In September 2022, Ford told its US dealer body that anyone who wanted to sell a Mustang Mach-E or an F-150 Lightning past January 2024 would have to enroll in Model e, install DC fast chargers, commit to no-haggle pricing, and put roughly a million dollars on the counter for the privilege. Twenty months later, in June 2024, Ford quietly threw the program out and opened EV sales to all 2,800 of its dealers.
The website was not the problem. The configurator worked. Ford.com could take an order. What broke was the operating model underneath — who owned the price, who honored it in the store, who trained the F&I desk, who serviced the vehicle on Tuesday morning. When those pieces did not line up, the certification became a tax the dealers refused to pay and the automaker eventually rescinded.
This is the shape of the mistake in automotive digital retail. Buyers have been unusually clear about what they want. Cox Automotive’s most recent car buyer study finds that 63% describe an omnichannel journey as ideal, roughly a quarter say all-online, and in practice barely one in fourteen finishes the entire transaction on a website. The measurable gaps between what buyers say they want and what they get are not at the top of the funnel. They are at the money end — credit, protection products, price finalization — and none of them are solved by another photo gallery.
The Metrics That Actually Move
Every OEM digital retail deck opens on the same slide: sessions, configurator completions, leads generated, time on page. It is a marketing slide. The problem is that none of the numbers on it are the numbers the business is trying to move. Comp sales per rooftop. Close rate on a submitted deal. Finance and insurance product penetration. Back-end gross per unit. Service retention on a delivered vehicle. Those are the metrics the operator answers for on the quarterly call, and none of them show up on the configurator dashboard.
The Cox Automotive data is a roadmap in plain sight. Applying for credit — 48% of buyers want to do it, 33% do. Selecting F&I products — the extended warranty, service plan, tire and wheel, gap coverage that carry most of the back-end gross — 40% want to online, 16% do. Finalizing the price — 37% want to, 19% do. Twenty-four points, eighteen points, fifteen points. That is the agenda in order of size, and every point of it lives at the intersection of the OEM, the lender, the F&I provider, and the store. A configurator does not close that gap. A shared view of the deal does.
Which is why the retailers who quietly move the metrics tend to have reshaped the operation and left the front end alone for a cycle. The ones who ship a redesigned website and declare digital retail complete discover, six months on, that the store metrics did not move. The comp is flat. The close rate is flat. The F&I penetration is flat. The digital property launched. The retail operation did not change.
One Place Where the Deal Lives
The architectural question underneath the entire agenda is old-fashioned. If the online price, the store price, and the contract price are three different numbers, no amount of interface polish will produce a customer who trusts the brand. And the price will differ — because the OEM’s marketing system, the dealer’s DMS, the lender’s decisioning engine, and the F&I menu tool were designed in different decades to answer different questions, and none of them were built to agree.
The retailers moving the metrics have picked one system as the source of truth for a specific deal — usually a shared deal object that the store, the site, the lender, and the F&I provider all read and write against — and rebuilt everything else as a view on top of it. That is not a website project. It is an operating decision about whose number wins when the store and the site disagree, and every downstream problem, from credit application friction to protection-product attach to price transparency, gets easier once that decision is made. Left unmade, every new digital surface is another opportunity for a buyer to catch two versions of the same deal, which is now specifically what regulators describe as deceptive.
Through the Network, Not Around It
The temptation, watching Tesla, is to conclude that the answer is to skip the network. Tesla has delivered more than six million vehicles without a franchised dealer, operates roughly 276 stores across 43 US states, and in states like Texas cannot legally quote a price inside its own gallery — the sale routes to an out-of-state entity. That is a real and impressive operating model. It is also a model built from scratch by a company that never signed a franchise agreement, and it lives inside a patchwork of state statutes that treat pure direct distribution as a special case rather than a preferred one.
For every other OEM, the network is the operation. Nearly seventeen thousand franchised US dealers sold more than sixteen million vehicles in 2025 and cleared more than a hundred and sixty billion in service and parts. That is the delivery system, the service footprint, and the legal structure combined, and reshaping the retail operation means reshaping how the OEM and the network actually run a deal together — not a Model e-shaped bet that the network can be reformed by contract on an aggressive timeline. Ford’s reversal was not a technology failure. It was a change-management failure attached to a large infrastructure ask, and the underlying question — how does an incumbent OEM sell EVs consistently through a network it does not own — is still on the table.
What Only the OEM Can Supply
Buyers visit four or five sites on the way to a purchase. Third-party sites — Edmunds, Cars.com, Kelley Blue Book, Autotrader — take most of the top-of-funnel traffic. Dealer sites take second. The OEM’s own site is one stop among several, and rarely the first or the last. Planning to win that ground back with a better brand experience is an expensive way to lose slowly.
The OEM’s real leverage is what only the OEM can supply, and what the network cannot replace: inventory truth, incentive integrity, and configuration accuracy, syndicated outward so the surfaces buyers already use are correct. The assistant a buyer opens on their phone will read the OEM’s pricing and incentive data whether or not the OEM intended it to, and will summarize whatever it finds. Nearly one in five buyers already uses an AI shopping tool, and one in four among new-vehicle buyers. Machine-readable and correct is now a merchandising requirement, not an IT preference. If the incentive the assistant reads is different from the one on the dealer’s ticket, the buyer arrives already annoyed, and the deal starts thirty minutes behind.
Pricing Integrity as the Compliance Frontier
The federal CARS Rule that specifically addressed advertised pricing and add-ons is gone — the Fifth Circuit vacated it in early 2025 on procedural grounds — and reading that as deregulation would be a mistake. The court did not find the conduct the rule sought to prohibit was lawful. The FTC has been explicit that its longstanding authority on unfair and deceptive practices has not changed, and in 2026 sent warning letters to nearly a hundred dealerships. The names were made public. State attorneys general have moved in the same direction, with multi-million-dollar settlements in several states and larger ones under discussion.
The conduct named in those letters reads like a specification for an honest digital price. Advertise the actual offering price, including all mandatory fees. Do not advertise a price that reflects rebates only some buyers qualify for. Disclose any additional down payment requirement. Do not make the advertised price contingent on dealer financing. Do not require add-ons that inflate the final number. Do not advertise vehicles that are not actually available. Read the list and it is difficult to build to it from two systems of record. Read it and it is straightforward to build to from one.
What Good Looks Like on Monday Morning
The OEMs who are quietly moving the metrics are not running an exotic playbook. They are doing four things, none of which are particularly photogenic.
They start with the retail number, not the digital number. Close rate on a submitted deal from twenty-two percent to twenty-eight. F&I penetration on the online-started deals from a third of the store average to the store average. Service retention on delivered EVs from fifty percent to seventy. The digital roadmap gets evaluated against those numbers, not against sessions or configurator completions. Programs that cannot name a retail number in one sentence rarely move one.
They pick the three commercial gaps the buyer data already named — credit, protection products, price finalization — and rebuild each one end-to-end with the store, the lender, and the F&I provider on the same line of the plan. Not a portal that hands off to a phone call at the moment the deal gets interesting. A workflow where the credit application submitted online is the credit application the dealer sees, the protection products chosen online are on the menu presented in the store, and the price shown on Ford.com or GMC.com or Toyota.com is the price on the buyer’s order at signing.
They put a business owner and a technology owner on the hook for the same number. The head of retail operations and the head of digital, accountable together for close rate on omnichannel deals, not co-chairs of a steering committee. When only one function chairs it, the other is not really committed, and it is better to find that out in the first meeting than in month eight of a program that has already spent the change budget on the wrong things.
And they design for the network rather than around it. The dealer body is where the vehicle gets delivered, serviced, and traded back in three years, and any program that treats the store as a legacy channel to be routed past rather than a partner to be rebuilt with tends to end the way Model e did — a large infrastructure ask, a smaller-than-expected opt-in, and a quiet reversal about eighteen months in.
Who Owns the Price the Customer Sees
Digital retail is not a website. It is an agreement about who owns the customer’s price, syndicated cleanly to every surface the buyer will visit, honored in whichever channel the buyer chooses to finish. The industry has spent a decade choosing platforms and answering a slightly different question — how do we make the front end more beautiful — while the metrics that pay for the business have sat where they were.
The unfinished decision is smaller and harder. Which organization owns the price the customer sees, and how does it stay the same number across every surface the buyer touches on the way to signing. Answer that, and the credit application, the protection-product menu, and the price finalization stop being three separate portals and start being one deal with three views. Leave it unanswered, and every new digital surface is another opportunity for the buyer to catch two versions of the same offer — which is now specifically the conduct the regulator is writing letters about, and specifically the reason the comp did not move.
The technology has never been the constraint here. The constraint is a decision about who runs the retail operation and whose number wins when the store and the site disagree. That decision belongs to the operator running the business.
