Why Are Used-Car Results Diverging Across Public Dealer Groups?
The quarterly earnings cycle for the publicly traded dealer groups just finished, and the used-car numbers told very different stories depending on where you looked. That gap is real, and the trade press has noticed.
Auto Remarketing reported in August 2026 that used vehicles offered one of the clearest examples of divergence in the quarter among the publicly traded groups. The headline language was direct about what happened across the six public groups:
"Used vehicles offered one of the clearest examples of divergence in the quarter among the publics." — Auto Remarketing
The industry's first move when a gap like this appears is to look at sourcing, at inventory mix, at which groups had access to cheaper auction units and which were stuck overpaying. That explanation has surface logic. But it also conveniently lets the marketing operation off the hook. The groups that diverged upward were not necessarily running better inventory. In many cases they were running better measurement, and measurement depth is the advantage that compounds invisibly until the earnings call makes it visible.
Why Does the Industry Default to Inventory as the Explanation?
Inventory is a comfortable explanation for used-car performance gaps because it is concrete, it is visible, and it has always mattered. The group that sourced well wins. That story is true as far as it goes.
It stops being sufficient the moment you ask a second question: if two groups are running similar inventory in overlapping markets, priced within a comparable band, why does one convert at a higher rate? Sourcing cannot answer that. Neither can the usual responses: more budget, better creative, stronger brand. Those are real levers, but they are not levers you can pull without knowing which vehicles are actually moving in the funnel and which are sitting on spend without producing VDP engagement or leads.
The inventory explanation also maps to how most dealer groups still build their used campaigns. A single ad set targets "used cars" as a category. The budget flows to that category. When results are strong, the category gets credit. When results are weak, the category gets the blame and the conversation turns to what the acquisition team paid at auction last month. The campaign structure itself provides no granularity that would let anyone draw a different conclusion. The measurement surface is designed to produce inventory explanations, so inventory explanations are what it produces.
This is not a sourcing problem. It is a measurement architecture problem. And the groups that are pulling ahead understand that difference, even if they would not describe it in those terms.
What Does It Actually Mean to Measure a Used-Car Campaign at the VIN Level?
Most dealers can tell you how their used-car campaign is performing. They can pull impressions, clicks, cost per click, and platform-reported conversions. What most cannot tell you is which specific units those conversions are attached to, which VINs are drawing VDP engagement without generating leads, and which vehicles have been on spend for three weeks without a single qualified interaction.
VIN-level measurement is the practice of answering those questions. It means connecting ad platform performance data back to individual vehicle records, so the unit on the lot has an attributed funnel, not just the campaign that points to the lot in aggregate. Getting server-side tagging installed is a start, but it does not automatically produce VIN-attributed attribution because the tag still needs to know which vehicle the converting session was looking at.
When you have VIN-level attribution, you can answer questions that category-level campaigns structurally cannot. Is this unit converting at a rate that justifies its current budget allocation? Is the problem with this VIN a price objection (high engagement, low conversion, no lead) or a visibility gap (low impressions relative to similar units on the same lot)? Is there a markdown effect visible: does engagement increase within days of a price drop, and does it translate to leads?
These are not exotic analytical questions. They are the questions a competent used-car manager would ask on a lot walk. VIN-level measurement is what lets the advertising operation answer them with ad-platform data instead of instinct. The variables that research cannot capture at the campaign level are exactly the ones that change your CPL.
How Does a "Used Cars" Targeting Bucket Blind You to What Is Actually Happening?
A used-car campaign structured as a single targeting bucket is an averaging machine. Every VIN's performance gets pooled into one set of reported metrics. The $14,000 economy hatch and the $42,000 certified late-model SUV are sharing the same budget logic, the same CPL calculation, and the same optimization signal. When the platform's algorithm decides where to push spend next, it is working from that averaged signal.
The consequence is not just imprecision. It is active misdirection. A unit that happens to carry a great thumbnail image and a well-written headline will pull engagement and skew the aggregate metrics upward, masking the fact that four other units on the lot are burning budget without producing any funnel activity. The campaign looks like it is working. The lot data says otherwise. Neither surface is obviously wrong, but they cannot be connected unless the ad platform data is resolved to the VIN.
This is how two dealer groups running comparable inventory in overlapping markets end up with materially different used-vehicle results. The one with VIN-level attribution can pull budget weight from units that are not converting and redirect it toward units where the funnel is moving. The one running the "used cars" bucket cannot make that distinction. It can only optimize the bucket as a whole, which means the bucket decides, based on averaged signals, where the algorithm thinks money should go next.
The divergence that Auto Remarketing is describing at the public-retailer level plays out at every scale, at the 10-rooftop group and the single point. When market conditions shift fast, the dealer advertising at last period's price points with last period's inventory assumptions is paying for that lag twice: once in wasted spend and once in units that age past their optimal turn window because the campaign never knew they were stalling.
There is also a price-history dimension that category-level campaigns cannot see. Attribution signal is leaking from dealer ad stacks from multiple directions simultaneously. For used-car campaigns specifically, the ability to tie a price move on a specific VIN to a change in that unit's engagement and conversion rate is valuable operational intelligence. You cannot produce that intelligence if you are measuring a category instead of a vehicle.
How AUTONOMi Solves This
AEGIS runs per-vehicle funnel intelligence as a standing nightly capability, unifying GA4 vehicle-detail-page engagement with per-VIN ad performance data from Google and Meta, joined to per-VIN price history from inventory captures.✓ Sep 12 The result is not a campaign report. It is a vehicle-level diagnosis: which units have a visibility gap, which have a price objection signature, which have strong engagement but are failing at the lead step.
The unification covers Google Shopping and Performance Max product reporting at the VIN level, and Meta catalog ad insights resolved to the individual product ID, which corresponds to the VIN.✓ Sep 12 Per-VIN ad performance granularity covers Google and Meta today; TikTok and Microsoft ad performance is not yet available at the VIN level, and AEGIS states that scope rather than claiming otherwise.✓ Sep 12
AEGIS runs a daily inventory-diff rebuild that re-scrapes each dealer's live inventory, diffs it VIN by VIN for arrivals, sales, and price moves, and rebuilds only the affected ad groups across the active campaigns.✓ Sep 12 This means the campaign structure is not a category bucket with a shared budget. Each VIN has its own representation in the ad account, and when a VIN exits the lot, the ad group reconciles. There is no stale creative pointing at a sold unit and no budget flowing toward inventory that no longer exists.
Because AEGIS holds per-VIN price history from successive inventory captures, it can surface markdown-effect reads: whether a price reduction on a specific used unit produced a measurable change in VDP engagement and lead rate within the days following the change.✓ Sep 12 That is the kind of feedback loop that lets a used-car operation make pricing decisions on evidence rather than elapsed time on lot.
On every inventory refresh, AEGIS re-judges the dealer's spend split across new, used, and CPO against the live stock mix, weighing current inventory composition and market conditions in a single reasoning pass rather than a static formula.✓ Sep 12 AEGIS also runs inventory analysis as a standing merchandising capability, surfacing actionable inventory-versus-demand findings that connect to the advertising allocation decisions.✓ Sep 12 The budget is not locked to last quarter's condition split; it moves as the lot moves.
Every allocation shift and every campaign adjustment AEGIS makes is hash-chained in the dealer audit trail, so the GM reviewing the used-vehicle results at month end can see exactly which vehicles received increased weight, when, and why.✓ Sep 12 The measurement surface that was producing inventory explanations by default now produces vehicle-level evidence instead.
What Separates the Groups That Close the Gap
The divergence in used-car results among public retailers is not going to resolve itself through better sourcing alone. The groups running ahead have an operational advantage that compounds: every VIN that is correctly measured produces a signal that improves the next allocation decision, which produces better data on the following cycle. The groups still running category campaigns are averaging that signal away on every iteration.
The good news for independent groups and smaller multi-rooftop operations is that this is a measurement architecture problem, not a scale problem. A 10-rooftop group with VIN-level attribution is running a materially more intelligent used-car operation than a 30-rooftop group whose campaigns are structured as category buckets. Scale amplifies the advantage once you have it, but it does not create the advantage.
The separator is whether the ad platform knows what it is advertising. A platform that knows it is advertising a specific 2022 certified late-model crossover with 28,000 miles, priced at a specific point relative to comparable units in the market, can optimize differently than a platform that knows it is advertising "used cars." That specificity is an architectural choice, and it is made at campaign setup, not at the earnings call.
The groups that will close the gap in the quarters ahead are the ones that treat the measurement infrastructure as the priority, not the downstream consequence. If your used-car campaign cannot tell you which VINs are stalling and why, the sourcing decisions you make at the auction are running on less information than they should be. Connect your inventory to a platform that resolves attribution at the vehicle level, and the sourcing conversation changes because the evidence underneath it changes.
Source: Auto Remarketing
