Attribution for Home Service Contractors — What Actually Works in 2026.

Not directly. CAPI itself is free. Setting it up typically costs 1-3 hours of developer time OR is included in any competent agency's onboarding. Meta also offers a free 'CAPI Gateway' tool for contractors who don't have dev resources.

Use our Account Health Score tool — it checks Pixel, CAPI, AEM, and attribution window configuration. If you're missing any, fix those before optimizing further. Good attribution setup typically recovers 15-25% of 'lost' conversions overnight.

Yes — they're complementary, not alternatives. Pixel catches most events on non-iOS. CAPI fills the iOS + ad-blocker gap. AEM handles iOS for Meta's own optimization. All three should be on by default in 2026. If any are missing, you're losing data.

Three-step check: (1) Events Manager → Test Events tab, submit a test lead, confirm both 'Pixel' and 'Server' columns light up (proves Pixel + CAPI firing in parallel); (2) check Events Manager → Diagnostics for any flagged issues; (3) compare Meta-reported leads to your CRM-received leads weekly — the delta should be under 15%. If delta is 25%+, something's broken in CAPI or AEM. Fix before blaming campaigns for 'underperformance.'

Yes, similar but less severe. Google has Enhanced Conversions (their CAPI equivalent) and Consent Mode v2 to recover some signal. iOS users still under-report on Google by 10-20% without Enhanced Conversions, vs 20-30% on Meta without CAPI. The fix is identical in concept: pair browser-side tracking with server-side fallback to get attribution close to whole.

Three-layer phone attribution stack: (1) Track phone clicks as a Pixel event in Events Manager — anyone who taps your tel: link from a Meta ad fires a 'Contact' event, no special integration needed; (2) Use call tracking software (CallRail, Twilio) that assigns a unique phone number to each ad source — lets you match calls back to specific creatives; (3) Train your phone-answering team to ask 'How'd you find us today?' on every call and log the answer in your CRM. Meta + CallRail handles the click-to-call attribution cleanly; the verbal 'how'd you find us?' captures the multi-touch attribution Meta will never see (i.e. someone saw your ad 3 weeks ago, didn't click, but called organically when their AC broke). Without all three, you'll under-attribute Meta by 30-50% on any trade where 70%+ of conversions happen via phone.

Three new ones specific to this era: (1) OVER-RELYING on Meta's reported attribution — modeled-conversions in Ads Manager are 30-50% modeled (not actual events) for iOS users, but contractors treat them as true. Truth: cross-reference Meta's reported leads against CRM-received leads weekly; the delta is your attribution gap. (2) IGNORING incrementality — many contractors report Meta-attributed conversions that would have happened anyway via referrals or LSA. Run a quarterly 'pause test' — turn off Meta for 7 days, see what happens to total business. The drop is your TRUE Meta contribution. Most accounts find Meta's reported ROAS overstates by 15-30%. (3) UNDER-MEASURING phone calls — 60-70% of home service conversions happen via phone, but most contractor accounts have ZERO phone-call attribution set up. Without it, Meta looks unprofitable when it's actually carrying the business. Fix by adding CallRail dynamic numbers to all Meta-driven landing pages.

Yes, every 6 months minimum. Structure: (1) PICK A QUIET WEEK — not peak season, not a holiday week, not right after a major creative refresh. Mid-fall or mid-winter typically works; (2) PAUSE META ENTIRELY for 7 days — keep all other channels (LSA, organic, referrals) running normally; (3) MEASURE TOTAL leads + booked jobs that week vs the prior 4-week average. The drop = your TRUE Meta contribution. Common findings: contractors who report 'Meta drives 60% of leads' often find true incrementality is 30-45% (some leads would have arrived via other channels anyway); (4) RESUME META and you'll typically see a 7-14 day re-warm period where CPL is slightly elevated. Total cost of the test: roughly 25-40% of one week's expected revenue. Total benefit: knowing whether your Meta ROAS reports are accurate or inflated. Worth it once a year minimum. Most contractors avoid this test because they're afraid of what it'll show — but ignorance isn't bliss when you're spending $5K+/mo.

Neither alone — use both as complementary views. META Ads Manager: optimistic on Meta's contribution (last-click + view-through within Meta's data ecosystem; doesn't see what happened on Google or before the Meta touchpoint). Typical bias: Meta over-credits itself by 15-30%. GA4: pessimistic on Meta (uses data-driven attribution by default; tends to credit Search + Direct heavily because those are last-touch). Typical bias: GA4 under-credits Meta by 20-40%. The truth lives in between. Best practice: (1) pull Meta's reported leads from Ads Manager weekly; (2) pull GA4's 'Meta-attributed' leads from Acquisition reports same window; (3) average the two for a directional 'true Meta contribution' estimate. (4) cross-reference with your CRM source field (every lead asks 'how'd you find us?'). Three data points always beat one. Most contractors trust whichever number is highest, which is usually Meta's — and end up over-spending against inflated ROAS.

Match the attribution window to your trade's typical sales cycle. Three rules: (1) EMERGENCY services (plumbing leak, HVAC outage) — sales cycle is 1-3 days; use 1-day click + 1-day view. Tighter window catches the urgency signal without crediting view-through noise; (2) STANDARD home services (roofing repair, painting, pest control) — sales cycle is 1-3 weeks; use Meta's default 7-day click + 1-day view. Captures multi-touch journeys without going wide enough to credit unrelated awareness; (3) LONG-CYCLE projects (full roof replacement, kitchen remodels, solar) — sales cycle is 4-12 weeks; use 7-day click + 7-day view OR consider 28-day click for awareness campaigns. Don't pick the longest window automatically — wider attribution = more conversions credited but also more noise. Test by running the same audience under 1-day vs 7-day attribution for 30 days; compare reported CPL + close rate. If close rate is similar, the wider window is just noise; if close rate drops materially, the wider window is catching real multi-touch conversions.

The 'how'd you find us' survey is the cheapest, most accurate method. Six-step setup: (1) Add a required 'How did you hear about us?' dropdown on every form with options: Facebook/Instagram ad, Google search, Google LSA, referral, drive-by, other; (2) Train phone-answering staff to ask the same question on every inbound call + log it in your CRM; (3) Track the answers in a single spreadsheet column; (4) Tally weekly: leads-by-source vs spend-by-source; (5) Calculate true source-level cost-per-lead + cost-per-booked-job; (6) Compare to Meta's reported attribution. The 'survey ratio' between Meta-reported and self-reported is your true attribution gap. This costs $0 + 30 minutes setup. Beats every fancy automation when properly executed because it captures multi-touch journeys Meta can't see (someone saw your ad weeks ago, didn't click, then searched your name on Google later — Meta misses this; the survey catches it). Most contractors at $5K+/mo Meta spend report that the survey reveals 20-40% MORE Meta-attributed leads than Meta itself credits — once you know your account's gap, you can correct your reported ROAS.

Treat 'don't remember / heard somewhere' as multi-touch attribution evidence — it usually means MULTIPLE touchpoints, not zero. Industry data on contractor 'don't remember' answers: roughly 60% of those customers actually DID see your Meta ad at some point but it wasn't the deciding-touch — they remember the deciding-touch (a referral, a search, a sign) but Meta played a role in priming. How to model: when computing per-channel attribution, distribute 'don't remember' answers proportionally to your KNOWN attribution percentages. Example: if 30% know-they-saw-Meta + 40% know-came-from-Google + 30% don't-remember, model the 30% don't-remember as roughly 30% Meta + 40% Google + 30% organic/word-of-mouth. Meta typically benefits from being a top-of-funnel awareness driver that doesn't get last-touch credit. Don't IGNORE the 'don't remember' bucket — interpret it as proportional fuel for ALL paid channels you're running. Most contractors discount 'don't remember' answers entirely, which underestimates Meta's true contribution by 15-25% in multi-channel setups.

Phone-call attribution requires deliberate setup — Meta can't see this without help. Three-layer approach: (1) DYNAMIC NUMBER INSERTION (DNI) — call-tracking software (CallRail $50-100/mo) replaces your phone number on your landing page with a unique trackable number when traffic comes from Meta. The customer sees your normal number; CallRail's system attributes the call back to source. Setup: 30-60 minutes; (2) CALL CONVERSION events — pass call data back to Meta via CAPI as a Lead event when calls last >60 seconds (filters out wrong numbers + sales calls). Restores phone-attribution into Meta's optimization signal; (3) MANUAL TAGGING — train phone staff to ask 'how'd you find us' on every call + log answer in CRM source field. Combines DNI's automatic tagging with manual verification. Without all three, contractors lose 30-60% of phone-attribution capability — especially painful for emergency-service trades where 70%+ of conversions happen via phone call. Set this up BEFORE scaling Meta past $3K/mo; the attribution gaps compound exponentially as you scale.

Expect continued attribution degradation; build resilience now. Three trends to plan for: (1) STATE PRIVACY LAWS multiplying — California, Colorado, Virginia, Connecticut, and others have implemented their own GDPR-style regulations. By 2027, expect 15-20+ US states with similar requirements. Each restricts what data Meta can collect; cumulative effect: 5-10% additional attribution loss per state added; (2) BROWSER-LEVEL changes — Safari already blocks most third-party tracking; Chrome's Privacy Sandbox rolls out fully through 2026, replacing cookies with Topics API. Meta retargeting will work differently; expect retargeting CPL premiums of 10-20% as the transition completes; (3) CAPI MIGRATION accelerating — first-party data + server-side tracking become non-negotiable. Contractors not running CAPI by 2027 will be at a 30-40% attribution disadvantage vs competitors. ACTION: implement CAPI now, harvest first-party email + phone data aggressively, build retargeting pools while it's still cheap. The contractors who treat privacy regulation as an existential threat are the ones who fall behind; the ones who treat it as competitive opportunity (because most competitors won't adapt) gain market share.