Automakers must scrap metrics for dealers

Auto manufacturers routinely assess dealer sales performance relying on metrics with no established or demonstrated statistical reliability or accuracy to gauge the minimum number of new cars dealers should expect to retail.

Factories deploy calculations ("sales expectations" vs. actual sales) to impact dealers in a variety of ways. Automakers require dealers to score "average or above" to be considered minimally satisfactory. Thus, as many as half a brand's dealers will be assessed as "underperforming," in breach of the dealer agreement at any — or every — particular moment in time.

Carmakers may threaten to terminate dealers with less-than-average numbers, while those who beat "expected" scores may receive greater incentive or bonus payments. Also, when dealers seek approval to acquire new dealerships, the reviewing brands consider, as a "gateway," the applicant's past sales performance based on the metrics from relevant brands.

Nothing wrong here, say the factories. They boast of "industry standards" that scientifically measure dealer retail sales performance against "expected" sales numbers. They maintain, erringly, that this meritocratic methodology yields unbiased counts and comparisons, generating reliable rankings of dealers from top to bottom. Let's take a look at an illustrative, hypothetical example:

Dealers A and B operate in New Jersey. Each is a franchisee of the same automaker, and each retailed 1,000 new units in 2016. But Dealer A's calculated "expected" annual retail sales nationwide was 2,000 units, and Dealer B's "expected" count was 500. Dealer A thus falls a whopping 50 percent below the brand's "sales effective" standard. Dealer B stands at an enviable 200 percent of the benchmark. (Note that these calculations may be more complicated when a brand competes in separately calculated "segments.")

Applying dealer scores, franchisors rank dealers from best to worst, nationally and across other geographical areas.

Franchisors arrive at their numbers this way: First, they count the number of competitive registrations in a dealer's assigned area of responsibility. This supposedly establishes the dealer's "opportunity" number. Then, they apply to that number a percentage stemming from the brand's market share in a given benchmark geographical area. In the above example, if the brand's market share is 10 percent, Dealer A's competitive registrations in its area of responsibility were 20,000, whereas Dealer B's were only 5,000. Dealer A "should have" retailed at least the "expected" 2,000 units (10 percent of 20,000), but Dealer B doubled his "expected" sales.

Here's the fundamental problem: The metrics are seductive because the arithmetic of the "expected" sales divided by the number of sales is undeniable. But the unreliability of ascertaining "expected" sales demolishes the methodology. For "apples-to-apples" comparative evaluations of same-line dealers' sales expectations, factories need to control for numerous variables such as unique locations, sizes and shapes of areas of responsibility, thoroughfares, demographics, local preferences, etc. — tasks that factories cannot resolve in a statistically reliable way.

Dealer challenges to such factory methodologies have garnered some modest successes in recent years. In 2016, Beck Chevrolet v. General Motors, a matter certified by the 2nd U.S. Circuit Court of Appeals to New York's highest court, decimated key aspects of the factory sales numbers game.

GM had used the whole of New York to formulate each dealer's "expected sales number ... based on an adjusted state market average." The state court held that GM's methodology violated New York law, observing that "GM may not rely on a standard that is unreasonable and unfair simply because of its prevalence within an industry the Legislature sought to regulate."

Subsequently, when Beck returned to the 2nd Circuit, that court noted "it is unlawful ... to 'use' [the average] standard — alone or in conjunction with other metrics — to assess an automobile dealer's compliance with its franchise agreement." GM purported to remedy its transgressions by changing its geographical metric assessment from statewide to local. But its new benchmark methodology does nothing to cure the inherently corrupt method of calculating "expected" sales in the first place.

Auto franchisors have a right — indeed, a responsibility — to require and enforce dealer compliance with reasonable and attainable standards, but they also must assure fairness and accuracy.

They need a methodology that contemplates and encourages compliance with reasonable standards by 100 percent of the dealer body. They should jettison the "industry standard" of zero-sum metrics that make it a certainty that "lower performing" dealers are unsatisfactory and in breach. They should assess each dealer on the outcome of specific on-site reviews and inspections to assure that dealers comply with reasonable requirements to, for example, hire and train sales personnel, technicians and capable managers.

Facilities must meet reasonable and nondiscriminatory standards. Dealers must stock adequate vehicle and parts inventories (assuming availability and supply by the carmaker).

In short, "tailored" assessments will eliminate the sweeping, misplaced reliance upon dealer-vs.-dealer numerical metrics that are inherently unreliable.