A facilities manager in Reading watches a 12-second reel of a warehouse NVR rack, then taps through. That is not the same person as a landlord in Croydon who just saw a Ring-style doorbell montage and wants a kit they can hang themselves. If your Meta account treats both as “security leads,” you are paying for scrollers, not for the van that needs a survey booked in pounds, not likes.
UK alarm and CCTV companies sit in an awkward auction. Search still captures “CCTV installer near me.” Meta captures the moment after a neighbour’s incident, a commercial insurance reminder, or a late-night property group thread. Industry aggregators such as WordStream and Triple Whale have put blended Meta ROAS in the 2-3x range versus 3.5-5x for Google Ads - but for trades businesses optimising on booked surveys rather than last-click revenue, these numbers are largely decorative. That gap is not a reason to abandon the channel. It is a reason to stop judging Meta like a last-click search engine and start treating creative, offers, and post-click proof as the targeting.
This piece is for marketing owners inside installation firms, not for homeowners shopping Amazon cameras. It covers how Advantage+ and broad delivery actually behave, which conversions you should feed the algorithm, and the UK-specific friction (NSI / SSAIB language, domestic vs commercial jobs, GDPR-sensitive CAPI) that decides whether spend becomes surveys.
Where UK Meta spend quietly funds DIY kits instead of engineer surveys
You will feel it first in CPM, not in lead volume. Creative fatigue in 2026 shows up as rising costs while the lead counter still looks “fine.” UGC-style clips often peak in 5-6 days. Ads that looked solid on Monday can be junk by Friday even if the dashboard still reports cheap form fills. Plenty of operators still read that as “Meta is volatile” and slash budget. The more useful read: your hook is attracting the wrong job type, and the pixel is celebrating it.
Domestic doorbell interest is cheap and noisy. Commercial multi-camera, access control, and monitored alarm work is slower and worth more. If one campaign, one landing page, and one “Get a quote” event sit over both, Advantage+ will hunt the easier conversion. Broad targeting is not the villain. Weak creative and a homepage form are. The 2020 playbook of stacked lookalikes and interest piles no longer moves the needle the way a systematic creative engine does. “Creative is your targeting” is not a slogan. In Andromeda-driven auctions, the first two seconds decide whether you get a facilities buyer or a weekend DIYer.
Attribution makes the argument worse. Teams lose 20%+ of conversion data to ad blockers and privacy changes. Meta’s in-platform CPA then diverges from what your survey diary shows. Last-click models double-count when someone also Googled your brand after the reel. You cannot fix that with a tighter interest list. You fix the signal: which event means an engineer is actually going out, and which page matches the ad they tapped.
A typical miss: a Midlands installer ran one Advantage+ campaign to the homepage, optimising for Instant Form submits. CPL looked respectable in pounds. The ops board filled with “need a camera for the shed” notes. When they split domestic monitored alarms from commercial CCTV, and only fired the purchase-quality event after a calendar booking, volume dropped and survey rate rose. That is the trade you want. A Leeds-based NSI installer told us CPMs climbed on day four while Instant Form CPL still looked cheap, until they stopped optimising for the form and CAPI fired on booked surveys rather than page views - volume dropped, survey rate rose. If your Meta traffic still lands on a generic homepage, that is the first leak to close, not another audience layer. HeyLead’s Meta Ads programmes are built around that split, not around chasing cheaper form volume.
Get a free marketing audit - we review your search, ads, and landing pages and send back what to fix first.
Get a free audit
How UK alarm and CCTV firms should structure Meta when creative is the targeting
Run fewer campaigns. Give each one a job type, a geography that matches where you actually install, and a conversion that ops will recognise. Greater London CPMs will not behave like a North East patch. Do not merge them to “get more data.” You will get blended nonsense. Keep Advantage+ Shopping or sales-style objectives only if you truly sell kits online. Most NSI / SSAIB installers should run leads or appointments with a quality event downstream of the form.
Copy and creative have to name the work. “Protect your home” is wallpaper. “Monitored Grade 2 alarm, engineer survey in Birmingham within 5 working days” is a filter. Commercial ads should show rack rooms, door controllers, and night-time loading bays, not a smiling family in a hallway. Offers that work: free site survey (not a free camera), insurance-ready spec, existing system takeover, 4K retrofit on a working NVR. Price-led “from £99” creatives fill the inbox with comparison shoppers you will never win against the multiples.
Audit scorecard
-
Job-type splitSeparate domestic monitored alarms, domestic CCTV, and com
Job-type splitSeparate domestic monitored alarms, domestic CCTV, and commercial / multi-site work. One ad set per job if volume is thin. Do not let a doorbell creative train the commercial campaign.
-
UK geo that matches the vanRadius or postcode clusters around branches y
UK geo that matches the vanRadius or postcode clusters around branches you can actually attend. Exclude regions you only quote at a loss. Meta will happily spend in Cornwall if your creative is strong and your installers sit in Leeds.
-
Broad delivery, strict creativeAdvantage+ and broad targeting outperform
Broad delivery, strict creativeAdvantage+ and broad targeting outperform 2024 audience stacking when hooks are specific. Do not “help” the algorithm with six overlapping interest piles. Help it with proof: NSI, SSAIB, insurance certificates, engineer on a ladder, control-room stills.
-
Event you optimiseFire a secondary event when a survey is booked or when
Event you optimiseFire a secondary event when a survey is booked or when CRM marks the lead as qualified. Do not optimise solely on Instant Form start. Event Match Quality is a hygiene check, not a profit forecast. EMQ scores do not reliably predict whether the job pays.
-
Creative rotation clockPlan replacements before CPM climbs
Creative rotation clockPlan replacements before CPM climbs. If a UGC clip peaked in under a week, have the next three variants (hook, offer, proof) already shot. Ad-hoc testing is why learning never compounds.
-
Policy and claimsDo not imply the police will arrive
Policy and claimsDo not imply the police will arrive. Avoid fear montage of children. Stick to property, commercial duty of care, and factual spec. UK financial and crime-adjacent claims get restricted fast.
-
Remarketing without a mazeMeta already routes a meaningful slice of broa
Remarketing without a mazeMeta already routes a meaningful slice of broad spend toward people who engaged. A single remarketing campaign with survey proof is enough. Do not build a 12-ad-set funnel that resets learning every time you tinker.
-
Budget patienceDaily tweaks reset learning
Budget patienceDaily tweaks reset learning. Low spend plus daily optimisation is why Meta looks “unreliable.” Give a job-type campaign a full learning window before you declare Advantage+ a failure.
Action checklist
- Export 90 days of Meta leads and tag each row: survey booked, quoted, won, DIY / no-show. That file is your real CPA, not Ads Manager.
- Kill any ad whose first three seconds could sell a consumer camera kit. Replace with engineer-on-site footage and a clear “survey, not a box” line.
- Build two dedicated landing pages: domestic monitored alarm and commercial CCTV. Match headline to the ad. Put NSI / SSAIB, response SLA, and a short form above the fold.
- Pass UTMs on every ad. Standardise source / medium / campaign / content so CRM can tell a Manchester warehouse reel from a Croydon landlord static.
- Send Conversions API events from the server for booked surveys. Browser pixel alone will keep leaking 20%+ of the signal you need for bidding.
- Check landing speed. Google's CrUX data consistently shows fewer than half of mobile origins pass all three Core Web Vitals - check yours before scaling spend. A slow survey form on mobile (most Meta traffic) trains people to bounce, then trains Meta to find more bouncers.
- Refresh creative on a 5-7 day cycle for winning UGC, longer for proof-heavy statics. Do not wait for lead volume to collapse.
- Hold structure still for a learning phase. Change one variable: hook, offer, or page, not all three.
Free tools - try these yourself
After the click, session recordings matter more than another interest. HeyLead Insights shows whether people stall on the accreditation block, abandon the postcode field, or never see the survey CTA on a phone. That is cheaper than guessing why CPL in Ads Manager and surveys in the diary refuse to match.
What a survey-grade Meta landing page actually has to prove
If they say they just send traffic to the homepage, you already know the outcome. Dedicated pages should repeat the ad’s job type in the H1, show a real install (not stock hallway), name the coverage area, and ask only for what dispatch needs: postcode, property type, existing system, preferred survey window. Long “tell us your life story” forms look thorough and kill mobile conversion. Mobile already converts worse than desktop by a wide margin on paid traffic. Do not add friction for sport.
Social proof that moves installers: named commercial sites you are allowed to show, engineer count, years NSI / SSAIB, response time for monitored alarms. A 4.9-star badge with no review platform, no reviewer name, and no job type named reads as invented. Link to your Google Business profile reviews or name the platform - Checkatrade, Which? Trusted Traders, or your NSI certificate page. Open Graph images should match the creative so the share preview does not swap a warehouse NVR for a smiling couple. Run the Open Graph preview before you scale a winner.
A South London commercial CCTV firm cut Instant Forms entirely after noticing 11 of 14 “hot” leads had no site access notes and never answered. They moved to a page with a 4-field form, a map of boroughs they cover, and a calendar. Volume fell. Booked surveys did not. That is the only ratio the owner should defend in a board pack. Tighter positioning and a clearer offer beat adding TikTok “because Meta is tired.” Adding platforms will not rescue a funnel that celebrates junk events.
Measurement should be weekly and ugly, not a monthly PDF. Line up Meta reported leads, CRM surveys, and jobs won. When those three disagree, fix tracking before you raise budget. Automated bidding will spend into a bad event with great confidence. That objection is fair: bidding inflates budgets when the conversion is a form start. It does not mean you should go fully manual forever. It means the event has to be the survey.
FAQs
Should UK alarm companies use Instant Forms or a website form?
Use Instant Forms only if every field maps into CRM and someone phones within minutes. For survey-led work, a fast dedicated page usually produces fewer, better enquiries because you can show accreditation and coverage before they submit.
Is Advantage+ too broad for NSI / SSAIB installers?
Broad is fine when creative and the landing page filter job type. Broad with a generic “security” video and a homepage form will buy doorbell shoppers. The risk is weak assets, not the delivery type.
How fast should we rotate Meta creative?
Treat short UGC as perishable (often days, not weeks). Proof-heavy statics and engineer footage last longer. Watch CPM and thumb-stop rate before you wait for lead volume to fall.
What ROAS should we expect versus Google Ads?
Directional 2026 benchmarks put Meta near 2.8x average ROAS and Google Ads near 4.2x. For installers, judge Meta on cost per booked survey and won jobs, not on last-click ROAS copied from Search.
Do we still need lookalike audiences?
As a primary targeting system, they are weaker than they were. Feed high-quality survey conversions, run broad, and let creative select the buyer. Keep a simple remarketing set for people who opened the survey page and left.
Putting it to work
Execution sprint
This week
- Pull 30-90 days of performance for UK alarm and CCTV installation firms (Search Console, ads, CRM, or call logs - whatever you have).
- Flag the top leak: wrong intent, weak page, slow response, or dirty conversion tracking.
- Ship one fix on the highest-traffic money path (page, campaign split, or response rule).
- Run the free tools below on that same URL or account and log the findings.
Free tools for this sprint
Next 30 days
- Expand the fix to the next one or two money paths only after the first one shows cleaner bookings.
- Align creative, keywords, or content with the same offer the page now states.
- Review booked outcomes weekly; cut anything that still only produces unqualified volume.
This week
-
Tag last 60 days of Meta leads as survey booked, quoted, won, or junk, then compute CPA on surveys only.
-
Pause ads whose hook could sell a retail camera kit.
-
Ship two message-matched pages (domestic alarm vs commercial CCTV) and rebuild UTMs with the UTM link builder.
-
Point bidding at booked-survey events, not form starts.
-
Run Core Web Vitals on both pages on a real phone.
-
Queue three replacement creatives before the current winners hit day five.
Next 30 days
-
Turn on Conversions API for survey bookings and reconcile weekly against the diary.
-
Split spend by branch geography if you have more than one van patch.
-
Use session behaviour to fix the highest-abandon field on the form.
-
Hold campaign structure still through a full learning window while you only rotate creative.
Pull your last 60 days of Meta leads and score them against booked surveys before you touch another interest list. If the gap is the handoff between a 12-second reel and a survey form that never states the job type, HeyLead can own that Meta creative, landing-page, and conversion-signal loop so your installers see diary slots, not DIY kit chatter. Chat with us on WhatsApp
Free marketing audit, or reach Martin directly:
Get a free audit Connect on WhatsApp · [email protected]