CarGurus
monetize a proprietary audience by reselling to suppliers a predictive intelligence fed by shoppers' behavioral signals
At its full-year 2025 results (released February 19, 2026), CarGurus quantified its AI monetization strategy: about 500 million first-party shopper signals ingested each day, turned into predictive intelligence and resold to dealers through products such as Dealer Data Insights (used by 60% of global paying dealers) and PriceVantage, with annual revenue of 907.0 million dollars, up 14%, and monetized products expected to grow roughly 15x in 2026.
Key points
- CarGurus resells to dealers a predictive intelligence fed by its shoppers' signals.
- CG Discover (natural-language search) and Dealership Mode instrument the audience: about 500 million first-party signals per day.
- Dealer Data Insights used by 60% of global paying dealers in 2025; monetized products expected at +15x in 2026.
- Proof A: full-year 2025 results, revenue 907M USD (+14%), 34,409 paying dealers, status confirmed.
Objective
Move CarGurus from a lead-generation marketplace to a day-to-day software partner for dealers, by reselling as subscriptions a predictive intelligence (demand, pricing, inventory) that the brand produces from the behavioral signals captured on its own shopping surfaces. The consumer AI experiences (Discover, Dealership Mode) attract and instrument the audience; that audience feeds the predictive platform sold to dealers.
The deployment
CarGurus sells its audience twice. On the shopper side, the brand has deployed AI surfaces: CG Discover, a generative search where the user describes in natural language what they want (budget, use, options) and receives a curated list, and Dealership Mode, an app mode that triggers by geolocation when the shopper is on a dealer's lot and shows the lot inventory, AI-generated pros and cons, a recommendation described as unbiased, and an all-in price. Dealership Mode is announced as live for all consumer app users. These surfaces capture on average about 500 million first-party signals per day in 2025. CarGurus translates these signals into predictive intelligence on demand, pricing, and inventory, then resells it to dealers as software products. Dealer Data Insights, intelligence reports for dealers, were used by 60% of global paying dealers in 2025; in October 2025 the brand added PriceVantage, an action-oriented pricing product. At the group level, full-year 2025 revenue from continuing operations grew 14% for the second consecutive year, to 907.0 million dollars, with 34,409 paying dealers. Management expects the monetized dealer products launched in 2025 to grow roughly 15x in 2026. In parallel, CarGurus divested its wholesale business CarOffer to focus on the marketplace and the dealer software platform.
Results Proof A
Figures from CarGurus' fourth-quarter and full-year 2025 results, published February 19, 2026: official brand press release (revenue 907.0M USD, +14%, 34,409 paying dealers) and the earnings call transcript where the CEO details Dealer Data Insights adoption (60%), the volume of first-party signals (about 500 million per day), and the expected growth of monetized products (about 15x). Financial-results source, with a primary press release and a concordant transcript.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- outil CG Discover Generative search experience where the shopper describes their priorities (budget, lifestyle, options) in natural language rather than through filters, and receives a curated list drawn from the selection of new and used vehicles. Presented as ramping up at the Q4 2025 earnings call.
- outil Dealership Mode Consumer app mode that, through geolocation, adapts the experience to the dealer where the shopper is located: lot inventory, vehicle details, AI-generated pros and cons, a recommendation presented as unbiased, and an all-in price estimate. Announced as live for all app users.
- plateforme Dealer Data Insights Intelligence reports built from shoppers' first-party signals, now central to the dealer workflow; used by 60% of global paying dealers in 2025.
- plateforme PriceVantage First dedicated software product, launched in October 2025, that moves dealers from passive data consumption to action-oriented pricing decisions.
How it runs, concretely
For ops teams-
1Capture of intent on the shopper side customer
The shopper describes their need in natural language in CG Discover, or enters Dealership Mode on a dealer's lot to see inventory, pros and cons, and a recommendation.
-
2Ingestion of first-party signals AI
Each interaction feeds the data foundation; about 500 million signals are ingested per day on average in 2025.
-
3Translation into predictive intelligence AI
The AI turns these signals into real-time insights on demand, pricing, and inventory.
-
4Delivery to dealers data team
The insights are packaged into software products (Dealer Data Insights, PriceVantage) integrated into the dealer workflow; 60% of global paying dealers used Dealer Data Insights in 2025.
-
5Action and monetization customer
The dealer adjusts pricing, inventory acquisition, and marketing based on the insights; the subscription to these products lifts revenue per dealer, tracked quarter after quarter.
Shoppers' first-party behavioral signals (searches, comparisons, on-lot interactions via Dealership Mode). This is the raw material of the predictive intelligence being resold: if the volume or quality of these signals drops, the relevance of the demand, pricing, and inventory insights degrades, and with it the perceived value of the dealer products.
How your customers perceive this type of use
Sourced studiesC'est la famille la moins acceptee : 68% des Americains jugent inacceptable un score financier personnel calcule par algorithme et 67% l'analyse video automatisee d'entretiens d'embauche (Pew Research, 2018). La demande d'explication et de recours est massive : 83% veulent savoir quelles donnees l'IA utilise et 91% veulent pouvoir corriger des donnees erronees (Consumer Reports, 2024). A l'echelle mondiale, seuls 46% se disent prets a faire confiance aux systemes d'IA et 70% jugent une regulation necessaire (KPMG / Universite de Melbourne, 2025).
Acceptance conditions
- Transparence sur les donnees utilisees : 83% des Americains la reclament (Consumer Reports 2024)
- Droit de correction des donnees erronees : 91% le demandent (Consumer Reports 2024)
- Explication de la logique de decision : 44% des consommateurs sont plus enclins a utiliser un agent IA si sa logique est clairement expliquee (Salesforce 2024)
- L'acceptabilite depend du contexte de la decision : 50% des Americains jugent equitable un score de risque criminel pour la liberation conditionnelle, contre 32% pour un score financier applique aux consommateurs (Pew Research 2018)
Red lines
- La decision opaque et sans recours sur l'emploi, le credit ou le logement : 45% tres mal a l'aise pour l'embauche, 39% pour le pret, 39% pour le logement (Consumer Reports 2024)
- Le scoring des personnes a partir de donnees comportementales : 68% le jugent inacceptable pour les offres financieres (Pew Research 2018)
Sources: Pew Research Center 2018 · Consumer Reports 2024 · KPMG / Universite de Melbourne 2025 · Salesforce 2024
How to replicate
Inference, not sourcedData prerequisites
- A high-volume proprietary audience that generates usable behavioral signals (searches, comparisons, interactions)
- A capacity to ingest and structure these signals continuously, at large scale
- An inventory and pricing reference on the supplier side to anchor the predictive insights
- A product content base rich enough to feed a natural-language search
Org prerequisites
- A two-sided marketplace where suppliers (here the dealers) already pay and can move up to software products
- A product and data team able to operate both consumer AI surfaces and a subscription-sold analytics platform
- A clear GDPR legal basis for reusing first-party signals and an AI transparency mechanism (AI Act) on the conversational surfaces
Possible stack
- Generative natural-language search engine over the catalog
- Recommendation and scoring on behavioral signals
- Predictive analytics platform (demand, pricing, inventory)
- Subscription software products delivered into the supplier workflow
The plan, step by step
- Step 1Map the first-party signals already captured by the shopping surfaces and assess which ones predict a decision on the supplier side (demand, price, inventory turnover).Deliverable: Inventory of usable signals, ranked by predictive value.
- Step 2Deploy one or two consumer AI surfaces (natural-language search, contextual mode) that enrich signal capture, with transparency about their AI nature.Deliverable: AI surfaces in production that increase the volume and granularity of the signals.
- Step 3Build a platform that translates these signals into real-time predictive insights, then package a first report for suppliers.Deliverable: First intelligence product delivered to a panel of paying accounts.
- Step 4Integrate the product into the supplier workflow (pricing, inventory) and measure adoption and revenue per account rather than usage alone.Deliverable: Product adopted in operations, reading of incremental revenue per account.
- Step 5Extend the range (pricing, inventory acquisition, marketing) and track the share of suppliers active on these products as a signal of scaling.Deliverable: Suite of monetized products, adoption rate tracked over time.
First step: Identify the densest behavioral signal in your audience (for example what shoppers compare before buying) and derive from it a first intelligence report useful to a supplier, tested with a small group of paying accounts before turning it into a product.
The same pattern, proven in other industries
American Express
real-time fraud detection with deep learning
DoorDash
Causal targeting of promotions (offer personalization)
Lemonade
AI-driven underwriting and claims management
Sources
- S1 CarGurus Announces Fourth Quarter and Full-Year 2025 Results Primary archive pending
- S2 CarGurus, Inc. (NASDAQ:CARG) Q4 2025 Earnings Call Transcript Established press archive pending
- S3 CarGurus Expands Big Deal Brand Campaign, Introducing AI-Powered Dealership Mode and CarGurus Discover Primary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.