A dealer group finally saw which ads sell cars after fixing cross-domain tracking across eleven rooftops

Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.

Where they started

The group sells and services vehicles across eleven rooftops spread over three states, each operating its own site on a vendor-supplied template and its own local ad budget. Showroom traffic, test drives, and phone calls are the currencies of the business, and the group reports monthly into manufacturer co-op programs with fixed definitions. Before us, the marketing team ran ads from eleven sets of platform reports and reconciled them by hand in a spreadsheet the CFO's analyst maintained. Nobody could follow a person from an ad click to a booked drive without losing them at the scheduling domain.

What it was costing

Each rooftop's site sat on a different vendor template, test drives booked on a shared scheduling domain, and GA4 saw eleven disconnected properties — so the group couldn't tell which campaigns produced showroom visits, and ad spend was argued from dealer-platform reports nobody trusted.

What they could see

• The monthly ad-spend review argued from dealer-platform reports that the general managers did not believe and could not check.
• Questions like which spring campaign filled the weekend drive calendar got answered with last-click screenshots instead of journeys.
• Eleven properties in GA4 meant eleven dashboards; comparing rooftops meant opening tabs, not reading a table.
• Test-drive bookings on the shared scheduler showed up as separate sessions with no memory of the rooftop site that sent them.
• Manufacturer co-op reports and internal reports disagreed, and the group had no way to say which was wrong.

The constraints we worked inside

• Vendor templates lock large parts of the page markup; tagging must work through a tag manager with what the templates allow.
• The scheduling domain is shared across rooftops; sessions must carry attribution across it without leaking one store's traffic into another.
• The group reports monthly to manufacturer programs with fixed definitions — measurement must match those definitions, not reinvent them.

What had been tried before

The media buyer had stitched the eleven properties into one Looker Studio blend so the leadership deck showed a single table.

Blending side-by-side properties can put numbers in one row but cannot stitch a journey; a click on one site and a drive on the scheduler stayed two anonymous strangers.

An outside agency standardized UTM naming across rooftops and pushed everyone to read the platforms' own attribution columns.

Each ad platform grades its own homework, so the columns disagreed with each other and with the scheduler's counts — and none of them matched the manufacturer program's definitions.

The group once paid for call tracking at two of the larger stores to see whether phones could be counted at all.

It counted calls but not context — no link to campaigns or to the rooftops' sites — and the per-seat cost never spread past the pilot stores.

What we proposed

We proposed collapsing eleven properties into one GA4 property with a store dimension and cross-domain measurement to the scheduler, so an ad touch on a rooftop site and a booked drive stayed one journey. Conversions — test-drive requests, calls, brochure downloads — would be defined once, on the manufacturer programs' definitions, and documented so monthly co-op reporting stopped being an argument. The group's real question is which campaigns sell cars, and only a single property with honest attribution and agreed definitions can answer it; everything else is reconciliation theater. Trust came first: events, consent coverage, and internal-traffic filtering would be audited against real traffic before anyone argued from the numbers.

Just as important is what we ruled out, and why:

• Eleven properties joined by dashboard blends — It preserves each vendor template's independence but still cannot stitch a journey across the scheduler, and every template change silently rewrites what a property counts.
• A warehouse-only identity-stitching build — It would have handled the scheduler hop in theory, but there was no first-party login across rooftop sites to anchor identities, so the joins would have been guesswork.
• Upgrading every rooftop to a unified vendor template — Tempting for tagging, but template procurement runs on the vendors' roadmaps and would consume the whole engagement without improving a single measurement definition.

How the work ran

Delivered by the growth pod — analytics specialist over 6 weeks, with working increments reviewed with the client every week.

The stack, and the reasoning

GA4

One property with a store dimension answers cross-rooftop questions directly; eleven separate properties would have kept every comparison manual and every journey broken at the scheduler.

Google Tag Manager

Vendor templates lock most markup, so a tag manager is the only sanctioned way to deploy consistent tagging without asking eleven vendors to change their code.

BigQuery export

Manufacturer co-op reporting needs fixed, repeatable definitions; the raw export lets the group rebuild any report from history when a program changes its rules.

Looker Studio

Store managers and the group's leadership already live in browser tabs, not warehouses; a shared dashboard per rooftop kept adoption high without new seats or licenses.

Consent Mode

Rooftops sell across state privacy regimes; Consent Mode let tagging respect declining visitors consistently while preserving the co-op reports the manufacturer definitions require.

What went wrong

How we worked together

Cadence

A 30-minute call each Tuesday with the group marketing director and one rotating general manager; written audit findings landed the same day for anyone who missed it.

Client side

The marketing director owned scope and sign-off; each rooftop's digital coordinator supplied template access and local campaign context; the CFO's analyst reviewed the definition sheet.

Decisions

Tagging calls we could demo, we made; anything touching manufacturer definitions or rooftop budgets went to the marketing director, who answered within a business day.

They provided

Admin access to each rooftop's tag manager, a test booking path on the scheduler, the manufacturer programs' definition documents, and one hour weekly from the coordinators.

What changed

The headline: single property reporting campaign-to-test-drive journeys per rooftop — 11 disconnected properties → 1 with attribution, read from GA4 conversion paths. A second check: rooftops reporting with agreed conversion definitions at 100%.

The monthly spend review changed character. Instead of relitigating whose platform counted what, the meeting started from one property's journeys and spent its time on budget moves — which rooftops were under-fed, which campaigns filled weekend drives. The general managers stopped requesting their own ad-platform exports because the shared dashboard answered them first, and the argument finally moved from whose numbers were right to which rooftops deserved the next dollar.

The result was read from GA4 conversion paths against the pre-engagement baseline over the stated window, with a guardrail check on rooftops reporting with agreed conversion definitions. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.

What they own now

• The consolidated GA4 property, its cross-domain settings, and the store dimension mapped to all eleven rooftops.
• A written conversion definition sheet matching each manufacturer program's reporting language.
• The migrated tag manager container with a naming convention and per-rooftop folders.
• The Looker Studio dashboard with store, region, and chain views, owned by the group's marketing team.
• BigQuery export access plus the audit checklist used to verify events before any future template change.