Sheppard

Value Creation

Marketing attribution for home services: connecting spend to booked revenue

Last-click attribution breaks down the moment a home services platform crosses three channels and two CRMs. The fix isn't a better dashboard — it's a reference architecture that ties GA4, CallRail, ServiceTitan, and ad-platform data into one source of truth your CFO can reconcile.

By Chris SheppardApril 29, 202611 min read

Ask any operating partner what's hardest about marketing in home services and they'll likely name attribution. Ad platforms claim credit. The CRM claims credit. CallRail claims credit. ServiceTitan claims credit. Add them together and you've doubled — sometimes tripled — the actual sourced-revenue number.

The fix isn't a better dashboard. It's a reference architecture that ties GA4, CallRail, ServiceTitan, and ad-platform data into one source of truth that the CFO can reconcile and the sponsor can audit.

Why attribution is harder in home services than in e-commerce

E-commerce closes the loop in the browser. Home services closes the loop weeks later, on a tech's tablet, in a customer's driveway, after a phone call that may or may not have been routed through call tracking. The conversion event isn't a button click — it's a service delivered, and the data lives in ServiceTitan or its equivalent, not in GA4.

The four sources of truth — and why they disagree

Every home services platform has at least four claim-systems for marketing attribution: GA4 (sees the website session), CallRail or WhatConverts (sees the inbound call), ServiceTitan or the dispatch CRM (sees the booked job and revenue), and the ad platforms (claim conversions based on their own tracking). Each uses a different attribution window, a different tracking identifier, and a different definition of 'conversion.' They will never agree. The job is to pick one as the system of record and reconcile the others against it.

A reference architecture

  1. Tracking IDs: GCLID for Google Ads, FBCLID for Meta, GBRAID/WBRAID for app campaigns — passed as URL parameters and captured in the CRM
  2. Dynamic Number Insertion (DNI) on the website to attribute calls to the originating session and channel
  3. UTM standardization with a documented schema (source, medium, campaign, content, term) — every paid placement uses it
  4. ServiceTitan as the system of record for booked revenue, with tracking IDs joined back to lead records
  5. A weekly export from ServiceTitan to a data warehouse (BigQuery, Snowflake, Postgres) where channel-level attribution is reconciled to financial system actuals

Choosing an attribution model

Last-click is misleading at scale because it over-credits bottom-of-funnel channels (paid brand search, retargeting) at the expense of top-of-funnel discovery (organic search, paid social). Multi-touch is better but expensive to implement well. Marketing mix modeling (MMM) is the gold standard for sponsors who underwrite multi-channel platforms but requires 18-24 months of historical data. Most home services platforms should run last-click for in-period decisions and MMM (or a position-based multi-touch) for board-level reporting.

Reconciling booked revenue to spend at month-end close

Marketing-sourced revenue should reconcile to ServiceTitan booked revenue and ServiceTitan booked revenue should reconcile to financial system invoiced revenue. Three reconciliation passes per month: marketing-to-CRM (channel attribution), CRM-to-revenue (booked vs. invoiced), and revenue-to-finance (invoiced vs. recognized). Each break needs a documented owner and a closure cadence.

Common attribution leaks

  • Form fills that create CRM records without tracking IDs — usually because the form vendor and the website don't share session data
  • Branded calls (someone Googles the brand and clicks the GBP phone number) credited to organic when they're really word-of-mouth
  • Repeat customers attributed to whichever channel touched them most recently, instead of credited to retention
  • Walk-ins and field-sales bookings that bypass the marketing funnel entirely — should be tagged as 'unattributable' instead of being defaulted to 'organic'

What 'good' looks like — the sponsor-facing dashboard

One page. Sourced revenue and spend by channel for the month and trailing 12. Cost-per-booked-call by channel. CAC by acquisition source, normalized for service mix. Attribution methodology documented in a footnote, with the audit trail of any methodology changes. The sponsor should be able to read it in four minutes and trust it for the rest of the quarter.

Attribution is one of the few marketing investments where the work is invisible until the audit. Build it before the buyer's diligence team asks.

Frequently Asked

More on value creation.

What's the right attribution model for a multi-channel home services business?

+
Last-click for in-period operating decisions, multi-touch or position-based for monthly board reporting, marketing mix modeling (MMM) for annual planning and exit prep. Each model serves a different decision cadence. Sponsors should expect the platform to use all three appropriately.

How do you tie ServiceTitan revenue back to specific ad campaigns?

+
Pass tracking IDs (GCLID, FBCLID, UTMs) through the website, capture them in lead records, and join them to booked revenue in ServiceTitan. Weekly export to a data warehouse where channel-level attribution reconciles to financial system actuals. The integration work is non-trivial but one-time; the reporting compounds for the rest of the hold.

Why is last-click attribution misleading in home services?

+
It over-credits bottom-of-funnel channels (paid brand search, retargeting) at the expense of top-of-funnel discovery (organic search, paid social, content). At scale this leads to under-investment in brand and demand-generation channels — which is exactly the gap that buyers diligence at exit.

How do you handle attribution for offline conversions?

+
Tag them explicitly as 'walk-in,' 'field sales,' 'referral,' or 'unattributable' rather than defaulting them to 'organic' or 'direct.' Word-of-mouth and repeat customers should also be tagged as separate categories — they distort CAC and LTV calculations when bucketed with paid channels.

What does a working marketing data flow look like?

+
Tracking IDs flow from ad platforms → website → form/call tracking → CRM (ServiceTitan or equivalent) → data warehouse → sponsor dashboard. ServiceTitan is the system of record for booked revenue. The data warehouse is where channel attribution reconciles to financial actuals. The dashboard is what the sponsor sees monthly.

Engage Sheppard

Have a deal that needs this work?

Pre-LOI, post-close, mid-hold, or pre-exit — the conversation starts with five questions and fifteen minutes on the calendar.