Shoot at least 11 photos per vehicle, and go up to about 30 where you can. AutoTrader.ca's marketplace data found used listings with 11 or more photos generated 205% more VDP views and 185% more leads than listings with 10 or fewer, and they recommend up to 30 when possible. The catch nobody mentions: 30 unoptimised photos will make that page slow enough to lose the buyer before the gallery renders. You can have both, but only if the gallery is built properly.
I build inventory platforms for dealers, so I've had to resolve this exact tension in code rather than in a blog post. This article is the resolution: what the photo-count evidence actually says, where it gets misread, and the specific image pipeline that lets a 30-photo vehicle page load fast.
What the photo-count data actually says
The most concrete public numbers come from AutoTrader.ca's own marketplace warehouse, analysed over January to July 2022. For used listings, 11 or more photos produced 205% more VDP views and 185% more leads versus 10 or fewer. For new listings, the same comparison produced 211% more VDP views and 167% more leads. Their recommendation is 11 as a floor and up to 30 where possible.
The same guidance says 75% of consumers find multi-angle exterior and interior photos useful, and 42% find video walkarounds purposeful. It also lists what to actually capture: all four corners plus head-on front and rear for a full 360, close-ups of hood, headlights, wheels, tyre tread depth and badges, the full dash with infotainment and climate controls, all four door cards, every row of seats, and the boot, engine bay and any unusual feature.
Three honest caveats before you treat 205% as a law of nature. First, this is marketplace listing data from AutoTrader.ca, not data from dealer-owned websites — the behaviour is probably similar, but it isn't the same measurement. Second, it's from 2022. Third and most important, it's correlational. A dealer who shoots 25 photos is usually also the dealer who reconditioned the car properly and priced it sensibly. AutoTrader's own guidance makes this point implicitly by opening the photo section with an instruction to recondition the vehicle before shooting anything. Photo count is partly a proxy for a dealer who does the whole job well.
So the useful version of the finding is: thin galleries signal a careless listing, and buyers treat them accordingly. Eleven is the point where a listing stops looking thin.
Why 30 photos usually breaks the page
Largest Contentful Paint measures the render time of the largest image, text block or video visible in the viewport. Google's threshold for good is 2.5 seconds or less, measured at the 75th percentile of page loads, split between mobile and desktop; above 4.0 seconds is poor. On a vehicle detail page, the LCP element is almost always the hero photo of the car.
That means your gallery is your Core Web Vitals score. And the typical dealer gallery does everything wrong at once:
The photos come straight off a phone or a DSLR at full resolution — often 4000px wide and 4–8 MB each — and get displayed in a container 800px wide. The browser downloads every pixel and throws most of them away. All 30 load on page load rather than on demand, because the gallery library requests them upfront. The images are JPEG rather than AVIF or WebP. There are no width and height attributes, so the layout jumps as each one arrives, which damages CLS as well. And frequently the gallery is JavaScript-rendered, so the hero image can't even start downloading until a script has parsed and executed.
Do the arithmetic on a typical case. Thirty photos at 4–6 MB each is somewhere between 120 and 180 MB of images if the page requests them all at once, and even the hero photo alone at that size is a heavy download on a mobile connection. Run the same photo through the pipeline below and the file the browser actually fetches for the hero is a fraction of that: an AVIF sized for the screen, typically a few hundred KB or less. Multiply that difference across the whole gallery and you have the entire gap between a page that passes the 2.5 second LCP threshold and one that doesn't.
You can see where your own pages stand in a couple of minutes. Open any vehicle page on your site in PageSpeed Insights, switch to the mobile tab, and look at three numbers: the LCP time, the total image weight, and the number of requests. Those are the three this article moves.
The pipeline that fixes it
Five decisions, in the order they matter.
1. Only the first photo loads eagerly. The hero image gets fetchpriority="high" and no lazy loading. Every other photo in the gallery gets loading="lazy". The thumbnails strip loads as small thumbnails — not as full images scaled down with CSS, which is the single most common mistake I see.
2. Serve modern formats with a fallback. AVIF first, WebP second, JPEG last, via <picture> so each browser takes the best format it supports:
<picture>
<source srcset="car-01.avif" type="image/avif">
<source srcset="car-01.webp" type="image/webp">
<img src="car-01.jpg" alt="2019 Golf GTI, front three-quarter, driver's side"
width="1200" height="800" fetchpriority="high" decoding="async">
</picture>3. Generate size variants and let the browser choose. A phone at a 3x device pixel ratio and a desktop monitor need genuinely different files. srcset with sizes lets the browser pick before it downloads anything:
<img src="car-01-800.jpg"
srcset="car-01-480.jpg 480w, car-01-800.jpg 800w,
car-01-1200.jpg 1200w, car-01-1800.jpg 1800w"
sizes="(max-width: 700px) 100vw, 800px"
width="1200" height="800"
alt="2019 Golf GTI, front three-quarter, driver's side">Note the interaction with pixel density: if you optimise only for high-DPR screens you end up shipping a heavy file to a small phone for no visible benefit. Size variants plus sizes handles both axes.
4. Always set width and height. This reserves the space before the file arrives and prevents the gallery from shoving the price and the enquiry button around the screen as photos load. It costs nothing and it's the cheapest CLS fix that exists.
5. Resize on upload, not on display. This is the part that actually survives contact with a dealership. The person photographing cars should not have to think about file size. The upload pipeline takes whatever comes off the phone and generates every format and every size variant automatically. If compliance with your image policy depends on staff discipline, your image policy will be dead within a month.
In practice the flow looks like this: a staff member uploads a 6 MB photo straight off a phone, the server keeps the original, and within moments generates AVIF, WebP and JPEG copies at 480, 800, 1200 and 1800 pixels wide. The gallery only ever references the generated files, so each visitor's browser downloads one small, correctly sized file per photo, and the 6 MB original never touches the public page.
What this means for your dealership
Shoot 11 photos minimum, aim for 20–30 on anything above your average price point, follow a fixed shot list so the set is consistent across your whole inventory, and recondition before you shoot. Then test one of your own VDPs on PageSpeed Insights, on mobile, and look at the LCP figure against the 2.5 second threshold.
If your platform can't tell you what it does with your uploads — whether it converts to AVIF or WebP, whether it generates size variants, whether the gallery lazy-loads beyond the first image — that's a reasonable question to put to your account manager, and the answer tells you a lot. This is one of the areas where the gap between platforms is real and measurable rather than a matter of marketing claims, which is part of the broader custom versus platform decision.
Where I'd tell you not to worry about this
If your total inventory is under roughly 20 vehicles and your traffic is mostly people who already found the car on a marketplace and clicked through, your image pipeline is not your bottleneck — your photography is. Hire someone to shoot the cars properly before you pay anyone to optimise the delivery of bad photos. A fast page full of dark, cluttered, badly-angled images sells nothing.
And if you're on a platform whose galleries already pass Core Web Vitals, this is not a reason to leave. Test it before you assume otherwise.
Sources
- AutoTrader.ca, Lift Your Leads with Better Merchandising (1 Sept 2022) — photo count, VDP view and lead figures, shot list, video and multi-angle research. Data: AutoTrader.ca Data Warehouse, Jan–Jul 2022; Vehicle Purchase Behaviour Research, Dec 2021.
- web.dev, Largest Contentful Paint (LCP) (last updated 4 Sept 2025) — 2.5s good / 4.0s poor thresholds, 75th percentile guidance, which elements count as LCP candidates.
- DebugBear, The Ultimate Guide to Responsive Images on the Web —
<picture>,srcset/sizes, device pixel ratio.
Send me one of your vehicle pages and I'll run the image audit and send you the waterfall. No pitch attached.