Applying sector benchmarks to diagnose your marketing metrics

A slide lands in your inbox: "industry average cost per acquisitioncost per acquisitionCost Per Acquisition: the total cost to generate one customer or conversion, computed by dividing total spend by the number of acquisitions.View full definition →: $640." Before that number reaches a board deck, three questions decide whether it is useful or radioactive. Who produced it, what denominator did they divide by, and from which year? Get those wrong and you will spend a quarter fixing a problem you do not have, or worse, congratulating yourself on a number that was never comparable. Your own figures, computed the way the sibling lessons set out, only become diagnostic once they sit beside a reference range built the same way.

Which benchmark sources hold up

Rank sources by how the number was produced, not by how confident the headline sounds.

Dealer financial panels. NADA compiles operating data from franchised US dealerships in its annual NADA Data report, including advertising expense per new vehicle retailed (several hundred dollars per unit in recent years, and it moves with volume). It comes off dealership income statements, so the numerator is money the store actually booked and the denominator is a retailed unit. That is the closest thing to an audited peer set for franchise retail. Its limit matters: it captures store spend, not OEM tier-one brand advertising, and it says nothing about a direct-sales brand with no dealer P&L.

Syndicated research and transaction panels. J.D. Power runs large owner and shopper surveys plus transaction-level data from dealer systems, which is where per-unit incentive spend comes from. Incentive money is marketing money that never appears in the marketing ledger. If you benchmark your media cost against a rival while ignoring a four-figure per-unit incentive gap between you, you are comparing half a budget to a whole one.

Registration-based loyalty data. S&P Global Mobility ranks brand loyaltybrand loyaltyYour customers' propensity to repeatedly purchase from you and resist competitive offers, driven by satisfaction, habit, trust, and switching costs.View full definition → off vehicle registration records, so it measures what owners did rather than what they told a survey they intended to do.

Vendor benchmark reports. Segment, which sells the customer data platformcustomer data platformSoftware that unifies customer data from every source into one persistent profile that marketing, sales and service teams can act on.View full definition → that produces many of these numbers, publishes engagement and conversion benchmarks drawn from its own customer base. Directionally useful, structurally biased: that base is instrumented, digital-first companies, few of which sell a $45,000 physical object through a third party who owns the customer relationship. A vendor also has an interest in the gap it reports looking fixable by the product it sells.

The failure mode here is benchmark shopping. Check four sources, adopt the one that flatters your quarter, and you have converted diagnosis into public relations. Choose the source before you compute your number, and write down why you chose it.

Normalise before you compare

Four alignments, in the order they usually break.

Denominator first. Per retailed unit, per household, per VIN and per lead are four different numbers, and blending new with used quietly destroys the comparison. Then scope of spend: does the benchmark include agency retainers, CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition → licences, sales commissions, third-party lead fees, OEM co-op reimbursement? Then vintage. Inventory shortages in 2021 and 2022 pushed gross profit per new vehicle to record levels; any efficiency ratio anchored to that period will make a normal year look broken. Last, market structure. A US franchise benchmark does not transfer to a European agency model where the manufacturer sets the price and owns the customer record.

Here is what the correction usually looks like on a mid-size dealer group's quarter:

Reported sales and marketing spend        $300,000
Less: used-vehicle campaigns              -$45,000
Less: OEM co-op reimbursement received    -$30,000
Add: commissions attributable to acquisition +$60,000
Normalised new-vehicle acquisition spend  $285,000

Raw:        $300,000 / 400 retailed units (new + used) = $750
Normalised: $285,000 / 300 new units retailed          = $950

The raw number sat comfortably inside the volume band. The normalised number sits at the top of it. Nothing about the business changed; the first calculation was simply answering a different question from the benchmark it was being held against.

Reference ranges, and their error bars

Precise current figures are rarely published by brands and vary by market and source. Treat these as directional.

• Cost per new-vehicle buyer, US: roughly $600 to $900 for volume brands and their dealers, roughly $1,000 to $2,500 for premium marques, and frequently well above $2,000 for early-stage EV brands building awareness from zero.
• Ratio of lifetime valuelifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → to acquisition costacquisition costCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition →: the cross-industry rule of thumb is 3:1 healthy, below 1:1 loss-making, much above 5:1 possibly underinvestment. See this HubSpot explainer on the LTV:CAC ratio for the general logic. Automotive ratios run far higher because service, parts and the replacement cycle stack up, so 10:1 is common here and not automatically a red flag.
• Lead-to-sale: internet leads commonly cited around 5 to 12 percent, walk-in showroom traffic nearer 20 to 30 percent.
• Brand loyalty: 40 to 60 percent for the strongest mainstream brands, materially lower for niche players.

How big a gap is a real gap

Within about 10 percent of the benchmark, do nothing. Sample composition and definition drift move numbers that far on their own, and chasing that gap burns analyst time you need elsewhere.

Between 10 and 30 percent, you have something worth a channel-level look.

At 2x or more, the odds strongly favour a definition mismatch over a performance collapse. Re-derive both sides before you fire an agency. The most common culprits: fleet units sitting in a retail denominator, tier-one brand spend landing in a store-level numerator, and a benchmark whose "customer" is a household while yours is a transaction.

The counter-example is the one nobody checks. A blended number sitting exactly on the benchmark can hide new-vehicle acquisition running 40 percent over while used runs 30 percent under. Match at the aggregate, split at the segment, always.

The same number, three verdicts

Three marketers look at an identical $1,900 per buyer.

The volume franchise dealer should be alarmed: that is double the healthy band and the margin per car cannot absorb it. Something upstream is broken.

The premium marque is comfortable, because gross profit per car and a high-spending owner base carry it.

The direct-sales EV brand expects it, and has a harder problem: no floorplan, no co-op, no franchise income statement, so NADA Data does not describe it and J.D. Power's dealer-derived transaction views only partly do. Its only honest benchmark is its own prior quarters. Watch the trend line, not the level, and set the expected rate of decline in advance so you can tell a plateau from progress.

Turning a gap into a fix order

The gap tells you which door to open first.

Cost per unit above band, conversion at band: the problem is upstream. Audit targeting, channel mix and what you pay third-party lead providers for stale inventory of the same shopper your rivals already bought.

Cost at band, conversion below band: the problem is the sales floor and response speed. More media spend makes this worse, because you buy more of the leads you are already failing to work.

Both at band but the value-to-cost ratio thin: look at the service-to-repurchase cycle the retention lesson tracks, since the assumption doing the most work in your ratio is how many cycles you actually get.

Everything at band and share still falling: benchmarks have stopped being the right tool. A competitor with a different economic model is buying growth you cannot match on efficiency grounds, and the decision is strategic, not diagnostic.

One second-order consequence worth naming: never wire variable compensation to an external benchmark. The moment a bonus depends on beating NADA Data, someone reclassifies fleet units as retail, defers spend into next quarter or moves an agency fee into a non-marketing cost centre. Benchmarks diagnose. Targets should be internal, auditable and owned by someone who cannot also edit the denominator.

Key Takeaways

• Source before number. NADA Data for franchise store economics, J.D. Power for shopper behaviour and per-unit incentive spend, S&P Global Mobility for registration-based loyalty, vendor reports like Segment's for direction only, with their sample bias stated out loud.
• Normalise four things before comparing: denominator, scope of spend, vintage (2021 to 2022 grosses distort everything) and market structure. A $750 blended figure became $950 per new unit in our example without the business changing.
• Gaps under 10 percent are noise. Gaps of 2x are almost always definition mismatches. Re-derive before you act.
• Directional US ranges: roughly $600 to $900 per buyer for volume, $1,000 to $2,500 premium, above $2,000 for early EV brands; internet leads convert around 5 to 12 percent, showroom traffic 20 to 30 percent; strong-brand loyalty 40 to 60 percent.
• A direct-sales brand has no valid external peer set. Benchmark it against its own trend and pre-commit to the decline rate you expect.
• Benchmarks are for diagnosis. Attach a bonus to one and you will get a gamed denominator instead of a better business.