We audited 495 US automotive websites on October 9, 2026: 319 car dealers and 176 auto repair shops. Three findings stand out. Four vendors run 85.7% of franchised dealer sites. Half of repair shops and 38.2% of independent dealers show no analytics or ad pixel of their own. And every DealerCenter site blocks GPTBot and PerplexityBot by default, while no dealer blocks OpenAI’s search crawler.
In short:
- Four website vendors (Dealer Inspire, Dealer.com, DealerOn and Team Velocity) run 150 of the 175 franchised dealer sites we analysed (85.7%), and 175 of all 208 sampled franchised dealers (84.1%) once unfetched sites are attributed by DNS.
- 55.4% of franchised dealer homepages (97/175) carry their own Google Ads, Meta, Microsoft or TikTok ad tag in the page code, against 21.5% of independent dealers (31/144) and 18.2% of auto repair shops (32/176).
- 86.2% of franchised dealer vehicle detail pages (119/138) give search engines the car, its price and its VIN in schema markup; only 39.0% of independent dealer vehicle pages (23/59) do.
- All 41 DealerCenter sites we read block GPTBot and PerplexityBot in robots.txt, which puts 29.3% of independent dealers (41/140) behind that block, while 0 of 286 dealer robots.txt files block OAI-SearchBot or ChatGPT-User.
- 50.0% of auto repair shop homepages (88/176) show no analytics tag and no ad pixel of their own, and only 6.2% (11/176) show a price for any named service.
This is Clickem’s own study, built to answer the questions dealers and shop owners keep asking us and that no vendor publishes: who actually builds dealer websites, what is installed on them, and whether AI assistants can read them. Clickem runs Google Ads and Meta ads for car dealerships, auto repair shops, body shops and auto lenders, and this audit is the baseline our automotive marketing agency work starts from. Every percentage below shows its numerator and denominator.
How we ran the audit
We sampled real businesses from Google Maps, fetched their sites from US servers on one day, and read the raw HTML. Here is the full method, so you can judge the numbers yourself.
- Sample source: Google Maps results pulled on October 9, 2026 for “car dealership”, “used car dealer” and “new car dealership” in 30 US metros, and “auto repair shop” in 15 of those metros. Metros ran from New York, Boston and Atlanta to Dallas, Denver, Phoenix, Seattle, Los Angeles and Boise.
- Filters: we dropped listings with no website (957 dealer, 564 repair), non-US listings, out-of-scope categories, and social, marketplace, national chain and corporate location domains (such as Facebook, Cars.com, CarMax, Carvana, Firestone, Midas and Jiffy Lube). One row per domain, so a group running several rooftops on one domain counts once.
- Segments: “franchised” means a Google category of “[Brand] dealer”, or a generic dealer category with an OEM brand in the name and no “used” or “pre-owned”. Every other dealer is “independent”.
- Pool and sample: 3,213 dealer domains (594 franchised, 2,619 independent) and 581 repair domains. A random sample of up to 7 per metro per dealer segment and 14 per metro for repair gave 418 dealers (208 franchised, 210 independent) and 210 repair shops.
- Fetching: requests went through DataForSEO’s US server pool (New Jersey). Per site we fetched the homepage with a standard crawler user agent, again with a Chrome user agent, once with GPTBot’s user agent, plus /robots.txt, /llms.txt and, for dealers, one vehicle detail page (VDP). 30 of the 495 analysed homepages came through a non-US or alternate pool after US requests failed.
- HTML only: no JavaScript was run. Anything loaded later by Google Tag Manager, a consent banner or a platform loader is invisible, so every adoption rate here is a lower bound: “visible in the page HTML”.
- Failures excluded: sites we could not read (bot challenges, rate limits, dead or parked domains) are left out of every denominator rather than counted as “no”.
The fetch outcome matters, because some platforms defend harder against data-center traffic than others.
| Segment | Sampled | Analysed | Not analysed |
|---|---|---|---|
| Franchised dealers | 208 | 175 (84.1%) | 33 |
| Independent dealers | 210 | 144 (68.6%) | 66 |
| All dealers | 418 | 319 (76.3%) | 99 |
| Auto repair shops | 210 | 176 (83.8%) | 34 |
What this means for you: two platforms are under-represented among analysed sites. 17 Dealer.com franchised sites stayed rate-limited, and 39 independent domains resolving to Carsforsale.com served a captcha to our servers. That is why the platform tables below also show shares with DNS attribution, which identifies the host of sites we could not read.
Detection rules, briefly: platforms were identified by fingerprints in the HTML (asset hosts, generator tags, footer credits), and on all 212 sites where both HTML and DNS named a platform, they agreed. “Own” tags exclude any tracking ID found on 3 or more sites run by 3 or more different businesses, since those belong to a vendor or OEM, not the store. We checked 20 sites by hand against their HTML and fixed six false-positive patterns before the final run.
Which website platforms do car dealers use?
Franchised dealers sit on four big vendors; independents sit on low-cost inventory platforms. The two markets barely overlap.
Franchised dealer platforms, analysed homepages and with DNS attribution for the sites we could not fetch:
| Platform | Analysed (n = 175) | With DNS attribution (n = 208) |
|---|---|---|
| Dealer Inspire (Cars Commerce) | 68 (38.9%) | 68 (32.7%) |
| Dealer.com (Cox Automotive) | 39 (22.3%) | 56 (26.9%) |
| DealerOn | 31 (17.7%) | 31 (14.9%) |
| Team Velocity | 12 (6.9%) | 20 (9.6%) |
| Next.js or custom build | 7 (4.0%) | 7 (3.4%) |
| Dealer Venom | 6 (3.4%) | 6 (2.9%) |
| Dealer eProcess | 0 | 3 (1.4%) |
| Other identified platforms | 10 (5.7%) | 10 (4.8%) |
| Unknown | 2 (1.1%) | 7 (3.4%) |
What this means for you: if you run a franchised store, your website vendor’s defaults set most of your technical SEO, tag setup and crawler access before you touch anything. The other identified franchised platforms were WordPress (3), Overfuel (2), Sokal (2), All Auto Network (2) and Remora (1).
Independent dealer platforms:
| Platform | Analysed (n = 144) | With DNS attribution (n = 210) |
|---|---|---|
| DealerCenter | 41 (28.5%) | 41 (19.5%) |
| Carsforsale.com | 10 (6.9%) | 50 (23.8%) |
| DealerCarSearch | 20 (13.9%) | 20 (9.5%) |
| WordPress | 11 (7.6%) | 11 (5.2%) |
| Other identified platforms | 51 (35.4%) | 56 (26.7%) |
| Unknown or not attributable | 11 (7.6%) | 32 (15.2%) |
What this means for you: DealerCenter, Carsforsale.com and DealerCarSearch together host 111 of the 210 independent dealers we sampled (52.9%). The independent market is also long-tailed: the “other” row covers more than 25 platforms, including Overfuel, AutoManager, AutoFunds and GoDaddy at 4 or fewer sites each, and 3 independents on Dealer Inspire.
Across all 319 analysed dealer homepages, the order was Dealer Inspire 71 (22.3%), DealerCenter 41 (12.9%), Dealer.com 40 (12.5%), DealerOn 31 (9.7%) and DealerCarSearch 20 (6.3%). Treat Dealer.com and Carsforsale.com as understated in any analysed-only figure.
How many dealer and repair shop websites have tracking and ad pixels?
Most franchised homepages carry their own analytics and ad tags; most independents and repair shops do not show them. The gap is widest on ad pixels, the tags that let Google and Meta learn who buys.
| Signal in homepage HTML | Franchised (n = 175) | Independent (n = 144) | Repair (n = 176) |
|---|---|---|---|
| Google Tag Manager container (own) | 81.7% (143) | 29.2% (42) | 24.4% (43) |
| GA4 (own ID) | 57.7% (101) | 38.2% (55) | 27.8% (49) |
| Google Ads tag (AW- ID) | 18.3% (32) | 4.9% (7) | 9.7% (17) |
| Meta pixel (own) | 50.3% (88) | 18.1% (26) | 10.8% (19) |
| Microsoft UET | 12.0% (21) | 1.4% (2) | 2.3% (4) |
| TikTok pixel | 2.3% (4) | 1.4% (2) | 0.6% (1) |
| Any own ad pixel | 55.4% (97) | 21.5% (31) | 18.2% (32) |
| Any own analytics | 83.4% (146) | 58.3% (84) | 47.7% (84) |
| Neither own analytics nor own ad pixel | 10.9% (19) | 38.2% (55) | 50.0% (88) |
| Old Universal Analytics (UA-) ID still in code | 3.4% (6) | 15.3% (22) | 11.9% (21) |
What this means for you: if you are an independent dealer or a repair shop running Google or Meta ads, there is a real chance the platform cannot see your sales or bookings at all. Check before you raise budgets. Our guide to Google Ads conversion tracking covers the setup.
Two notes keep this fair. First, Dealer.com loads GA4, Google Ads and chat through its own loader, so an HTML read cannot see them: 0 of 40 Dealer.com homepages showed a GA4 or Google Ads ID. Excluding Dealer.com, the Google Ads tag appears on 23.5% of franchised homepages (32/136) and GA4 on 74.3% (101/136). Second, franchised homepages load a median of 4 Tag Manager containers, and 115 of 175 (65.7%) load 3 or more. Many of those belong to the website vendor, the OEM and third-party tools, which is why we separate “own” tags from shared ones.
The 15.3% of independents still carrying a UA- ID matter for a simple reason: Google says standard Universal Analytics properties stopped processing hits on July 1, 2023, so that code collects nothing. A leftover UA tag is a sign nobody has reviewed the site’s tracking in years.
Do dealer websites use call tracking and chat?
Chat is common on franchised sites; hard-coded call tracking is rare everywhere. That second number surprised us, given how much of a dealer’s business still starts with a phone call.
| Signal in homepage HTML | Franchised (n = 175) | Independent (n = 144) | Repair (n = 176) |
|---|---|---|---|
| Call tracking script | 5.7% (10) | 0.7% (1) | 3.4% (6) |
| Chat or messaging widget | 45.7% (80) | 19.4% (28) | 6.2% (11) |
| Clickable tel: phone link | 79.4% (139) | 86.1% (124) | 74.4% (131) |
What this means for you: a clickable phone number is not call tracking. Unless calls are tracked through Google forwarding numbers, a call tracking provider or imported call outcomes, Google Ads counts form fills and misses the phone, which is often the bigger lead source. The car dealership lead generation guide shows how to put calls, forms and chats on one cost per sale.
The vendors we found: franchised chat was led by Gubagoo (32 sites), Fullpath (29), CarNow (20), Podium (8) and ActivEngage (6). Independent chat was mostly Hammer (14) and Podium (6). Call tracking scripts were CallMeasurement.com (6), Marchex (2), Invoca, CallSource and CallRail (1 each) on franchised sites, CallTrackingMetrics on one independent, and CallRail on all 6 repair shops that had one. We counted vendor scripts only, not a vendor’s name in text, and excluded a chat vendor named in a Dealer.com feature list on every Dealer.com site and a disabled call tracking URL on Team Velocity sites.
How many dealer sites have VDP schema?
On franchised vehicle detail pages, structured data with price and VIN is the norm. On independent VDPs it is the exception, and within any one platform it is all or nothing.
Homepage markup first:
| Homepage structured data | Franchised (n = 175) | Independent (n = 144) | Repair (n = 176) |
|---|---|---|---|
| Any schema.org markup | 98.9% (173) | 81.9% (118) | 75.6% (133) |
| AutoDealer (dealers) or AutoRepair / AutomotiveBusiness (repair) | 97.7% (171) | 44.4% (64) | 27.8% (49) |
| LocalBusiness family or Organization | 97.7% (171) | 79.9% (115) | 60.2% (106) |
| AggregateRating or Review | 5.1% (9) | 5.6% (8) | 14.8% (26) |
| FAQPage | 6.3% (11) | 2.8% (4) | 3.4% (6) |
| WebSite SearchAction | 66.3% (116) | 13.9% (20) | 20.5% (36) |
Then vehicle detail pages. We reached and fetched a VDP on 138 franchised sites and 59 independent sites.
| VDP structured data | Franchised (n = 138) | Independent (n = 59) | All dealers (n = 197) |
|---|---|---|---|
| Car or Vehicle type | 90.6% (125) | 54.2% (32) | 79.7% (157) |
| Offer with a price | 86.2% (119) | 50.8% (30) | 75.6% (149) |
| vehicleIdentificationNumber | 90.6% (125) | 42.4% (25) | 76.1% (150) |
| Vehicle or Product + price + VIN | 86.2% (119) | 39.0% (23) | 72.1% (142) |
By platform, the pattern is stark:
| Platform | VDPs checked | Vehicle or Product type | Price | VIN |
|---|---|---|---|---|
| Dealer Inspire | 65 | 64 | 59 | 64 |
| DealerOn | 30 | 30 | 30 | 30 |
| Dealer.com | 19 | 19 | 19 | 19 |
| DealerCarSearch | 16 | 0 | 0 | 0 |
| Team Velocity | 13 | 0 | 0 | 0 |
| WordPress (other) | 9 | 6 | 4 | 4 |
| DealerCenter | 8 | 8 | 8 | 8 |
| Next.js or custom build | 7 | 7 | 7 | 7 |
| Overfuel | 6 | 6 | 4 | 6 |
What this means for you: VDP markup is a platform setting, not something your team adds car by car. If your platform ships it, you have it on every car; if not, you have it on none. We reached Dealer.com VDPs on only 19 of 40 sites and DealerCenter VDPs on 8 of 41, so read those two rows as small samples.
Be clear about what markup does and does not buy you. Google phased out Vehicle Listing structured data from Search results in 2025, and its AI features guidance says there are “no additional requirements to appear in AI Overviews or AI Mode”. Markup still helps other systems parse a car’s price and VIN. For paid vehicle ads, the rule that matters is visible content: Google’s vehicle ads requirements say the price, VIN and dealership details must be “clearly visible upon loading the page” without extra clicks. Our Google vehicle ads guide covers the full VDP checklist.
Do car dealers block AI crawlers?
Mostly through one platform template, not by choice. All 41 DealerCenter sites ship a robots.txt that blocks GPTBot, PerplexityBot, Google-Extended, CCBot, Bytespider and Applebot-Extended. Franchised dealers almost never block AI crawlers in robots.txt.
The base here is every site whose robots.txt returned a readable file or a 404: 146 franchised, 140 independent and 174 repair. 29 Dealer.com robots files stayed rate-limited and are excluded.
| robots.txt fully disallows the site for | Franchised (n = 146) | Independent (n = 140) | Repair (n = 174) |
|---|---|---|---|
| GPTBot (OpenAI training) | 1.4% (2) | 29.3% (41) | 1.7% (3) |
| OAI-SearchBot (ChatGPT search) | 0.0% (0) | 0.0% (0) | 1.7% (3) |
| ChatGPT-User | 0.0% (0) | 0.0% (0) | 1.7% (3) |
| PerplexityBot | 0.0% (0) | 29.3% (41) | 1.7% (3) |
| Google-Extended | 0.7% (1) | 29.3% (41) | 1.7% (3) |
| ClaudeBot | 1.4% (2) | 0.0% (0) | 1.7% (3) |
| Claude-SearchBot | 0.0% (0) | 0.0% (0) | 1.7% (3) |
| CCBot | 2.1% (3) | 30.7% (43) | 1.7% (3) |
| Any of the 11 AI crawlers we checked | 2.1% (3) | 30.7% (43) | 4.0% (7) |
| File names at least one AI crawler | 37.0% (54) | 43.6% (61) | 6.3% (11) |
What this means for you: the crawlers differ in what blocking them costs. OpenAI says GPTBot crawls content “that may be used in training” and OAI-SearchBot is what surfaces sites in ChatGPT search, so blocking GPTBot alone does not remove a dealer from ChatGPT search. Google says Google-Extended “does not impact a site’s inclusion in Google Search”. PerplexityBot is different: Perplexity says it is “designed to surface and link websites in search results on Perplexity”, so a PerplexityBot block takes a dealer out of the crawler behind those answers.
Across all 286 dealer robots.txt files, OAI-SearchBot, ChatGPT-User and Claude-SearchBot were blocked on 0. The 3 repair shops that block them block every crawler, Google included, with a site-wide “Disallow: /” for all user agents. A newer line, Content-Signal with ai-train=no, appeared on 16 franchised files (11.0%), 14 of them on DealerOn sites.
If you are on DealerCenter, the block is not necessarily wrong. Opting out of training crawlers is a legitimate choice. But the PerplexityBot line has a direct visibility cost, and it is worth deciding on purpose rather than inheriting. Our automotive SEO guide shows which assistants cite dealer websites and which cite third-party sites instead.
The firewall decides more than robots.txt
We also requested each homepage with GPTBot’s user agent string. Read this test carefully: it came from a cloud IP using GPTBot’s user agent, not from OpenAI’s IP ranges, so a 403 shows a firewall rule against that user agent from an unverified address. It does not prove the real GPTBot is blocked, because a firewall can let verified OpenAI addresses through. OpenAI publishes its crawler IP ranges for exactly that verification.
| GPTBot user agent test (cloud IP) | Franchised (n = 140) | Independent (n = 144) | Repair (n = 176) |
|---|---|---|---|
| GPTBot user agent got 403, 429 or a challenge | 52.9% (74) | 54.9% (79) | 9.7% (17) |
| …and Chrome was blocked too (generic bot defence) | 50.0% (70) | 31.9% (46) | 0.6% (1) |
| …but Chrome was served (GPTBot-specific rule) | 2.9% (4) | 22.9% (33) | 9.1% (16) |
The base is homepages that served cleanly to our default crawler user agent (140 franchised, 144 independent, 176 repair). Across dealers, 153 of 284 (53.9%) refused GPTBot’s user agent, but 116 of those also refused a Chrome user agent from the same IPs. Every Dealer Inspire site (71/71) and every DealerCenter site (41/41) behaved that way, which points to generic protection against data-center traffic, not an AI policy. DealerCarSearch was the opposite case: 19 of 20 sites served Chrome but refused GPTBot’s user agent. DealerOn (0 of 31), Team Velocity (0 of 13) and Carsforsale.com (0 of 10) served GPTBot’s user agent normally.
What this means for you: 105 of 276 dealer sites (38.0%) whose robots.txt allows GPTBot still refused its user agent from a cloud IP. Your robots.txt states a policy; your platform’s firewall enforces its own. If AI visibility matters to you, ask your vendor one question: does the firewall verify AI crawlers against their published IP ranges, or block by user agent?
llms.txt is a platform default
Most llms.txt files we found were generated by the website platform, not written by the business.
| Platform | Sites serving llms.txt |
|---|---|
| Dealer Inspire | 71 of 71 (same template on every site) |
| DealerCarSearch | 20 of 20 |
| Team Velocity | 12 of 13 |
| Overfuel | 6 of 6 |
| Dealer Venom | 5 of 6 |
| Dealer.com | 10 of 40 |
| DealerOn | 3 of 31 |
| DealerCenter | 1 of 41 |
| Carsforsale.com | 0 of 10 |
Overall, 65.7% of franchised sites (115/175), 30.6% of independents (44/144) and 36.4% of repair shops (64/176) serve one. Among repair platforms: Kukui 10 of 10, Duda 8 of 8, Hibu 6 of 6, Wix 15 of 20, GoDaddy 6 of 9 and WordPress 7 of 44. What this means for you: do not pay extra for llms.txt. Google says you “don’t need to create new machine readable files, AI text files, or markup” to appear in its AI features, and the file’s presence here tracks the platform, not the effort of the store.
Homepage basics: titles, headings, descriptions
Franchised platforms get the basics right by default. Independent and repair sites are uneven, and the misses are cheap to fix.
| Homepage check | Franchised (n = 175) | Independent (n = 144) | Repair (n = 176) |
|---|---|---|---|
| HTTPS final URL | 100.0% (175) | 99.3% (143) | 93.2% (164) |
| City in the title tag | 91.4% (160) | 56.2% (81) | 64.2% (113) |
| Title under 30 characters | 1.7% (3) | 41.7% (60) | 18.8% (33) |
| Title over 60 characters | 43.4% (76) | 16.7% (24) | 33.5% (59) |
| Meta description present | 98.3% (172) | 68.1% (98) | 76.1% (134) |
| No H1 heading | 8.0% (14) | 34.7% (50) | 22.2% (39) |
| Exactly one H1 | 76.6% (134) | 50.7% (73) | 53.4% (94) |
| Median homepage HTML size | 452 KB | 170 KB | 141 KB |
What this means for you: an independent homepage titled with just the store name (median title length 36 characters, against 59 for franchised) gives Google no city and no “used cars” to match. Putting “Used Cars in [City], [State]” in the title is a five-minute change. Mobile viewport is not reported, because Dealer.com’s server HTML carries no viewport tag and an HTML-only read cannot judge it fairly.
What do auto repair shop websites miss?
Repair shop sites are the least instrumented in the study. Half show no tracking of their own, very few take bookings online, and almost none show a price.
- Tracking: 50.0% (88/176) show neither their own analytics nor their own ad pixel. Wix injects tracking at runtime, so Wix sites look worse than they may be (only 2 of 20 showed own analytics). Excluding the 20 Wix sites, 44.9% of repair shops (70/156) still show neither, our own cut from the same data.
- Booking: 11.9% (21/176) embed an online booking or shop-system widget. Tekmetric led with 8, then Mitchell 1 (3), Square, Steer and AutoVitals (2 each), and Shopmonkey, Acuity, Openbay and Calendly (1 each). Another group only says “schedule service” in text: 57.4% (101/176) had that wording, and 60.8% (107/176) had either.
- Prices: 6.2% (11/176) show a dollar price next to a named service, such as an oil change or brake job, in visible text.
- Schema: 27.8% (49/176) mark themselves up as AutoRepair or AutomotiveBusiness, but 14.8% (26/176) carry review or rating markup, more than either dealer segment.
- Platforms: WordPress 44 (25.0%), Wix 20 (11.4%), Kukui 10 (5.7%), GoDaddy 9 (5.1%), Duda 8 (4.5%), then Repair Shop Websites, Hibu, Squarespace, Net Driven and Autoshop Solutions at 4 to 6 each. 31 (17.6%) were unidentified.
What this means for you: a shop that adds a booking widget and shows starting prices for its three most common jobs is doing something roughly nine in ten local competitors do not. The auto repair shop marketing guide covers which jobs to price and how to measure bookings, and Google Ads for auto repair shops covers the paid side.
Do busier businesses have better website setups?
For independents and repair shops, yes: the stores with the most Google reviews are about three times as likely to run their own ad pixel. For franchised dealers, setup is flat. This is correlation only; it does not show that the setup causes the reviews or the other way round.
| Segment and review tercile | Google reviews | Median rating | Own ad pixel | Own analytics |
|---|---|---|---|---|
| Franchised, low | 2 to 1,886 | 4.4 | 56.9% (33/58) | 75.9% (44/58) |
| Franchised, high | 3,817 to 25,691 | 4.6 | 57.6% (34/59) | 91.5% (54/59) |
| Independent, low | 2 to 65 | 4.6 | 10.6% (5/47) | 36.2% (17/47) |
| Independent, mid | 66 to 230 | 4.5 | 22.9% (11/48) | 50.0% (24/48) |
| Independent, high | 230 to 14,653 | 4.7 | 31.2% (15/48) | 89.6% (43/48) |
| Repair, low | 2 to 96 | 4.6 | 5.3% (3/57) | 15.8% (9/57) |
| Repair, mid | 96 to 250 | 4.7 | 19.3% (11/57) | 49.1% (28/57) |
| Repair, high | 255 to 1,850 | 4.7 | 31.0% (18/58) | 79.3% (46/58) |
What this means for you: ratings barely move across terciles (4.4 to 4.7). Review volume, not rating, is what tracks a measured, advertised website. The franchised middle tercile (51.7% ad pixel, 82.8% analytics) sits between the two rows shown. Median review counts in the full sample were 2,856 for franchised dealers, 117 for independents and 141 for repair shops.
How many Google profile website links are broken or hijacked?
A small but real share. 3 of the 418 dealer websites and 1 of the 210 repair websites linked from Google Maps in our sample now serve gambling spam, so 4 of 628 sampled businesses send profile visitors to a hijacked domain. Another 20 (5 dealer, 15 repair) link to a parked, expired or empty site.
What this means for you: this can happen when a domain lapses and someone else registers it while the Google Business Profile still points to it. It takes thirty seconds to check: open your profile, click your own website button, and set the domain to auto-renew. We are not naming the businesses.
What should you fix first?
Start with whatever stops your ad platforms from seeing results, then the fixes your platform can switch on for every page at once. Here is the order we would work in, by business type.
| Business type | Fix first | Why (our data) | Guide |
|---|---|---|---|
| Franchised dealer | Confirm your own Google Ads conversion tag and import sold deals from the CRM | Google Ads tag visible on 23.5% of non-Dealer.com franchised homepages (32/136) | Google Ads for car dealerships |
| Franchised dealer | Track calls as conversions, not just clicks on a number | Call tracking script on 5.7% (10/175) | Lead generation and cost per sale |
| Franchised dealer | Ask your vendor how its firewall treats verified AI crawlers | 50.0% (70/140) refused both GPTBot’s and Chrome’s user agents from a cloud IP | Automotive SEO and AI answers |
| Independent dealer | Install your own GA4 and at least one ad pixel | 38.2% (55/144) show neither | Used car dealer marketing |
| Independent dealer | Get price and VIN on the VDP, visible and in markup | 39.0% of independent VDPs (23/59) carry vehicle, price and VIN markup | Google vehicle ads |
| Independent dealer | Decide on the PerplexityBot block instead of inheriting it; fix title, H1 and meta description | 29.3% (41/140) block PerplexityBot; 34.7% (50/144) have no H1 | Automotive SEO |
| Auto repair shop | Install analytics and an ad pixel you own | 50.0% (88/176) show neither | Auto repair shop marketing |
| Auto repair shop | Add online booking and starting prices for common jobs | Booking widget on 11.9% (21/176); prices on 6.2% (11/176) | Auto repair shop marketing |
| Any business | Click your own Google profile website link and set the domain to auto-renew | 4 of 628 sampled profile links lead to gambling spam | Automotive SEO |
What this means for you: none of the top fixes is a redesign. Most are a tag, a setting or a question to your website vendor. For the paid side, the automotive Google Ads benchmarks show what clicks and leads cost once tracking works.
Use this data
You are free to cite, quote or chart any number on this page, in articles, decks or vendor evaluations, as long as you link to this page as the source. Suggested citation: “Clickem’s 2026 audit of 319 US dealer websites and 176 auto repair shop websites, October 9, 2026 (https://clickem.io/car-dealer-website-study/)”. If you are a journalist or researcher and want a cut we have not published, such as one metro region or one platform, ask through our contact page.
Limits of this study
- HTML only. Tags fired inside Google Tag Manager, server-side tagging, tags held back by consent banners and widgets added by a loader are invisible. Every adoption rate is a lower bound.
- Dealer.com is understated on tags and chat. Its loader injects GA4, Google Ads and chat after page load. Dealer.com and Carsforsale.com are also under-represented among analysed sites, which is why platform shares are given with DNS attribution.
- The GPTBot test is not OpenAI’s crawler. We used GPTBot’s user agent from a cloud IP. A 403 shows a firewall rule, not proof the real crawler is blocked.
- One page per type. One homepage and at most one VDP per site. A store’s other VDPs may differ, though within a platform they rarely do.
- Who is in the sample. Businesses with no website on their Google profile, national chains and marketplaces were out of scope. Sites that failed to load were excluded, not counted as “no”.
- Association, not cause. The review terciles show which setups go together, not what drives results.
- A snapshot. Everything was fetched on October 9, 2026. Platform defaults change, sometimes overnight for thousands of sites.
Common questions
Which website platforms do car dealers use most?
Among franchised dealers we analysed, Dealer Inspire ran 38.9% of sites (68/175), Dealer.com 22.3% (39), DealerOn 17.7% (31) and Team Velocity 6.9% (12), 85.7% together. Independents mostly use low-cost inventory platforms: DealerCenter, Carsforsale.com and DealerCarSearch host 111 of 210 sampled independent dealers (52.9%) once DNS attribution is included.
Do car dealers block AI crawlers like GPTBot?
Mostly not by choice. In robots.txt, 1.4% of franchised dealers (2/146) block GPTBot against 29.3% of independents (41/140), and all 41 of those independents are DealerCenter sites sharing one template that also blocks PerplexityBot and Google-Extended. No dealer robots.txt in our sample (0/286) blocks OAI-SearchBot or ChatGPT-User, which OpenAI uses for ChatGPT search and user actions.
How many dealer websites have VDP schema with price and VIN?
86.2% of franchised vehicle detail pages (119/138) and 39.0% of independent ones (23/59) carried vehicle or product markup with a price and a VIN. It is set by the platform: DealerOn, Dealer.com and DealerCenter VDPs we checked all had it, while DealerCarSearch (0/16) and Team Velocity (0/13) VDPs had none.
How many car dealer websites have a Meta pixel or Google Ads tag?
A Meta pixel of the store’s own was visible on 50.3% of franchised homepages (88/175) and 18.1% of independents (26/144). A Google Ads tag was visible on 18.3% of franchised (32/175) and 4.9% of independents (7/144). These are lower bounds, since tags fired through Tag Manager or a platform loader do not show in page HTML.
Does a 403 to GPTBot mean ChatGPT cannot read my site?
Not necessarily. Our test sent GPTBot’s user agent from a cloud IP, not from OpenAI’s addresses, and most of the sites that refused it also refused a normal Chrome user agent from the same IP. GPTBot also covers training, not ChatGPT search, which uses OAI-SearchBot. Ask your vendor whether the firewall verifies crawlers by their published IP ranges.
Do auto repair shops need online booking on their website?
Only 11.9% of repair shop homepages we read (21/176) embed a booking or shop-system widget, though 57.4% (101/176) mention scheduling service in text. A widget connected to your shop system lets customers book without a call and gives you a booking you can count as a conversion. That only helps if analytics is installed, and half of the shops we read (88/176) showed no analytics or ad pixel of their own.
If you want to know where your own setup sits against these numbers, start with your tracking. Clickem runs Google Ads and Meta ads for car dealerships, auto repair shops, body shops and auto lenders, and the first step with any account is a review of what is being measured and where the spend goes. Our automotive marketing services page explains how that works, and you can ask for an account review whenever it suits you.
