On this page
Why retargeting carries trade accounts
Consider what a re-roof, a new kitchen or a set of bifold doors actually involves for the household buying it. Somebody notices a problem. Weeks pass. They look at a few websites one evening. More weeks pass. They mention it to a neighbour. They gather two or three quotes. Somewhere in the third month, they sign with one of them.
Almost every advertising channel is built to capture the moment they were ready. Retargeting is the only one that keeps you present through the two months before that. For long-consideration trade work, that is not a supporting tactic; it is where a large share of the value in a paid social budget sits.
It is also the cheapest inventory you will buy. Reaching somebody who has already read your pricing page costs a fraction of finding a stranger, because the audience is small and the relevance is high. A firm running no cold advertising at all can still run retargeting profitably against its existing search and organic traffic, and for a business with a tight budget that is often the right first move on the platform.
The catch is that it only exists if the plumbing was right from day one. A pool you did not start collecting six weeks ago is a pool you do not have, and there is no way to backfill it. This is the argument for putting the tracking in place before you are ready to advertise rather than afterwards: the cost is an hour of setup, and the alternative is waiting two months for an audience you could already have had.
One more thing worth stating early, because it shapes every decision below. Retargeting is not a way of pestering people into buying. For work of any size the household is not resisting you; they are working through a list of decisions, most of which have nothing to do with which firm they pick. Being present and useful while that happens is a different job from applying pressure, and accounts built as though it were the latter reliably underperform and generate complaints.
Building pools big enough to matter
The commonest reason retargeting disappoints is that the audience is too small to deliver. A trade site getting a few hundred visitors a month, with a thirty-day window, produces a pool of a couple of hundred people, and Meta cannot spend meaningfully against that. The result looks like failure and is actually arithmetic.
Build wider sources than website traffic alone.
- Website visitors, segmented by what they looked at. Everyone who reached a service page is one pool; everyone who reached pricing or contact is a much hotter one.
- Video viewers. Anyone who watched a meaningful share of a video. This is the most under-used source in trade accounts and it scales with your cold spend rather than with your site traffic.
- Page and profile engagers. People who interacted with your Facebook page or Instagram profile. Broader and cooler, but it fills a pool quickly.
- Instant form openers who did not submit. They opened it and stopped, which is a stronger signal than most of what you collect.
- Your own customer list, uploaded within the marketing permissions you hold, used both for repeat work and as the seed for a lookalike.
Build lookalikes from completed jobs rather than from raw enquiries wherever you have the volume. A lookalike modelled on people who actually paid you is a materially different instruction from one modelled on everyone who ever filled in a form, and the difference shows up in enquiry quality.
Pixel, Conversions API and the gap between them
A browser pixel alone no longer sees everything, and the losses are not evenly distributed.
Apple's app tracking prompt means many iPhone users decline cross-app tracking outright. Browsers have restricted third-party cookies and shortened the lifetime of the first-party ones. Ad blockers stop events firing at all. The consequence for you is twofold: your reported conversions understate reality, and, more damagingly, the optimisation system learns from a partial picture and therefore delivers worse.
The Conversions API addresses both. Your server or your CRM sends conversion events to Meta directly, matched on hashed details such as email and phone number, rather than depending on the visitor's browser. Run it alongside the pixel with proper event deduplication, so the same conversion counted twice does not distort the data you have just improved.
Two things worth doing while you are in there. Send offline events for jobs that were won on the phone or on site, so the system learns from real outcomes rather than form submissions. And keep the event payload generic: an event name should not describe somebody's health, finances or personal circumstances. Where we build this alongside a CRM, the CRM is usually the cleanest place for it to live, because the CRM already knows which enquiries turned into money.
Window lengths for slow decisions
Default retargeting windows are short because the defaults were designed for online retail. A person considering a pair of trainers decides in days. A household considering eight replacement windows does not, and a thirty-day window drops them halfway through.
Match the window to the decision, not to the default.
- Small reactive work — gutter clearing, minor repairs, a service call. Seven to fourteen days. After that the job has been done by somebody.
- Mid-ticket planned work — driveways, landscaping, tree reduction, external rendering. Thirty to sixty days.
- High-ticket considered work — glazing, kitchens, bathrooms, re-roofs, solar, extensions. Sixty to one hundred and eighty days, and yes, the longer end is deliberate.
- Seasonal work. Long enough to reach the season. Somebody researching garden rooms in November should still be reachable in March.
Long windows cost very little, because the audience is small and the inventory is cheap. What they buy is presence at the moment of choice, which is usually weeks after the moment of interest. Layer within the window as well: somebody who visited yesterday and somebody who visited ten weeks ago are at different stages and should not see identical advertising.
What a warm audience should actually see
The single biggest retargeting mistake in trade accounts is showing warm audiences the same advertisement that brought them in. They have seen it. It worked. Repeating it tells them nothing new and irritates them into hiding your ads.
Warm audiences have moved past awareness and into evaluation, so the questions have changed. They are no longer asking whether they want the work. They are asking whether you are the right firm, what it will cost, what could go wrong, and how disruptive it will be.
- Proof. Completed jobs explained properly, with the specific problem and how you solved it. One job told well beats a gallery of thumbnails.
- Reviews, shown as content. Real reviews with names and places, not a star rating graphic.
- Process. What happens between the enquiry and the finished job. Survey, quotation, lead time, payment stages, guarantee. Uncertainty about process stalls more high-ticket jobs than price does.
- The people. Who will actually turn up. Faces do disproportionate work at this stage.
- The objection, addressed directly. Whatever you hear most often on the phone, say it in an advertisement before they have to ask.
Save the direct ask for the hottest layer only: people who reached a quote or contact page and did not complete. There, be plain. Here is what happens next, here is the number, here is how long it takes.
Frequency, fatigue and exclusions
Small audiences burn out fast. If a pool of eight hundred people has a budget of thirty pounds a day pushed at it, they will see your advertisement several times a day and start to resent it. Cap frequency, keep retargeting budgets proportionate to pool size, and refresh creative on warm layers more often than on cold ones, because the same people are seeing it repeatedly.
Exclusions matter just as much and are more often neglected.
- Exclude converters from every prospecting campaign. Paying to advertise to somebody who already enquired last week is pure waste and is irritating.
- Exclude existing customers from acquisition messaging, and advertise to them separately about repeat and complementary work.
- Exclude hot-layer audiences from cold campaigns so you are not paying prospecting prices to reach people you already own.
- Check what Advantage+ audience expansion is doing to your exclusions. Its expansion can quietly undo the separation you carefully built, and this wants reviewing monthly rather than at setup.
Also exclude people who applied for a job, anyone who enquired and was disqualified, and, for seasonal work, people who bought last season and will not buy again this year.
Measuring it without fooling yourself
Retargeting flatters itself in reporting. It reaches people who were already considering you, so it takes credit for conversions that might have happened anyway. Any channel positioned at the end of a decision looks brilliant in a last-click report.
Three habits keep it honest. Compare retargeting-attributed conversions against your total enquiry volume rather than in isolation, so you can see whether the account is producing more work or merely re-labelling it. Run an occasional holdout: switch retargeting off for two or three weeks and watch what happens to total enquiries, which is blunt and revealing. And ask every enquirer where they came from, because their answer, while imperfect, is generated outside the system that wants to take the credit.
The honest summary is that retargeting is usually the highest-return part of a trade Meta budget and it cannot be the only part, because a pool has to be fed by something. Cold prospecting, search and organic visibility all pour into the same pool. Retargeting converts what they gathered. Judge the system rather than the campaign, which is the principle the whole methodology rests on, and check the glossary if any of the terminology here is new.
Rather have someone else do this?We run exactly this for trade and construction firms across the UK. Call and we will tell you what is realistic in your area, including when the answer is that it is not worth it.
Call 07443 392243WhatsAppCommon questions
How much of a Meta budget should go on retargeting?
For most trade accounts, somewhere between a quarter and a half, though it depends entirely on how much traffic you have to retarget. The constraint is pool size rather than preference: if your audiences are small, pushing budget at them just raises frequency until people hide your ads. Grow the pool first through cold prospecting and search traffic, then raise the retargeting share behind it.
Is retargeting worth it if we get very little website traffic?
On website visitors alone, probably not yet, because the pool will be too small to deliver. Video viewers and page engagers are the way around this: they fill quickly and cost little to build. Run cold video for a few weeks to create an audience, then retarget it. That sequence works on a small budget where straight website retargeting would stall.
Do long retargeting windows annoy people?
Only if the frequency is wrong or the creative never changes. A hundred-and-eighty-day window with a low frequency cap and rotating material is barely noticeable. The same advertisement six times a week for a fortnight is what generates complaints, and that is a budget-to-audience-size problem rather than a window-length one.
What about people who already had a quote and said no?
Keep them in a separate audience and treat them differently. Some declined on price, some on timing, and the timing group is worth reaching again in three or six months with something that acknowledges the gap. Repeating the original advertisement to somebody who already said no is the fastest route to a hidden ad and a poor impression.