CarFinder Research and Sources
A comprehensive evidence base supporting the case for conversational AI in automotive search. Compiled April 2026 from industry-leading studies including Cox Automotive, Google Think Auto, Baymard Institute, McKinsey, and more — covering buyer behavior, search UX, AI performance, and market trends.
Research Compendium
April 2026
Car Buyer Behaviour and Decision-Making
The modern car buyer enters the market with remarkable openness — and the data confirms it consistently across multiple major studies. The vast majority of shoppers have not made up their minds before they start searching, making the discovery and guidance experience critical to conversion.
71%
Open-Minded Buyers
Entered the process unsure of what vehicle to buy (Cox Automotive, 2025)
~900
Digital Interactions
Typical number of digital touchpoints in a single car-buying journey (Think with Google)
14.3h
Online Research Time
Average hours spent online during the car-buying journey (Demand Local)
2
Dealership Visits
Average number of dealerships visited before purchase — down from 4.5 (Cox Automotive)
Undecided at the Start
Only 29% of buyers were certain about a specific vehicle when beginning their search — down from 37% in 2020. Over 70% of Cars.com shoppers are undecided on make and model. 6 in 10 car shoppers enter the market unsure which car to buy (Google Think Auto).
Key Behavioural Signals
  • 52% undecided about repurchasing the same brand
  • 63% of European consumers open to switching brands
  • 66% considered both new and used vehicles (up from 57% in 2024)
  • Millennials consider 18–20 different vehicle models before purchasing
  • 92% of all car purchases begin with online research
  • Only 1% of buyers are fully satisfied with the overall buying experience (McKinsey)
Filter-Based Search: Drop-Off and Abandonment
Filter-based search — the dominant paradigm across automotive marketplaces — is fundamentally broken for the majority of users. Baymard Institute, CXL, and UX benchmarking studies reveal systemic failures that actively drive shoppers away.
67-90%
Abandonment Rate
On sites with mediocre product list and filter implementations (Baymard Institute)
57%
Poor Filtering UX
Of e-commerce sites have poor or mediocre filtering UX (CXL)
16%
Good Filter Experience
Only 16% of major e-commerce sites provide a genuinely good filtering experience (CXL)
53%
Mobile Abandonment
Of mobile users abandon if filtering feels tedious (Wizzy)
Structural Gaps in Filter Design
  • 42% of sites lack category-specific filter types for main product categories (CXL)
  • 32% of top 60 e-commerce sites don't display an overview of applied filters (Baymard)
  • 36% of sites have usability issues severe enough to be described as "downright harmful" (Baymard)
Automotive-Specific Evidence
  • Cars.com filters described as "just too much" by UX test participants (MeasuringU)
  • Autotrader scored in the 69th percentile — lowest among 8 automotive sites tested (MeasuringU)
  • Only 17–33% abandonment on sites with good filtering — the gold standard rarely reached
Search Abandonment and Lost Revenue
When shoppers can't find what they're looking for, they don't wait — they leave. The financial and conversion consequences of poor search are enormous, and the data from Google Cloud and retail benchmarks underscores just how much revenue is being lost across the industry.
$300B
Annual revenue lost by US retailers due to poor search experiences (Google Cloud)
53%
Of US consumers abandon and go elsewhere when they can't find an item via search (Google Cloud)
88%
Of consumers say good search is "very important" or "essential" to their experience (Google Cloud)
12%
Only 12% of shoppers get exactly what they're looking for every time they search (Google Cloud)

64% of lost conversions occurred because users never even started searching — overwhelmed before they began. (GetGoBot)
Further compounding the problem: 12% of shoppers immediately abandon a website when search fails, while 48% finish their purchase on a competitor's site. 60% of visitors quit entirely if they cannot locate what they were looking for (NewMedia).
The Paradox of Choice
Academic and commercial research converges on a clear finding: more options do not lead to more conversions. The landmark Iyengar & Lepper study and subsequent commercial replications show that reducing choice dramatically increases purchase intent — a principle with direct implications for how automotive inventory is surfaced.
The Landmark Study
24 options produced a 3% conversion rate. 6 options produced a 30% conversion rate — a 10x difference. (Iyengar & Lepper, Columbia/Stanford, via CXL)
Conversion peaks at 4–6 options and drops 10–20% once choices exceed 12–15 items. (Cognitive Clicks)
Commercial Evidence
Dr. Scholl's Case Study: Reducing from 5 products to 3 resulted in 38% higher revenue and 58% more conversions. (BuildGrowScale)
Miller's Law: The human brain can process only 7 items (±2) at any given time — a hard cognitive ceiling with direct implications for inventory browsing. (Cognitive Clicks)

The implication for automotive marketplaces: surfacing 200+ vehicles in a filtered list is not helpful — it is a conversion killer. Intelligent curation to a shortlist of 4–6 highly relevant matches is the evidence-backed approach.
Conversational AI vs. Filter Search: Performance Data
The most compelling evidence for conversational AI comes from live deployment data. Cars.com's Carson AI assistant and broader e-commerce benchmarking studies show consistent, significant lifts in conversion, engagement, and lead quality when conversational AI replaces or augments traditional filter-based search.
4x
Conversion Lift
AI chat users convert at 12.3% vs. 3.1% for non-engaged shoppers (BigSur AI)
47%
Faster Purchases
AI assistance helps shoppers complete purchases faster (BigSur AI)
25-50%
AOV Increase
Average order value rises through context-aware AI recommendations (BigSur AI)
23%
Overall Conversion
Higher overall conversion on websites with conversational AI (BigSur AI)
Carson (Cars.com) Live Data
~30% Higher Conversion
From search to vehicle detail pages for Carson users vs. standard search (Cars.com Investor Release)
3x More Vehicles Saved
Carson users save 3x more vehicles and generate 2x more leads per session (Cars.com)
2x Return Rate
Carson users return to the platform twice as often as other shoppers (Cars.com)
84% Satisfaction
Of mostly-digital buyers who engaged AI assistants reported high satisfaction (Cox Automotive 2025)

73% of AI users say it saves time to have AI turn conversational queries into results. 19% of all buyers used AI tools in 2025 — the first year Cox Automotive tracked this metric. (Cars.com AI Survey; Cox Automotive)
UX and Search Design Research
UX researchers at Nielsen Norman Group and academic studies in consumer psychology provide the theoretical foundation for why filter-based automotive search underperforms — and what a better paradigm looks like. The core insight: buyers think in needs and lifestyle, not technical specifications.
Mental Model Mismatch
People's search mental model is shaped by conversational engines like Google. Filter-based systems cannot accommodate this expectation. Users expect to describe what they want — not configure a matrix of dropdowns. (Nielsen Norman Group)
Buyers Think in Needs
Actual buyer criteria are: lifestyle fit, safety, comfort, economy, and aesthetics — not make, model, or fuel type. Yet most automotive search interfaces lead with exactly those technical dimensions. (PMC / NCBI Study)
Filter Relevance Principle
Filters should prioritize relevance to user needs — e.g. "Interior Capacity" — over raw technical specifications. Most sites invert this principle, leading to cognitive overload. (Nielsen Norman Group)
EV Complexity Acknowledged
Autotrader UK explicitly acknowledges that EV-specific filters add significant complexity to an already difficult search journey — further validating the need for guided, conversational approaches. (Autotrader UK Insight Blog)
Swedish Automotive Industry Trends 2026
Data from Framtidens Bilhandel's 2026 outlook reveals both the gaps and the opportunities in the Swedish automotive market. Lead leakage, slow response times, and AI adoption patterns point toward a pivotal moment for dealers who move early.
Lead Management Crisis
23.5% of leads don't receive a response within 24 hours. A further 13.3% disappear entirely before being registered in the CRM. This represents a structural revenue leak that AI-driven lead management directly addresses.
AI-driven lead management delivers 33% shorter sales cycles and 40% higher conversion to booked viewings. (Framtidens Bilhandel)
AI Adoption Signals
  • 60%+ of first-response interactions handled by AI agents projected by end of 2026
  • 85% of salespeople with AI agents say it frees them for higher-value tasks (Salesforce)
  • 81% of US car dealers plan to increase AI budget in 2025 (Fullpath)
  • 19% of all buyers / 25% of new car buyers used AI tools in the process (Cox Automotive)
  • 80% of buyers and sellers are open to AI involvement; final deal typically in-person
  • 63% of car buyers want a combined online + physical buying experience
First-Party Data and Personalisation
The business case for investing in first-party data activation and personalised experiences is well-established at the highest level of commercial research. For automotive marketplaces, this represents a significant untapped revenue lever — particularly as third-party cookie deprecation reshapes digital advertising.
1.5–2.9x Revenue Uplift
Achieved by brands with advanced first-party data activation strategies, compared to peers with limited data maturity. (BCG / Google)
5–8x Higher ROI
Personalised campaigns deliver 5–8x higher return on investment versus generic, untargeted campaigns. (BCG / Google, cited in Framtidens Bilhandel)

Conversational AI search is a natural first-party data engine: every interaction captures intent signals, preference data, and behavioral patterns that can power personalised remarketing, inventory recommendations, and dealer lead prioritisation.
Amazon / E-Commerce Search Benchmarks
Amazon's search performance data provides the gold standard benchmark for what effective search can deliver. These figures establish the upper bound of opportunity and set a clear target for what automotive marketplaces should aspire to — particularly as car buying increasingly mirrors e-commerce behavior.
Amazon Conversion Spike
When shoppers on Amazon use the search function, conversion spikes 6x — from 2% to 12%. This single data point illustrates the transformational impact of effective search on revenue outcomes. (NewMedia)
Search Users Buy More
Shoppers who use search buy 10–25% more items per order than those who simply browse. The intent signaled by search behavior translates directly into higher basket value — a principle equally applicable to automotive leads. (NewMedia)

The parallel for automotive: a buyer who uses natural-language search to describe their needs is signaling purchase intent. Capturing and converting that intent efficiently — rather than losing them to filter overwhelm — is the core opportunity CarFinder addresses.
Sources and Methodology
This research compendium was compiled in April 2026. All links were verified at time of research. Where sources reference paywalled reports, summary data has been cited from publicly available extracts only.
01

Compiled April 2026. Links verified at time of research. Some sources reference paywalled reports — summary data cited from publicly available extracts only.