MSO Marketing for Auto Repair: An Operating Model from 31 Instrumented Locations - Enterprise Digital Marketing

An operating model for auto repair MSO marketing built from live telemetry across 31 automotive service locations: what actually moves in the first 90 days after a marketing takeover, where the call-handling leak sits, which measurement layer to build first, and the portfolio economics underneath.

August 17, 2026 · 5 min read · By Nick Martinelli

Marketing advice for auto repair MSOs generally arrives from one of two sources: software vendors describing what their dashboard measures, or single-shop playbooks scaled up by assertion. This paper is built from a third source: the live telemetry of the 31 automotive service locations we manage, spanning single shops and operator groups of up to five stores, run on shared infrastructure and instrumented from ad click to repair order.

The evidence base:

What follows is the operating model that telemetry supports: four layers, in the order an MSO should build them, with the measured numbers that justify the ordering.

Layer 1: Demand is the first thing that moves, and the only honest early KPI is callers

When we take over a location's marketing, the earliest defensible signal is not revenue, not average repair order, and not platform conversions. It is unique inbound callers and first-time callers. Across ten onboarded locations, the first-90-day fingerprint was consistent: unique callers rose at six of ten (mean +10.6 percent) and first-time callers at five of ten (mean +8.7 percent), while repair-order count, sales, and average ticket were mixed around zero in the same window.

Two disciplines follow from that:

Ignore the click explosion. A campaign rebuild produces a click surge at essentially every takeover, often on less spend (one location: clicks +431 percent on 20 percent lower spend). That is a restructuring artifact, not an outcome. Any partner presenting click growth as early proof of performance is presenting the cost side of the ledger as the benefit side.

Track whether the store converts the demand. The same 90-day dataset splits cleanly into locations that turned caller growth into repair orders and locations where callers rose while repair orders fell. That divergence is not a marketing variable. It is the front counter, which is layer 2.

Layer 2: Call handling is a larger lever than media at most locations

At one four-location operator group we studied over six months: 11,867 inbound calls, 87.2 percent answered. The 12.8 percent miss was not random; it was structural, concentrated at the edges of the day (as low as 60 percent answered at 7am, 28 to 100 percent missed after 6pm) and on weekends (Saturday 38.9 percent answered). Within the window, 395 callers abandoned without ever being recovered, 339 of them first-time callers, worth an estimated $66,000 to $122,000 in lost repair-order revenue at that group's tickets.

For an MSO, this is the uncomfortable arithmetic: a media optimization that improves cost per lead by 10 percent is worth less than answering Saturday's phone. Marketing spend that lands on an unanswered line is not underperforming marketing; it is a staffing decision wearing a marketing budget. We treat answered rate, by hour and by day, as a standing report at the same altitude as ROAS, and our 90-day onboarding data shows answered rate does not improve on its own: it was essentially unchanged at eight of ten locations post-takeover, because nobody was managing it.

Layer 3: Measurement, in a specific order

The auto repair stack has a broken join in the middle by default. Our audit found a Google click ID survived into the call record in 0 of 11,867 tracked calls, and the default interaction-to-revenue link rate across 1.45 million records rounded to zero. The full mechanics, and the hashed-identity architecture that closes the loop anyway, are in the closed-loop attribution paper; the sequencing argument for which KPI to run while that build completes is in the net revenue lift paper.

The order that works at MSO scale:

1. Net revenue lift first. Same-store, same-months, year over year, from the POS the operator already trusts. Available on day one, no integrations. Across the mature book this KPI measured a 13 percent lift, roughly $3M against ~$500K of media cost.
2. Caller telemetry second. Deduplicated unique and first-time callers per location, plus answered rate. This is the leading indicator layer.
3. Closed-loop feeds third. Completed-transaction revenue uploaded to the ad platform against hashed identity. This is the end state; live builds in the book run at 7.1x measured ROAS floor at one location and carried $438,398 of repair-order revenue in a first scheduled upload at one operator group.
4. Never desk tags for channel math. The employee-typed marketing-source field undercounts digital by an order of magnitude at stores where tagging discipline has decayed, which in our audits is most stores.

Layer 4: The economics only work if the operating model matches the location count

Multi-location marketing costs do not decline smoothly with scale. The driver-based model we built from this book's actual labor consumption shows a sweet spot around nine locations, a structural dead zone from roughly 10 to 59 locations where coordination overhead outruns shared-services savings, and enterprise economics only past that. The mechanics are in the cost-to-serve paper, and the pricing consequences (floors, not rates) in the pricing paper.

For an auto repair MSO evaluating partners, the practical test is whether the proposal prices the coordination layer at all: cross-location budget arbitration for sibling stores in overlapping markets, per-location attribution maintenance, and reporting that reconciles to the POS. Proposals priced as N times a single-shop retainer have not modeled the actual work, and the gap surfaces later as either quality decay or renegotiation.

What this model does not claim

Honesty about limits is cheap insurance. Our own cohort analysis shows the onboarded stores grew 9.9 percent year over year on revenue, with eight of nine stores up, but growth was roughly in line with comparable stores in the book over the same window. A strong automotive year lifts everything; the marketing-specific claims we stand behind are the ones this paper is built on: demand telemetry, call-handling exposure, and closed-loop traced revenue. The full measurement protocol, including the control comparison, is in the revenue lift measurement paper.