On this page
The short answer
For a decent minority of trade businesses, Facebook and Instagram advertising is the best-performing channel they run. For another decent minority it is a straightforward waste of money, and no amount of competent management changes that. The rest sit in between, where it works as a supporting channel provided somebody keeps feeding it material.
The variable that decides which group you are in is not your budget, your agency or your targeting. It is whether the job you sell already exists as a formed intention in the customer's head. A feed cannot arrive at the moment somebody's boiler fails. What it can do is reach the thousands of people who have been aware of a problem for years and have never once done anything about it. That second group is invisible to search, and it is much larger than the first.
Everything else in this guide follows from that one point, including the uncomfortable parts.
Which trades fit, and which genuinely do not
Three questions sort it quickly. Is the job urgent? Is the result visible from outside the house? Would the owner recognise the need without being prompted?
Work that is urgent belongs to search. Emergency plumbing, locksmiths, drainage, electrical faults, gas safety, boiler breakdowns, a tree across the drive. These are decided in minutes by whoever appears first, and advertising them in a feed is buying attention that expires before the need exists. We routinely tell firms in this position to put every available pound into search advertising and their Google Business Profile, and to come back to social later or not at all.
Work that is visible and deferred is where the channel earns its reputation. Driveways, resin surfacing, landscaping, garden rooms, external rendering, fascias and guttering, outdoor lighting, tree reduction. The defining feature is that the customer sees the problem daily and has never typed a word about it into a search engine. Demand for these jobs is not waiting to be captured. It has to be manufactured, and a photograph of a finished job two streets away is the most efficient manufacturing tool available.
Work that is planned, expensive and researched sits in the middle. Kitchens, bathrooms, extensions, glazing, solar, full re-roofs. Social gets in early and cheaply, then hands over to search when the customer finally starts comparing firms. The risk here is impatience: the enquiry arrives weeks or months after the advertisement that started the process, and firms who assess the account at thirty days switch it off just as it begins to pay.
And there is a fourth category people skip. If your customer is another business making a procurement decision, paid social is weak. A contracts manager specifying scaffolding for a site is not going to appoint you because they saw a rig on Instagram. There are narrow exceptions for recruitment and for staying in the memory of local trades, but not for filling an order book.
Creative is the lever, not targeting
Owners who have run search advertising arrive expecting the same controls, and the biggest adjustment is discovering that the settings barely matter. In a well-built feed campaign, the creative accounts for most of the variance in results. Two campaigns with identical targeting, budget and structure can differ by a multiple based on nothing but what is in the video.
What consistently performs for trades is unglamorous and cheap.
- The work itself, filmed on a phone. A machine doing something satisfying, a surface going down, a tree section on ropes. Thirty seconds, no music, no editing.
- Before and after from the same position. The oldest format in the trades and still the most reliable, provided it is genuinely your work.
- A person talking. Whoever will actually turn up, saying one specific thing about one job. Trust in this industry is personal and a logo carries none of it.
- Honest numbers. A price range with the variables named removes the main reason people never enquire about deferred work.
The corollary is a resourcing fact rather than a marketing opinion: a feed consumes creative in a way search never does. The same photograph shown to the same town for six weeks stops working and cannot be rescued by budget. If nobody in the firm will reliably take pictures before packing up, this channel will not work for you. That is worth deciding before you appoint anybody, not four months in.
Why broad targeting now beats interest stacks
There is a persistent belief that skilled paid social means assembling elaborate audiences: homeowners, interested in home improvement, engaged shoppers, in-market for renovation. That approach is largely obsolete and frequently counter-productive.
Two things changed it. Meta removed large numbers of detailed-targeting options, particularly sensitive ones, so many of the stacks people built no longer exist. And its delivery system improved to the point where, given a clean conversion signal and enough budget to gather one, it finds buyers more efficiently than a human-defined audience does. Narrowing the pool mostly narrows the auction you are allowed to compete in, which raises your costs without improving who you reach.
Advantage+ tooling sits on top of this. Its audience features help most on cold prospecting, where the system has room to explore and you have no reliable way of defining the buyer yourself. Where it hurts is when it is left to expand into audiences you have already paid to build, so that prospecting budget gets spent re-reaching your own website visitors at prospecting prices. Explicit exclusions matter more than ever, and they want checking monthly rather than at setup.
The practical rule: keep the geography tight, keep exclusions deliberate, let the creative do the qualifying, and stop trying to out-guess the delivery system on who is in the market.
Why your reporting under-counts what happened
If your tracking is a browser pixel and nothing else, your account is being measured through a narrowing window and the numbers in the dashboard are lower than reality.
Several forces push the same way. Apple's app tracking prompt means a large share of iPhone users decline to be tracked across apps, so a meaningful portion of conversions cannot be tied back to the advertisement that caused them. Browser restrictions on third-party cookies and shortened cookie lifetimes cut into it further. Ad blockers remove another slice. The result is systematic under-reporting, and it falls hardest on exactly the long-consideration jobs where the gap between the first advertisement and the enquiry is measured in weeks.
The Conversions API is the standard fix. Instead of relying solely on the visitor's browser to fire an event, your server, or your CRM, sends the conversion to Meta directly, matched on hashed customer information. It recovers a share of the events the browser loses, and because the optimisation system learns from the events it receives, better data produces better delivery rather than merely a better-looking report. Set it up alongside the pixel with proper deduplication, not instead of it.
Even then, treat platform-reported conversions as an indicator rather than a ledger. The number that matters is the one in your own system: enquiries received, quotes issued, jobs won, and where each of them said they came from. Our methodology page sets out how we reconcile the two, and the glossary covers the terminology if any of this is unfamiliar.
Lead quality, said plainly
Meta leads are usually worse than search leads. This is not a criticism of the platform, it is a direct consequence of how they were created. Somebody who typed your trade into a search engine had already decided they had a problem. Somebody who tapped a form during an advertisement break may have liked the photograph. Both are recorded identically as a lead.
Expect a lower proportion to answer the phone, a higher proportion to be out of area or out of scope, and more people who do not remember enquiring. Instant forms make this worse because a submission can take two taps with pre-filled details. That is precisely why they are cheap.
The fixes are known and all of them cost you volume on purpose. Add qualifying questions. Use the higher-intent form format with a review step. Send warmer audiences to a proper page instead. Most importantly, respond in minutes: speed to first contact does more for realised lead quality in this channel than any targeting change, because you are catching people while they still remember the advertisement. That is a CRM and automation problem more than an advertising one.
What the published numbers can and cannot tell you
You will be shown benchmark tables. Treat nearly all of them with suspicion.
The most widely circulated set is WordStream and LocaliQ's Facebook Ads benchmarks for 2025, covering 1 April 2024 to 30 June 2025. Across all industries it reports median figures of $27.66 cost per lead, a 7.72% conversion rate, $1.92 per click and a 2.59% click-through rate. Useful for a sense of scale, with two heavy caveats: the figures are American, in US dollars, from US auctions, and the leads-objective sample is only 726 campaigns spread across fifteen industries. Divided down, the industry-level rows rest on a handful of campaigns each, which is nowhere near enough to plan a budget against. Anybody quoting you a specific industry cost per lead from that dataset is overstating what it can support.
There is no published independent UK cost-per-lead benchmark by industry for Meta. Not a cautious one, not a partial one. Anything presented as such is either an agency's own client data or American figures relabelled.
What is solid is the market context. IAB UK put total UK digital advertising at £40.5bn in 2025, of which social accounted for £11.5bn, roughly twenty-eight per cent, having grown twenty-one per cent year on year. That tells you the channel is substantial and getting more competitive. It tells you nothing about what a driveway enquiry should cost in Doncaster. For that, your own first sixty days is the only benchmark with any authority, which is why we publish our working on the benchmarks page rather than a table of borrowed numbers.
If you are going to do it
Start with the honest decision. If your work is urgent and searched for, fund search properly first and leave this alone. The comparison between the two channels sets out the dividing lines trade by trade.
If your work is visible and deferred, go the other way and treat social as the primary channel, because there may be very little search demand to capture in the first place.
Either way, three things want settling before a pound is spent: who is taking the photographs and how often, whether the Conversions API is properly connected, and who answers an enquiry within ten minutes. Accounts fail on those three far more often than they fail on structure, and none of them are things an agency can do on your behalf.
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 long before we know whether it is working?
Six to eight weeks for a quick-turnaround trade, and a full quarter for anything with a long decision cycle such as glazing, kitchens or re-roofing. The common error is assessing at thirty days, when the enquiries generated in week two are still gathering quotes. If the channel is genuinely wrong for your trade, though, that is usually obvious within a month and you should stop rather than persevere.
Is it cheaper than Google Ads?
Clicks are cheaper. Customers are frequently not. A feed click can cost a fraction of a search click and be worth a fraction of it, because the person had no intention five seconds earlier. Compare cost per job won rather than cost per click or even cost per lead, and keep the two sources separated in your own records so the comparison is possible at all.
Can we just boost posts instead?
Boosting optimises for engagement, which is a different objective from enquiries and generally produces likes rather than work. A properly built campaign optimises for a conversion event, excludes existing customers, and retargets people who visited and did not call. The gap in cost per enquiry between the two approaches is usually large enough to fund the difference several times over.
Do we need Instagram as well as Facebook?
They are the same advertising system and in most trade accounts you let placements run across both rather than choosing. Instagram tends to carry more of the weight for visually strong work and a younger audience; Facebook tends to reach the older homeowner demographic that commissions most trade work. Restricting placements manually usually costs more than it saves.