MSO Local Search: Telemetry from 36 Local Services Ads Enrollments - Enterprise Digital Marketing

What local search actually produces at MSO scale: portfolio benchmarks from 36 Local Services Ads enrollments (blended CPL, pacing, producing share), a profit-verified LSA ROI measured to the POS, the local-actions measurement trap, and why branded gains are usually capture rather than demand.

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

Local search is where multi-location marketing claims are least audited. Every location has a business profile, most have some form of Local Services Ads, and the reporting for both arrives pre-aggregated by the platform with no revenue attached. This paper reports what those surfaces actually produce across the 31 automotive service locations we manage, including a portfolio-wide Local Services Ads sweep and one location where we matched LSA leads all the way to point-of-sale revenue.

The MSO-relevant finding up front: at portfolio scale, local search performance is dominated by administrative state and measurement scope, not by auction tactics. The stores that underperform are rarely losing auctions. They are unenrolled, paused, unpaid, or mis-measured.

Local Services Ads at portfolio scale: the constraint is administrative

We sweep every Local Services Ads enrollment under management on a recurring basis, because the platform's own API surfaces only a fraction of them and no consolidated view exists. A representative full-portfolio sweep:

That last row is the one that matters. The portfolio was spending at roughly half its budgeted weekly rate, and the shortfall decomposed almost entirely into administrative failure states, not weak auction volume: a payment method that could not be charged (account dark for days), an insurance document lapsed in re-verification, accounts paused and forgotten, and one abandoned enrollment shell. A third of enrollments producing zero leads is the norm we observe, not the exception, and none of it is visible unless someone reads every account, because LSA sends no consolidated alert.

For an MSO the implication is blunt: before any conversation about LSA optimization, the highest-yield work is an enrollment census. Verified, funded, unpaused, documents current, at every location. It is unglamorous, and in our experience it moves more lead volume than any bidding change.

LSA unit economics, measured to the POS at one location

Platform lead counts say nothing about profit, so at one single-location shop we matched the LSA lead inbox against the shop's management system by normalized phone number, with an exact-name fallback:

• 249 leads over the audit window at a $34.74 cost per charged lead
• 83 leads matched to a customer record (33 percent, a floor: unmatched is not proven non-customer)
• 43 matched leads reached a completed repair order within 90 days, worth $48,898
• At the shop's parts-and-labor margin, that is a 4.4x return on LSA spend measured in gross profit, with a strictest-case floor of 2.1x counting net-new customers only
• 90-day value of a net-new LSA customer: $1,184, roughly four times the average ticket of the shop's direct-mail-sourced customers in the same system

Two operational notes travel with those numbers. First, the return depends on lead hygiene: this operator actively disputes junk leads (wrong service, robocalls), and 71 of the 249 were never charged or were credited. Second, the match is name-and-phone based, so every figure above is a floor by construction. The methodology matters more than the multiple; an LSA ROI claim with no POS join behind it is a lead count wearing a dollar sign. The join mechanics are the same ones covered in the closed-loop attribution paper.

The local-actions measurement trap

The second local-search surface, the business profile and its "local actions" (calls from the profile, direction requests, store-visit estimates), is where multi-location reporting most reliably inflates. The distortion is not fraud; it is scope.

At one location we audited across an agency transition, counted conversions rose from roughly 25 per month to 163 per month. On inspection, 80 to 85 percent of that rise was a change in which action types were included in the counted column: profile calls, direction requests, and modeled store visits had been tracked all along but excluded from the previous agency's counting. Apples to apples, on the one metric defined identically in both eras, ad-driven phone calls went from 47 to 85 per month on 16 percent less spend. Real improvement, roughly 2x, against a headline that implied 6x. The same account carried $491,000 in platform-reported conversion value that was 99.6 percent modeled store visits at a default $100 apiece.

The MSO defense is a standing rule: any conversion trend that crosses a vendor transition, a tagging change, or a counting change is invalid until re-based on a metric defined identically on both sides. Local actions are the most common place that rule gets violated, because adding them to the counted column is one checkbox and it only ever moves the number up.

Branded search: capture is not demand

The last local surface is the branded query space around each location's name. It produces the most seductive chart in local marketing: branded clicks climbing month over month. At the location above, branded organic clicks tripled. Branded impressions stayed flat, at roughly 600 per month, across the entire period.

That distinction settles what a marketing program can and cannot claim. Flat impressions mean brand demand did not grow; the program got dramatically better at capturing the demand that already existed (overall organic clicks rose 66 percent as average position improved from 29 to 13). Capture efficiency is real money. But once capture saturates, more search budget cannot manufacture branded demand, because search is downstream of it. At that point the growth lever moves off-search entirely: reputation, retention, and awareness channels. A local search program that cannot tell you whether its branded gains are capture or demand cannot tell you when it has hit that wall.

Operating rules for MSO local search

1. Census before optimization: every enrollment verified, funded, unpaused, documents current. Re-check on a schedule; failure states are silent.
2. Price LSA against POS-joined gross profit, not platform lead counts, and dispute junk leads as a standing process.
3. Freeze conversion definitions across vendor and tagging transitions; re-base any trend that crosses one.
4. Split branded reporting into impressions (demand) and clicks (capture), and claim only the one the data supports.
5. Treat the profile-call, direction, and store-visit numbers as directional telemetry, never as valued conversions.

The wider measurement architecture these rules plug into, and the portfolio evidence behind it, are in the auto repair MSO operating model and the portfolio teardown.