Retention is a lagging indicator. The experience signals that drive it are leading ones.
The headline from the 2026 Reynolds Retention and Defection Report is easy to read as reassuring. Overall brand retention across the automotive industry held at 43.9% nationally, roughly where it was the year before. Stable. No alarm bells.
Look harder at what sits beneath that number and the picture is considerably less comfortable.
Of the 38 vehicle brands tracked in the report, 20 saw their retention rate decline year over year. Only five brands now keep more than half of their customers when those buyers return to market, down from six brands in each of the two years prior. Eight mainstream brands lost retention ground in 2025 alone. Volkswagen dropped 4.5 percentage points. Dodge dropped 3.7. Six mainstream brands now sit below their 2021 retention levels, with most of the long-term slide concentrated in Stellantis nameplates including Fiat, Dodge, Jeep, and Ram.
The national average is stable because the gains of a few high-performing brands are masking the losses of many others. For the brands on the losing side of that equation, the number is not stable at all.
Why Retention Erodes Before You See It Eroding
The structural problem with retention data is timing. By the time a brand's repurchase rate moves, the customer experience decisions that drove that movement are months or years in the past. A customer who had a poor service visit in 2024, didn't say much about it, and quietly bought a different brand in 2025 has already left. The 2026 retention figure captures the result but tells you nothing about the experience that produced it.
This is not a hypothetical. J.D. Power's 2026 Customer Service Index found that when overall dealer service satisfaction reaches 950 or above, 86% of mass market customers say they will definitely return to the dealer for paid service. That connection between service experience and repurchase behaviour is well established. What's less established is the feedback infrastructure that would allow an OEM to act on service quality signals in time to influence the outcome.
Most post-visit feedback mechanisms are designed to report on experience after it has happened, not to surface the patterns that predict whether a customer will still be in the brand family two years later. By the time those patterns become visible in retention data, the window to intervene has already closed.
The Regional Problem Inside the National Average
The Reynolds report highlights something else worth attention for OEMs managing dealer networks at scale: retention is not a uniform experience across geography.
Toyota leads mainstream brands nationally with a 62.6% retention rate. But in the Western US, that rate climbs to 63.1%. Honda's nationwide retention of 52.2% looks different in markets where its model mix skews toward higher-volume segments. Chevrolet and GMC have become the truck brands most likely to lose buyers nationally, having overtaken Ford on defection rates. Yet in the Midwest, 66% of Chevrolet and GMC owners stay with those brands when they trade, and among those who buy another truck, 82.4% remain loyal to the same badge.
The same brand can be winning in one region and losing in another, often for reasons that have nothing to do with the vehicle itself. Local service quality, advisor consistency, network density, complaint resolution speed — these are all experience variables that behave differently across a dealer network. A national retention average smooths over all of that variance. For OEMs, the useful signal is not the average; it is the deviation from it.
What the Brands Holding Above 50% Understand
It is worth asking why Toyota and Honda are the only mainstream brands to maintain retention rates above 50% in a year when most others declined.
Product quality and model breadth are part of the answer, but they are not the complete one. Both brands have also invested consistently in understanding the ownership experience at a granular level. They have not treated customer feedback as a compliance requirement satisfied by annual survey scores. They have treated it as a management tool, one that surfaces where the experience is breaking down before that breakdown shows up in repurchase data.
The brands experiencing multi-year retention slides share a different pattern. The signals of deteriorating customer experience were almost certainly present in their feedback data before the retention numbers moved. The question is whether those signals were visible in a form that allowed them to be acted on at the level of the individual location, service advisor, or interaction type.
Retention is a lagging indicator. The experience signals that drive it are leading ones. OEMs that have built feedback systems capable of connecting those two things, surfacing location-level experience patterns before they become brand-level retention losses, are not reacting to the 2026 report. They are already working on 2027.