinary, honest variations outperforms occasional polished productions.
Authenticity. People scrolling identify stock imagery instantly and discount everything around it. A phone photograph of an actual finished driveway, deck or kitchen — ideally somewhere locally recognisable — will beat a professionally produced generic image most of the time. The phone in your crew’s pocket is the most underused marketing asset in the trades.
The measurement trap
This is where Meta campaigns die, almost always for the wrong reason.
Someone sees your work on a Tuesday, follows the page, thinks about it, mentions it to their partner, and inquires three weeks later — possibly by searching your name, which attributes the lead to Google.
Judged on a seven-day window alongside a search campaign, Meta looks like it failed. The budget moves to search. The top of the funnel quietly empties, search performance declines a few months later for reasons nobody connects back, and the conclusion is that Meta does not work.
Avoiding that requires tracking built for the actual decision length and the discipline to fund the two channels separately. They are not competing for the same job.
The Pearl Motors campaign is the clearest version of this we have run: the figures that mattered were vehicle purchases connected to the campaign, not inquiry volume. The case study has the detail.
The landing page has to honour the ad
Search traffic arrives warm — they went looking. Meta traffic arrives cold, mid-scroll, mildly interested.
That means a far shorter patience window and a much heavier burden on whatever they land on. If the ad showed a finished patio and the page is a generic homepage with a navigation bar and a contact form at the bottom, the interest the ad created is spent before anything happens.
The ad and the page it lands on have to be built together — same work shown, same claim, one obvious next step — and the follow-up has to catch the inquiry while the interest is still warm.