Independent newsroom The Wyre News Network OpEd desk

Perspective 5 min read

Your Reporting Is Lying to You

Add up the conversions your six platforms claim and you sold more cars than your DMS has ever heard of. Why attribution broke for good / self-graded homework, double-counted buyers, MTA models with opinionated decimals, and a showroom no pixel can see / and the inference-based playbook dealers use to allocate budget without believing a single dashboard.

By Amber Stone for the Wyre OpEd desk

Run this experiment after your next month-end: take the conversions each of your six platforms claims it drove, and add them up. Compare that to the number of cars you actually sold. The total will be more / sometimes comically more / than reality. Six report cards, all A's, written by six students grading their own homework, describing a month your DMS says never happened.

Nobody is technically lying. Every number is computed exactly as documented. And yet the report as a whole is a lie, because it answers a question no dealer is asking / "what can this platform take credit for under its own rules" / while wearing the costume of the question you are asking: "what actually sold cars." Here is what broke, why the fix everyone sells does not work, and what a dealer can actually do.

Why attribution broke / and stays broken

The old promise of digital advertising was a clean chain: impression, click, lead, sale, each link logged. That chain has been dissolving for years. Privacy rules and platform lockdowns killed most cross-site tracking. The walled gardens do not share user-level data with each other, or with you. Your buyer researches on a phone, gets retargeted on a tablet, and walks into the showroom on a Saturday / and the sale lands in the DMS, offline, weeks after the first touch, sometimes under a spouse's name and a different email. Automotive is close to the worst case for attribution: a months-long consideration journey, a six-figure basket, and a purchase that happens at a desk, not a checkout page. The tracking layer that attribution software assumes simply does not cover the journey your buyers actually take.

The specific lies in your stack

Last-click hands the trophy to the harvest. Whatever touched the buyer final / almost always branded search or a bottom-funnel retargeting click / gets the whole sale. The upper-funnel work that put your store in the buyer's head months earlier gets zero. Budget follows the credit, so last-click reporting systematically starves the planting to feed the harvest, right up until the harvest mysteriously shrinks / because nobody planted.

Self-attribution double-counts the same buyer six ways. Each platform runs its own attribution window and claims any sale it touched within it. One buyer who saw your ads everywhere / which is exactly what a good campaign does / is claimed as a conversion by every platform that grazed them. Nothing deduplicates across gardens, because the gardens do not talk. That is the arithmetic behind the impossible month-end total, and it means summing platform dashboards is not a rough measure. It is a category error.

View-through credit is generosity dressed as measurement. An autoplayed impression from four weeks ago "assisted" the sale, says the platform that sold you the impression. Sometimes it genuinely did / upper funnel is real, we have argued so in this very blog. But the entity measuring the assist is the entity billing for it, and its window settings are tuned by its incentives. Treat view-through numbers as a vendor's opening offer, not a finding.

Multi-touch attribution software promises math it cannot have. Here is the uncomfortable one, because MTA is sold as the grown-up fix. Fractional credit / forty percent to the video view, thirty to the click / requires observing individual buyers across every platform and device. That user-level, cross-garden data is precisely what no longer exists. So the models allocate by assumption: position rules, decay curves, statistical inference over gaps. The output arrives with decimal places and the confidence of arithmetic, but the decimals are opinions. An attribution model does not discover where credit belongs. It distributes credit according to what it was built to believe.

The showroom is a tracking black hole. Even perfect digital measurement would miss the point of sale itself. The buyer walks in, the deal happens in the DMS, and unless someone deliberately matches sales records back to marketing exposure / carefully, with all the fuzziness of changed emails and household buying / the systems literally never learn the story ended. Six platforms optimizing toward form fills, while the actual conversion event happens at a desk none of them can see.

What multi-touch reality actually looks like

Accept the hard truth first: you will never get deterministic, per-sale truth across six platforms again. The dealers measuring sanely have stopped chasing it and switched from accounting to inference / triangulating with instruments that do not depend on the broken chain.

The scoreboard is the DMS, full stop: units, gross, and a monthly matchback of sales against marketing exposure, done with honest fuzziness. Platform dashboards get demoted to thermometers / useful for optimizing inside a platform, banned from settling budget fights between platforms, because their numbers are not comparable and never will be. The tiebreaker is incrementality: the only attribution question that ever mattered is "what happens if we turn it off," and you answer it by actually testing / pause a channel in one market and watch, hold out a region, change one budget at a time and read the DMS thirty days later. Slow, unglamorous, and the only method on this page that produces evidence instead of allocation. And the early-warning gauges are brand instruments / branded search volume, direct traffic, store visits / which move when upper funnel is working long before any dashboard awards credit correctly.

Our position

Attribution is not broken like a bug that ships a fix next quarter. It is broken like privacy won / structurally, permanently, and mostly for good reasons. The reporting layer built on the old assumptions did not gracefully degrade; it kept printing precise numbers with the confidence of an era that ended, and precision without truth is the most expensive kind of lie, because it feels like rigor while it quietly misallocates your budget toward whatever measures easiest.

So stop buying certainty / nobody sells it anymore, whatever their deck says. Buy better decisions instead: a DMS scoreboard that counts what is real, platform numbers used only where they are honest, incrementality tests for the questions that move real money, and brand gauges for the work that pays off in quarters. That is multi-touch reality for a dealer on six platforms. It has fewer decimal places than the dashboards / and unlike the dashboards, it will not tell you a story about a month that never happened.

Related Reading

More Opinion

From the same desk

Perspective

A Rounding Error

Over twenty-four hours the single largest AI crawler on one consultant's website arrived roughly 1,500 times under a nonprofit research archive's name, sent back nothing, and asked for his SSH keys, his AWS config and his Firebase service account key. Across a hundred such paths: 1,028 requests, 6.7 megabytes, zero referrals, and requests for anything he had actually written rounding to nothing. It appeared in no security log, because security logs record rule trips and he was not blocking it. It appeared in exactly one place: the AI crawler dashboard, counted as an audience. Cloudflare's CFO says humans will be a rounding error on the internet within five years and admits he has been wrong every time, always low. Machine traffic passed human traffic in May 2026, two years earlier than the company forecast. Why robots.txt is a request that OpenAI's own documentation says may not apply, why blocking the obvious way costs you the visibility and keeps the fetches, what it means that agent config files are now on the standard secret-scanning wordlist, and the afternoon of work almost no company has done.

13 min

Perspective

Cut The Price Last

When demand softens, the first lever everybody reaches for is price. It is also the most expensive lever on the board, the hardest to reverse, and almost never the one that was actually broken. Redfin reports buyers have negotiating power across most of the country, the median housing payment fell to $2,575 because sellers cut asking prices to a one-year low, and pending sales hit their lowest since early April. Underneath it, July payrolls came in at minus 23,000 against an expected plus 83,000 with another 103,000 revised away. So the softening is real, and the question is which lever you pull and in what order. Why cutting price assumes a diagnosis you have usually not earned, the cheaper levers ranked, why the gap between the decision and the execution is where the margin leaks, and the measurement problem underneath all of it: cost per lead holds beautifully steady while cost per closed sale doubles.

6 min

Perspective

The Layer Nobody Looks At

A man in Connecticut typed instructions into his own court filing in three-point white text on a white background, where no judge would see them and any software would. They told the system to agree with him and to disregard the court's earlier denials. He was caught, warned, did it again, and lost his e-filing privileges. In the same seven days Google made the visible watermark on its AI images optional while keeping the invisible one, Anthropic shipped an API letting third parties detect text its models probably wrote, a new benchmark found that no frontier model clears sixty percent on basic visual perception with counting, localization and OCR among the weakest skills, and OpenAI started turning Mac users' clicks and keystrokes into a searchable memory their assistant can read, while warning in its own documentation about prompt injection. Five stories, one story. Every document, image and screen you produce now has two readers and they do not see the same thing. What the judge worked out on his own about where these attacks get caught, why provenance just moved into the layer humans cannot check, what a watermark hit actually proves and why a clean result proves nothing at all, and the ten-line diff almost nobody runs.

18 min