Key Findings—scaled to average rooftop ad spend
Andrew Reutsky,
ORCID iD icon
Marina Sorokina and Nadya Bel · Reviewed by John A. Deighton, Harvard Business School

New Banner-VIN-Profit audit shows sixty-two percent of a dealership’s ad spend lacks a defensible link to profit.

Research examines 100 ad listings over Q1 2026, finds $17,308 of $27,814 in ad spend carried no defensible link to a dollar of profit, and open-sources the Banner-VIN-Profit attribution model used to trace each ad to the vehicle it sold and the profit it produced., transforming systemically wasteful advertising into a liquidity lever amid economic volatility.

CONFIRMATION: Your submission has been received.
Andrew will follow up with a copy directly to your inbox. We've also trained an AI agent on it — ask it to check a number, challenge a finding, or walk you through the method.
Oops! Something went wrong while submitting the form. Please refresh the page and try again.
Reviewed by John. A. Deighton,
Harvard Business School, Professor of Business Administration & Management, Emiritus, and
Authority on Digital Marketing
Interviewees from
2 years
of field research
57
industry interviews
Countless
academic papers read
Zero
$ sponsored
Manufacturers, including General Motors and Ford
Regulators, incl. NADA and regional dealer boards
Dealerships comprising over 80 franchised rooftops
Advertising agencies, including Leo Burnett
Lenders, media, privacy professionals and academics

01 — THREE FINDINGS, EACH WITH ITS RECEIPT

62%
$17,308 OF $27,814

Sixty-two percent of audited spend produced no defensible link to profit. Over the quarter 62 tracked vehicles sold; of 646 verified leads, 19 sales were defensibly attributable, returning $10,507. Under the most conservative statistical reading, the range is 45–62%.

41%
646 VERIFIED OF 1,102 REPORTED

Forty-one percent of platform-reported leads could not be confirmed against the dealership’s own air-gapped records. Asked how its figure was derived, the media partner cited a "proprietary" algorithm. The paper names no platform: the critique targets the systemically wasteful ad spend, not a vendor.

$51/lead
EACH RETURNED ROUGHLY $16

Premium, boosted placement consumed 69% of audited spend and quadrupled leads. The reported engagement lift was true. It was also not worth it. The dealership cut most boosted listings immediately — the audit paying for itself in a single budget decision.

PODCAST OVERVIEW
A professional headshot of Andrew Reutsky in a dark navy blue blazer and baby blue shirt, photographed outdoors with a blurred green tree in the background. He is facing the camera with a slight smile.

“Retail advertising in a category as complex as automotive isn't unmeasurable so much as under-measured in connection to the bottom line.”

Business student at Harvard; over three years producing national retail advertising as Art Director for Volvo and General Motors — thousands of the very banners this paper audits — before resigning in August 2024. Learn More ↗

TEST YOUR OWN DIGITAL AD SPEND

What the audit’s range would mean for your quarter.

The study found between 45 and 62 percent of one dealership’s audited quarterly digital ad spend carried no defensible link to profit. Enter your own quarterly figure to project that range in dollars.

QUARTERLY AD SPEND
DEFENSIBLE LINK 38–55% $10,506 TO $15,352
NO DEFENSIBLE LINK 45–62% $12,462 TO $17,308
ACROSS A YEAR
$49,848–$69,232

Indefensible at the audit’s range, annualized — the working capital you cannot currently defend.

LEADS YOU COULD NOT CONFIRM
456 of 1,102

Scaled from the audit’s 1,102 vendor-reported leads, of which 456 could not be confirmed against the dealership’s own records.

RETURNED AS ATTRIBUTABLE GROSS
$10,507

Scaled from the $10,507 of front-end gross the audit defensibly traced back to advertising.

03  —  REPLICATION OFFER

APPLICATIONS CLOSE IN 00:00:00:00

ESTIMATED VALUE

$76,100

The estimate doesn't include efficiency gain that comes from finally knowing which ads bring profitable customers — and which do not.

One dealership,
one quarterly audit, free.

To extend the evidence base, one North American dealership will receive a complete Banner-VIN-Profit audit on live advertising through Q4 2026 — in exchange for anonymized data contributed to the follow-up study, which reports practices that improve advertising ROI rather than merely diagnose the waste.

ESTIMATED VALUE

$76,100

The estimate does not include efficiency gains that come from knowing which ads bring profitable customers.

Apply to participate
Applications close September 14, 2026

04  — CONTACT

Tailor the benefits of this research to your team's needs.

The method in this paper was built to be run, not cited. Bring your own numbers and we will find if your advertising traces to profitability—and what to do about it if it doesn't.

CHOOSE A ROUTE

A — BANNER-VIN-PROFIT AUDIT APPLICATION

Fill out the form below—the submission is immediate. We are interviewing on a rolling basis and will notify everyone by September 14, 2026 with the results—you are free to exit consideration pool at any time.

CONFIRMED: Contact details saved.
Please move on to the five qualification questions below.
Oops! Something went wrong while submitting the form. Please refresh the page and try again.
B — TEAM WORKSHOP

Please provide your onboarding information and Andrew will contact you to confirm and discuss the details.

Your submission has been received! Andrew, lead author, will get in touch within 24 hours.
Oops! Something went wrong while submitting the form.
C — PRIVATE RESEARCH REVIEW

Pick a slot from Andrew’s live calendar. Confirmation is immediate.

AVAILABLE ON A FIRST-COME BASIS UNTIL SEPTEMBER 14, 2026. OPEN FULL PAGE↗

D — BY PHONE

+1-888-837-7738

For any questions, feel free to call Andrew between 8 AM and 8 PM any day (ET). United States and Canada Toll-Free.

PROJECT YOUR DIGITAL AD SPEND AT AUDIT'S RANGE

AUDIT YOUR REAL SPEND — REPLICATE THE STUDY

STATUS QUO VS. BANNER-VIN-PROFIT

What changes when the ledger is yours.

Conventional automotive attribution measures activity and asks the dealership to accept the vendor’s arithmetic. Banner-VIN-Profit measures money, and every step of it can be reproduced from records the dealership already owns.

STATUS QUO
BANNER-VIN-PROFIT
Leads reported by the platform being graded
Leads air-gapped in the dealership’s own records
Impressions, clicks, and engagement as the unit of success
Profit on the VIN the ad helped sell
Online activity and showroom customers held in separate systems
Every lead a source-tagged CRM record, traced to a person on the lot
Lead collection practices described as proprietary; the arithmetic cannot be checked
Open-sourced attribution model published in full, with replication guidance and boundary conditions publicly available
New software, new contracts, vendor cooperation required
A phone extension, an inbox, a link tag, and a spreadsheet
Advertising as a marketing cost — cut blind when margins compress
Advertising as an audited liquidity position — cut with evidence or doubled down on

THE LEAD AUTHOR WILL NOT ASK TO BE TAKEN ON FAITH

“I am an ad guy arguing that advertising can be a liquidity lever. Maybe I just want a job or a contract after I graduate. The truth is, I do. But if that were all I wanted, I would have stayed at my well-paying job and done just fine. I resigned because I no longer believed automotive clients were getting their money’s worth, and I have spent two years since testing whether that belief survives evidence. At one dealership, it did — and it made me believe we, the ad industry, can do better. I invite you to read this paper as skeptically as you would a vendor’s report card. Unlike one, it shows its work.”

Business student at Harvard; over three years producing national retail advertising as Art Director for Volvo and General Motors — thousands of the very banners this paper audits — before resigning in August 2024. Learn More ↗

QUESTIONS AND ANSWERS

Questions worth asking before you read.

A finding is only as good as the questions it survives. Here are ours.

What is the Banner-VIN-Profit attribution model?

A method for tracing an individual advertisement to the vehicle it sold and the profit that vehicle produced. Each lead is captured in the dealership’s own records before the campaign runs, followed into the CRM as a source-tagged customer, and matched to the Vehicle Identification Number of the car that customer bought. The listing’s prorated cost is then set against that vehicle’s front-end gross.

The result is a per-ad profit figure rather than a per-ad engagement figure, computed from records the dealership controls.

How long does an audit take to run?

A quarter of live spend, because the trace has to follow leads through to delivered vehicles. Setup takes a day. The free audit on offer runs through Q4 2026.

How do I set up Banner-VIN-Profit?

Four things, none of them software purchases: a dedicated phone extension, a dedicated inbox, a link tag on the listings you are testing, and a spreadsheet. The audit in the paper used nothing else.

The paper includes the full replication guidance — what to record, in what order, and where the method breaks down.

Do I have to hand over customer data to run it?

No. The method runs inside your own systems, on your own records. Nothing has to leave the dealership for the arithmetic to work. Where a dealership does choose to contribute data to the follow-up study, it is anonymized first.

Why should I trust the lead author?

You should not, on faith. Andrew Reutsky spent over three years producing national retail advertising for Volvo and General Motors — thousands of the banners this paper audits — before leaving in August 2024, and is now a business student at Harvard. He is an ad person arguing that advertising can be a liquidity lever, and he says so in the paper.

What is offered instead of trust is the working: the method, the boundary conditions, and the limitations are all published so you can check them.

Is this backed by a consultancy trying to sell me something?

The research is independent and names no platform. The critique is aimed at how automotive attribution is structured, not at any single vendor.

The commercial interest is stated plainly rather than hidden: the author intends to work in this field after graduating. One free audit is on offer to extend the evidence base, in exchange for anonymized data.

Can a student paper be as good as consultancy research?

Judge it on the same terms you would judge any vendor’s report: is the method stated, are the assumptions declared in advance, and can someone else reproduce the result?

This paper draws on interviews with fifty-seven manufacturer, regulatory, dealership, and advertising executives; one quarter of audited spend at one dealership; and hundreds of academic works through Harvard’s libraries. Harvard Business School faculty reviewed it. Its limitations are published next to its findings. Most vendor reporting passes none of those tests.

More questions answered
What is the actual goal — is this paper a funnel for your business?

In part, and the paper says so. The stated goal is to establish whether advertising spend at a dealership can be traced to profit at all, and to publish the method so others can test it. The author’s interest in working in the field afterwards is disclosed rather than concealed. Read it as skeptically as you would a vendor’s report card.

Why 62 percent and not the more conservative number?

Both are published. Under the audit’s most conservative reading the untraced share is 45 percent rather than 62. The range was registered before the results were known, so neither figure was selected after the fact to make a point.

One quarter at one dealership. Why should that generalize?

It should not, and the paper does not claim it does. The finding is that the trace is possible and that at this dealership most spend failed it. The method is the transferable part; the percentage is not. That is why the follow-up study is recruiting another rooftop for replication.

Does this account for brand advertising that does not convert immediately?

Not fully, and that is a declared limitation. The audit measures what can be defensibly traced within the quarter. Advertising can do work the trace does not capture. The argument is narrower than “untraced means worthless”: untraced spend is spend a dealership cannot defend when it has to choose what to cut.

Which platform is being criticized?

None is named. Asked how its lead figure was derived, the media partner in the audit cited a proprietary algorithm — a structural condition, not a single company’s failing. Naming a vendor would let the industry treat the problem as someone else’s.

Downloads Now Available
Read the Paper in Full

Full methodology, replication guidance, and the complete findings.

CONFIRMATION: Your submission has been received.
Andrew will follow up with a copy directly to your inbox. We've also trained an AI agent on it — ask it to check a number, challenge a finding, or walk you through the method.
Oops! Something went wrong while submitting the form. Please refresh the page and try again.