Recommendations are built from your last 20 page views, weighted most heavily toward the last 5, with older signals decaying to zero influence after 30 days. Explicit filters you set (body style, budget band, fuel type) always override inferred behavior. No sensitive categories (health, finances beyond budget band, family status) are inferred, and you can reset your recommendation profile at any time from Custom Dashboard settings.
At a glance
| Rule | Value |
|---|---|
| Signal window | Last 20 views, top 5 weighted highest |
| Decay | Linear decay to zero influence after 30 days |
| Explicit override | User-set filters always beat inferred behavior |
| Diversity cap | No single source type exceeds 40% of a feed screen |
| Reset | One click, clears inferred profile, keeps explicit filters |
What feeds the recommendation engine
Three inputs drive what you see: pages you've viewed, vehicles or topics you've explicitly followed, and filters you've set in your profile (budget band, body style, fuel type, region). Explicit signals always outrank inferred ones — following an EV brand will surface EV content even if your last ten page views were about pickup trucks.
How weighting works
- Your most recent 5 page views carry the strongest weight in the recommendation model.
- Views 6 through 20 contribute a smaller, decaying weight.
- Any signal older than 30 days is dropped entirely — the engine does not use a permanent behavioral profile.
- A single research binge (e.g., 15 minivan pages in one evening) is capped so it can't crowd out every other category for weeks.
Diversity and anti-repetition rules
To avoid a feed that only ever shows one manufacturer or one content type, no single source type — reviews, comparisons, news, or buying guides — can exceed 40% of a given feed screen, and the same article cannot reappear within 14 days of being shown once, whether or not you clicked it.
What the engine will not do
- Infer or use protected categories such as health conditions, immigration status, or family composition.
- Use browsing data from outside True Motion Auto (no third-party ad-tech data feeds this engine).
- Prioritize sponsored or dealer-paid content in the same slot as organic recommendations without a clear "Sponsored" label.
Transparency and controls
Every recommended module carries a "Why am I seeing this?" link that names the one or two signals behind it — for example, "Because you viewed 2026 compact SUVs" or "Because you follow Brand X." You can mute a topic, reset the entire inferred profile, or switch to a chronological, non-personalized feed from account settings; explicit follows and saved filters are unaffected by a reset.
Resetting your recommendation profile does not remove saved vehicles, follows, or wishlist items — it only clears the inferred browsing signal, so your feed will look more generic for a few sessions until new signals build up.
Frequently asked questions
Why do I keep seeing the same type of car?
Can I stop personalized recommendations entirely?
Does True Motion Auto sell my browsing data to advertisers?
How do I see why an article was recommended?
Do sponsored articles count as recommendations?
Sources & further reading
- Nielsen Norman Group — UX Guidelines for Recommended Content
- Nielsen Norman Group — Individualized Recommendations: Users' Expectations & Assumptions
Figures, prices and policy details were current at the last-updated date above. Automotive pricing, incentives and regulations change frequently — verify time-sensitive details with the linked primary sources. Read our editorial policy and fact-checking standards.
