Close icon to exit the modal
Search icon
Close icon to exit the modal
Sign up to Publift Scoop
Subscribe to the Publift newsletter for all the latest news, updates and insights from across the industry.
Play video
Play icon
Play video

Layout Teardown: Episode 3 — Cars.com's Sophisticated Ad Strategy

Layout Teardown: Episode 3 — Cars.com's Sophisticated Ad Strategy

Next arrow icon
Back

Layout Teardown: Episode 3 — Cars.com's Sophisticated Ad Strategy

Layout Teardown is a video series where Ben and Brock breakdown the ad strategy and website layout of popular publishers. In this week's episode, we're taking a look at Cars.com, one of the biggest auto marketplaces in the US. Ben and Brock speak about the detailed insights and suggestions for optimising cars.com’s ad revenue. From tweaking the desktop layout to maximising mobile monetisation with various ad units, stay tuned to learn valuable tips to develop your ad strategy and improve overall performance.

Key Takeaways

  • Ads come second to leads, and that is the right call. Cars.com makes its real money when a buyer fills out a form and gets passed to a dealership. A single lead is worth far more than an ad click, so the ad strategy stays conservative.
  • The hero banner is being wasted. The big desktop billboard is a premium slot that car brands would pay well for, but eager buyers scroll straight past it before it has finished loading. Viewability sits around 40% to 50% as a result. Keeping it pinned for just the first 100 pixels of scroll would fix that almost immediately.
  • The image gallery is a huge missed opportunity. Buyers spend a lot of time flipping through vehicle photos, which makes it one of the most attentive moments on the entire site. Currently, that space is ad-free. Slipping an ad into the gallery rotation every ten slides would deliver near-perfect viewability without meaningfully disrupting the experience.
  • Car brands have video budgets burning a hole in their pocket. Manufacturers spend enormous amounts on video advertising, and Cars.com is sitting on the perfect placement to capture some of that spend. An outstream video unit inside the photo gallery, matched to the exact make and model the user is already looking at, would be a compelling pitch to any automotive brand.
  • Mobile is being underleveraged. There is no sticky footer, no 300x250 within the listings, and no real effort to monetise the casual browser who is just doing research on their phone. That audience may never fill out a lead form, but they are still valuable and currently generating almost nothing.
  • The map is creating a lot of noise. Every time a user drags or zooms, the page reloads the listings and fires off a fresh round of ad requests. Most of those go nowhere. People interact with maps constantly, so this adds up quickly and needs to be cleaned up before it distorts performance data.

What makes Cars.com interesting from an advertising perspective?

Ben: Hey and welcome to episode 3 of Layout Teardown, where we bring you through the layouts and programmatic strategy of some of the biggest publishers globally. Today we're going to walk through Cars.com, which is one of the biggest auto marketplaces in the US. This is an interesting one because it's a different type of publisher — they have a different business model. They're generating most of their revenue through basically leads: people buying cars online. And really advertising — or "media" as they call it — is normally seen as a supplementary business. Now we know from a lot of our bigger publishers that we work with that it's a growing area within their businesses, but they need to make sure that those ads aren't impacting what's really driving the revenue. So pretty unique from that perspective — also just in terms of the user type, the data that they have, and the auto segment having so much revenue and money in it in terms of advertising. There are just a lot of ways that they can make money. So it's an interesting type of website to look at from that perspective. My name is Ben.

Brock: Hey, I'm Brock, and we will be jumping into this one. We'll kick it off by looking through the site on desktop and mobile, showing you guys what we think is good, where we see some improvements, and generally giving you guys some tips as we go.

What does Cars.com's traffic look like, and what does the revenue mix between leads and media look like?

Ben: Okay, so we are now going to jump in and have a look at the data. Obviously we haven't seen any of the real data, but we're using a few Chrome extensions and tools to give us an insight on what the traffic looks like. We're pulling up SimilarWeb within here, so we can see they've got a really significant amount of monthly page views coming through on site and a really strong visit duration — this is kind of what you would expect on a marketplace like this. You've got people browsing and looking through sites a lot. In terms of geography, Brock — we've got almost all traffic really coming from the US.

Brock: Yeah, it makes sense to me. I would imagine these guys, who were born and bred in the United States, have got a US-centric strategy. I think that will spill into their media strategy as well. When you think about why Cars.com existed in the beginning, they would have been all about just getting the user to actually buy that car. So they have — I'm guessing — a very substantial lead product that goes out to all the dealers around the States. And now what you're seeing is a blended media strategy that's very much focused on the US. Plus I guess for users, they want to go to the place that has the most cars, the most choice. So it makes sense that you'd get that wheel of growth — the more cars you have, the more users you get, that kind of stuff. That's easier to do in one country.

Ben: The conversions as well I would add to that — if they've got a nice wheel that's churning out users into these dealerships, or even an online form, I'm sure a lead cost form on the dealer side would be very, very lucrative.

Brock: Definitely. And then just on traffic sources, it's pretty split between search and direct. So that's interesting — they've obviously got a pretty good search strategy in place, and I guess there are people coming to the site anyway, but it seems pretty heavily on the search side from what I would have anticipated.

Ben: Yeah. "Cars.com" — that's the complete URL — I feel like that's going to be feeding into their search strategy.

Brock: Definitely. And I'm sure with that kind of brand recognition they're going to have a lot of people coming back, and that builds up your direct channel over time. So in terms of the makeup — what does the revenue breakdown look like between their actual lead product and then within media, what's that like direct versus programmatic?

Ben: Yeah, I would imagine it follows a very standard marketplace model, even outside of the automotive space. They would have been born directly looking to get that lead revenue and would have probably been 100% automotive revenue driven — as in sales. And then over time some media starts to creep its way in: put an ad somewhere, put an ad next to the gallery or whatever they do. And then very, very quickly the media part becomes like a massive piece of their P&L. I would think now it's probably getting closer to 50/50, although that's fairly much guesswork.

Brock: Yeah. But when a company like that has to go through such a transformation, it becomes really important to have a nice blended strategy and a blended team — and that's not that easy to actually execute. Really any website where you've got such close buyer intent, where people are literally searching and in the idea of making a purchase — advertisers want to be across this. And unlike, say, property sites, which are similar, you don't have ads for one house, whereas with cars you've got brands and you've got specific vehicles. So there's a lot of revenue and money put behind these kinds of ads.

Ben: So there's a lot of revenue available. As we know, the product guys don't particularly like the ads because they think it's pulling away from the end objective. But we've seen some really slick strategies that allow you to keep ads relatively simple but use all that really rich buyer intent data to drive revenue — not just open auction, but selling it in a variety of ways.

Brock: Yeah, I would be very surprised if Cars.com haven't really spent millions of dollars developing out their audience strategy, and that is then fuelling their media strategy. If they're able to tap into that buyer intent and collect it in a privacy-safe way and make it available to buyers, I have no doubt they'd be able to turn a tidy profit out of that.

Ben: Yeah. If I could tell Volvo that this person was interested in a Volkswagen — that's some super valuable information, and I'm sure you could charge a pretty penny for it as well.

What are the two most important page types for ad revenue on Cars.com?

Ben: Okay, so we will now jump into looking through the layout specifically. Really on any of these types of sites, there are two core areas that drive 70 to 80% of the revenue. We've got our listing section — or search section — where we've searched for a type of vehicle that we want. And then we've got the details page: once we've gone through the listings page and we're looking at more information. These types of sites will have a lot of other content, but really that's just about bolstering traffic, SEO, experience — those kinds of things. The bread and butter, the revenue, is coming from these two page types. And really from the search listings from a media perspective — the reason being, once you get to a details page, you're at an even higher phase of intent, and so the model of getting leads is even more important at that point. So it's normally a bit more protected.

Brock: Yeah. I think conceptually it's probably basic enough marketing funnel stuff. You've gone from search/purchase consideration into buying intent into action behaviour, and hopefully for these guys — purchase.

Ben: Exactly.

What does the desktop search/listings page look like, and what are the optimisation opportunities?

Ben: So we'll start off on the search page. First thing we can see is we do have a header bidding auction running — they've got a bit of demand in there. From your analysis, they've actually done their own Prebid build here, by the looks of it.

Brock: That's what it looks like. I also noticed that you must have a VPN on because obviously we're in Australia and I didn't see any Prebid. So it looks like they have a localised strategy — that might just be because of some of the requirements with some of their OEM deals that they have to manage, and of course privacy around the world is not as easy and simple as it can be.

Definition — OEM: Original Equipment Manufacturer; in this context, major car brands like Ford, Toyota, or BMW that advertise on automotive platforms.

Ben: That's true. We've seen a lot of publishers just cut off bothering with European ads just because of the effort involved. Plus we've got 93% coming from the US and Canada, so maybe they've just decided the risk isn't worth the return.

Brock: Yeah. The other thing is they could really use that almost 10% of non-US traffic as just a test bed. Looking at SimilarWeb, they're getting some Australian traffic — they're going to get decent CPMs, they'll have enough volume to actually try something. That might be something they look at in the future.

Ben: Yeah. First thing we see obviously is this huge billboard — a 970x250 — really nice position. I would guess this is sold out most of the time direct. That's kind of a pretty premium spot.

Brock: Yeah, I think so. I think this is your centrepiece. I would imagine they have a homepage solution as well that they might be shipping out for direct. But you know, search is where the action happens. It's above the fold, so it inherently has a bit of extra value. It's a 970x250, and if you've got OEMs looking to put in a beautiful, nice rich media ad, you can charge a pretty penny for that for sure.

Ben: Any optimisation opportunities? The major thing will be speed in this case. If you want to refresh that one there — they've done a pretty good job at optimising the page, and as you can see, you're also loading a well-designed and probably heavy creative. You might have a bit of user banner blindness from that one — off they go, they start scrolling, they look at cars. There was a bit of a delay there. I guess the one tweak we could do is try to make this unit a bit more valuable and a bit more viewable. From similar units that we've seen, particularly at that kind of loading speed, you're at kind of a 40 to 50% viewability rate. So it looks like it should be a lot higher than that.

Brock: What we've seen other car marketplaces do is just have a small sticky function — so that unit sticks for 100 pixels — so we're just a little bit more likely to get that user's eyeballs on it. It rolls away straight away and we're back to the same content. I think that would probably be a small piece we could do there, but it could have a big impact. Even on the direct side, if they're serving roadblocks across a search result or key category, they could have that headline billboard and then like a 728x90 sticky that runs for the first say five or six images or links — that might be a way to keep that nice premium-looking top-of-page experience and then of course get a little bit of extra cash from a nice viewable unit.

Ben: Next we've got our within-listing units. It looks like a lot of these are being lazy loaded as well, so we've only got certain ones actually loading as we scroll. Lazy loading actually looks pretty good — it seems to be catching them. We've got this super nice native unit — presumably direct?

Brock: It looks like it, yeah. You can count how many actual listing positions there are — so in four. They've got a nice car image, all the text matches, same colour for the call to action, and it's served through GAM as well. Keeps it nice and centralised, easy for the media teams to manage — looks pretty slick.

Definition — GAM (Google Ad Manager): Google's ad server used by publishers to manage, serve, and report on ad inventory.

Ben: And then we've got our sidebar sticky on the left-hand side — a 300x600. Really nice. We've got this unit that looks like a shoppable format, gallery-type unit — maybe custom?

Brock: I'd be direct only on that, and I'd get a very nice CPM from it. It's good that they are allowing something interactive in there as well — not only does it break up the content, it actually gives the brand the ability to be quite creative and try and drive some sort of return on investment on what they're actually buying from Cars.com.

Ben: And clearly they've bought out the whole section here. Any optimisations you can see at the moment?

Brock: I like this placement, by the way — just at the kind of page transition section. We often see this is like the second most valuable location on these types of pages, just because users very often are changing page. Normally the further down an ad is, the less likely it is to be viewed — but for this unit it's actually got pretty high viewability and also high engagement. You've got users clicking and making some kind of action here, and they're more inclined to click on or engage with that unit at that point. Anything else you would change, do, or add differently?

Ben: Yeah, it's very hard to say when you've got a nice roadblock there. But the real question would be: how often are buyers going through the direct channel and then taking everything? So we want to be looking at what our sell-through rates are from a direct point of view and making sure that we give at least very valuable programmatic buyers an opportunity to participate. Some methods could be having standalone programmatic units — you could leverage first look through the ad server.

Brock: Sure. I think in this case it might be null and void because you've obviously got a nice direct roadblock there. But under, of course, approved situations where there might just be a standard line item — basically not what you see on the screen right now — you'll allow Google to take a piece of inventory that would have gone to standard line items at a pre-conceived rate. So we might get to a point where we think: okay, we're going to get $50 CPM out of this standard line item, but we'd be happy to give that to Google at $25. What that does is break up your reserved and unreserved programmatic stack. What we want to do is give Google an opportunity to diversify — let them pay when they're going to pay — but not go in and actually ruin the serving outcomes of your direct campaigns. So basically: normally if you have a direct campaign, it gets to serve above everything and that's always going to happen. But with first look, we're saying to GAM: if we get such a high CPM, we'll let it fill for these impressions. It still lets your direct campaigns fill as planned, so your advertisers are happy — but you're also getting a high CPM through if it can compete at that level.

Definition — First look: a setup in GAM that allows a demand source like Google to bid on reserved inventory at a pre-set price floor, giving high-value programmatic demand access to premium placements without disrupting direct campaign delivery.

Ben: That's right. And the difference is essentially the open auction value and the price that you put in at the rule level. Easy way to squeeze a few extra dollars at scale — we've seen that work super well for direct-heavy publishers. That can be a really quick win.

Brock: The other thing I'm not seeing is refresh. You often don't want to refresh your direct ads because advertisers don't particularly want that to happen. But a lot of the time we see premium enough publishers just get a huge amount more inventory — both for direct and programmatic — through strategic uses of refresh. Maybe not refreshing sponsorship, but having an option to refresh direct. You just get 2x the inventory to sell, so it's a really quick way of increasing what your direct team can sell but also what you can sell programmatically as well.

Ben: Often that's a pretty big issue you get with a premium brand — they just run out of things to sell. And it always comes back down to what your sell-through rate is. There are ways around that: if you know who's going to be buying an impression, particularly if it's going direct, a simple key-value plus refresh solution will maintain the likelihood of the direct campaign serving and not being refreshed away for something that's less valuable.

Brock: Yeah. Last thing is maybe just A/B testing the number of ads we have within the listings, their sequence, those kinds of things. We've been able to get some gains in A/B testing these things — potentially looking at the lazy loading that we actually have on page and trying to make that a bit more scroll-speed-based. Often we find you'll get one user on this type of site who is flicking through the cars very quickly, and another who is analysing every thumbnail to the nth degree. We want to be able to get both. So that we've got high viewability rates — adjusting our lazy loading depth: normally we just pick a pixel point above the unit to begin loading it. If we massively increase that for someone scrolling quickly and shrink it for someone scrolling slowly, we're more likely to get high viewability for both.

Ben: Yeah. The final idea is probably not going to be liked by the guys at Cars.com, but — AdSense for Search.

Brock: It's not well liked amongst non-media product teams, but it does pump. It's very clickable — it's basically linked ads, very much tied to the context — and Google will pay plenty for those clicks. You'd stick it right by the pagination at the bottom of the page. It'd be a perfect A/B test candidate — particularly if you've got the purple link colour, maybe an off-white background — you might be able to pick up a few people.

Definition — AdSense for Search (AFS): a Google product that places contextually relevant search ads on a publisher's site, typically at the bottom of a page near pagination or search results, earning revenue on a cost-per-click basis.

Ben: Of course, the thing you really have to measure is how much that's going to cost you by sending a user somewhere else. So what we've found with any of these types of sites is there's always hesitation to do any of these optimisations because anything that's going to impact people moving from search to details page is a negative. So we want to be able to A/B test anything we're doing — show that there's an uplift, but also show there's no impact to the wider business. Check: are we making more revenue, or are we impacting what we hypothesise would change? And are we impacting other areas — is there a change in pages per visit, or in the number of people moving into the next part of the funnel? If we can show there's no real change there but we've got our AdSense for Search revenue — then it's a win.

Brock: This is the most sensitive area of the website for ads. This is where buyer intent is super high and we really want to drive them to that lead.

Ben: What are people paying for these leads, do you know? How much is the pricing here?

Brock: I have no idea, but I bet it's more than the ads.

Ben: Yeah, it's more than the ads. I think it can be upwards of $200 — we might need to check that. But have a look at the price of this car — how much are they willing to pay? And what's their rate of conversion?

Brock: Pretty high, definitely.

What does the desktop details page look like, and where is the biggest ad opportunity?

Ben: Then in terms of ads that we actually have on the details page — we've got our top unit here, which is serving a 728x90. It's got a few other sizes in there — a 970x90 slightly longer version — and then we've got a mobile size there as well. As we scroll down, we've got the gallery. This is often a pretty hot area of the page for users, and we've got a really nice position there. My guess is view time on that is pretty high — viewability is going to be close to 100% as well. What's interesting is on the previous page, before you put the gallery into motion, you can still cycle through the images quite easily, and in this case they've very cleverly decided to lazy load that ad — it isn't loaded until you've actually opened up the page.

Brock: That's right. What I've seen across any of the verticals — particularly real estate — they want you to look at the images. And that might be a bit overblown in the real estate vertical because you're going to buy something you have to look at every day, right? You want to actually have a nice full screen. They've got massive refresh opportunity — the question would be how many people are actually going to the semi-full-screen mode, and what's their refresh opportunity on that?

Ben: Definitely. And obviously they're not refreshing — but we often think that would be a pretty big opportunity. So the idea would be: if we have X percent of people just using it as a slideshow, we could actually have some of these images be ads essentially as well, couldn't we?

Brock: That's right. And that could be standard programmatic, or of course they can wire that up to be part of their roadblocks — so one in every ten within the gallery, they surface an ad. We've seen really strong viewability and CPMs on those kinds of units. You're catching people in flow state — they want to see all the images and they'll click consistently, bang bang bang, and there you go.

Ben: Next up we've got a hardcoded direct unit, a few nice features for the user, and then we've got two units at the bottom — both lazy loaded. I don't know how much sizing we have on these; I think they're fairly fixed. This one can do a 300x250, and then our bottom one has the same sizes as the top one. There are a few opportunities with those too. I mean they're not going to be your money makers.

Brock: Yeah, but that being said, you can get away with a lot more. Having a latent 970x250 on this page is always going to be more valuable than not. There's no real harm in this unit being a billboard — being significantly bigger real estate. It's already at the bottom, it's already basically by the T&Cs. So that's a pretty quick win — we always suggest for our team and our customers, if you can get a 970x250 in even if it's in a lower position, then that's worth doing. The difference in performance is kind of night and day.

Ben: Cool. So that's the page. What would we look to enhance? The dream would be to be able to get some kind of placement in here where we've got user attention — users are going to go from the gallery down to the details to make their decision. This is where all of the engagement is. But at the same time, this is also where they're making their purchase decision.

Brock: Yeah. This is really challenging. This contact seller form is the gold nugget — this is where they want all their users to go. They want people to put their names in, collect that data, charge for that lead. Trying to get an ad around there is always going to be tricky internally. But that's where the performance is going to be — may test it, see it, we'd always suggest that.

Ben: Of course. The other thing is they've got this really nice navigation bar — very easy for you to flick to the next in the series. The real question will be: how many people are using that? Do they have telemetry on that usage? And if so, could they make it more prominent and move the ad down a little bit? It's nice how it pulls you back up to the top of the page — I don't know if this is on purpose, but you get a view of that ad every single time. I'm here, I don't particularly need to be pulled back up to the top again, but I do — and I get to see that ad again. It's nice for our ad viewability. I wonder from a user point of view — particularly if you're maybe not looking to really buy and you're just looking at cars for a hobby or you're at the start of the funnel — this might be something that they transfer to a SPA or single-page application type model where you can flick without having the full reload.

Brock: I mean, that's an investment — that's actually quite a massive investment, and I would never suggest that without a real need and use case. But for comfortable usage it might be a benefit.

Ben: Yeah, definitely. Quick wins: have that unit stick for 100 pixels like we said for the last one. Apart from that, it's going to be hard to pull any ad units up — potentially this unit could be closer to where the content is going to be, but this is going to be a difficult place to actually get to optimise. Blowing that out a bit, making it a bit bigger, if we could A/B test getting a small unit within content here — and then obviously our gallery optimisation. I think that's where the magic happens.

Brock: Yeah. That's where the rubber meets the road.

What does the mobile layout look like on the search page, and what's missing?

Ben: All right, we will jump into the mobile side of things next. First things first — we've got our top unit here, it's a standard 320x50. Pretty normal mobile banner. As we flick through the site, then we've got our native direct products running through GAM as well — look really nice, very clickable. As we continue, it's pretty light on ads and we've got these really small 320x50s coming through within the listings. They've kept it very simple — there are two, they're lazy loaded.

Brock: Yeah, there's another one there and then we're at the bottom unit. We've got a 300x250 there at the pagination section.

Ben: It feels super light to me.

Brock: I think you're right. It's counter to what I would have thought. I would have definitely thought that a desktop user is just more likely to transact. There's a human hang-up on buying something worth $120,000 through your phone, perhaps. So I would have thought mobile gives you a bit more flexibility in that sense.

Ben: On the flip side, it could be that that's their growing audience. So you're saying mobile users are less likely to convert — so maybe we've got more option to monetise them in a different way.

Brock: That's what I would have guessed. Particularly for a car site. I think that that's less true for a fast-moving consumer good or something like that, but for a car — particularly a house — I would have thought you'd be on your computer.

Ben: Yeah, but I guess in this case you're probably just filling in a lead form. And that's why we put that disclaimer at the start. The other piece is why they wouldn't use 300x250s in content — which is interesting, because the actual site is built around almost those card-image proportions. I would have thought it actually blends in quite well and I'd recommend pumping them out to a 300x250 pretty quickly. I think it would have a massive impact on performance.

Brock: There's also the notable omission of the mobile sticky footer. It's a volume gainer — it's all about just getting impressions. It gives them a little bit more to sell, particularly for their direct strategy. There's refresh opportunity when you do that, of course. Highly viewable, decent CPM — it's more or less a standard across the internet these days. And it drives up overall viewability across the site as well, so you're really helping everything perform a little bit better.

Definition — Mobile sticky footer: a 320x50 ad unit fixed to the bottom of the screen on mobile, remaining in view as the user scrolls. Widely used to boost impression volume and overall page viewability.

Ben: Yeah. I mean, how valuable that is is arguable, but when you're trying to tell a brand a story — particularly these car brands — you want to be telling them a story with data. Viewability is one of those key data points that you want to make sure keeps going up.

Brock: What else have we seen perform well on this type of site? The real question comes down to what the user is doing. If there's healthy scroll, I think you could probably get away with even direct or programmatic — some sort of interscroller unit. What I've seen work quite well in the direct sense is full-screen video. So when I say full screen — remember we're on a mobile device, scrolling down between two car listing pages. You'll have a nice immersive video, and these OEMs are going to really splash the cash at that kind of position.

Definition — Interscroller: a high-impact mobile ad format that reveals as the user scrolls, typically occupying the full viewport briefly before the user continues scrolling past.

Ben: So I scroll down within the listings and basically get a full-page video from a car brand in my face. Industrial kind of job.

Brock: 100%, yeah. It's going to be intense, there's no doubt about it. But you're charging for that. And on the flip side, you trade off accordingly — you'd serve this video unit and you wouldn't have other in-content units, but you would have a sticky footer as well. And we can use that position programmatically as well — you can do programmatic interscrollers. You'll be controlling the direct creative mainly from the ad server, and perhaps you might need to load in some JavaScript, but these days most of it just runs out of a standard requested size anyway.

Ben: Anything else you're testing or adding? Again, it's about getting that 300x250 demand. If we're wanting to really boost what's left over — remnant, or what comes to us in the open auction — we want to be getting the best sizes in there, and for mobile that's 300x250. Once we've got that space we can keep the 320x50 running anyway, keep that competition in there. Key thing we'd probably do there is try and fix that container and just get smaller units scrolling within that position so we get a little bit more viewability — our product is called the mini sticky. You're getting about an 8% viewability bump from that. Looks a little bit nicer.

Brock: Anything else? Yeah, I think those are the main three that I'd push. Might A/B test the ad sequence as well — that kind of stuff.

Ben: Of course. And that's a whole other area really. What the real question is: what is the media team trying to achieve with the site? If it's volume, there are ways to get around that with new units, and that's going to spill over into the direct stuff. If it's just CPM viability, then that's the path you probably take.

What does the mobile details page look like, and what would you change?

Ben: Okay, so any last optimisations that you can see here, Brock?

Brock: Yeah, when I was playing around with the site before, I noticed that even on the search page you can actually flick through images. So it's interesting — I'm hoping that they have some sort of measurement on that, maybe some sort of event telemetry system that allows them to see if that's being used. If it is, you could sneak an MREC in there perhaps.

Definition — MREC: a 300x250 medium rectangle ad unit, one of the most common and high-performing display ad sizes.

Ben: Yeah, definitely. It looks like you only get a few, but when you click through it brings you to the details page. So that's probably a nice transition — let's have a look through this section of the site. Similar enough to what we were looking at on desktop. We've got a top unit — standard mobile banner size — and then as we scroll down we've got a 300x250, and I think there's one more at the bottom. We also got a 320x50 filling in there. Let's just check the sizes — we also got a native right next to us, yes. I don't think that is actually serving through GAM as well — I think that might be hardcoded. So we do have multiple sizes coming through here.

Brock: The first thing that jumps out to me is you've got two ad units down the bottom that are next to each other. If I remember correctly on desktop, they had: big image, form, written content, form. And now on mobile they've got obviously the gallery or big image, initial form at the top, or you click the button — and they've kind of mixed up the order. That kind of makes me think maybe what we said before about mobile being less important from a conversion point of view. But what it probably is, is the template they're using for maintaining their website from a layout point of view is just sort of shifting all the stuff that was on the right-hand side and pushing it down — just wherever it goes.

Ben: Yeah, you know, maybe there's some optimisation they can do there. What are you changing or adding here? Let's have a look at the gallery to begin with.

Brock: I mean, this is the centrepiece of the page. You're able to scroll horizontally, plus we can click in — so we've got two options, same as desktop. As we're scrolling, we have no ads, but within here I actually really like how you flick down pretty quickly and we do have ads lazy loaded within here. It's loading, or else I've maybe hit a floor here. Assuming you haven't hit the floor and your internet is reliable, then maybe there's some lazy loading buffer optimisation they could do, or even just have the first unit render immediately.

Ben: But this is a long page — we've got quite a bit of scroll depth here. I think a mobile sticky probably wouldn't hurt too much in terms of the actual conversion rate. We've got the user not quite beside the contact form anymore — we've got that viewability and view time. I guess, had we been able to expand the image to full screen — I'm not getting that here — or at least maybe a View button. Either way, as a user, how am I getting the full immersive experience? I'm not really sure. So I guess we're saying: one, we could probably harness this section a little bit better — having one of these be an ad unit maybe every five or ten images, test that. Two — we click into the actual gallery, we've got scroll ability. I guess users are probably in here for quite a period of time. So we're taking advantage of that with a mobile sticky unit and potentially some lazy loaded units within there as well.

Brock: What they also could do is embed their lead form as a native style or some sort of creative template that takes those positions — so they can at least see if they can capture people at the end of the image cycle.

Ben: Yeah, that's a really good point. We've then got our next important unit here — a 300x250 — right by the contact seller button, which actually seems like a pretty good position. You'd want it higher up the page, but if you can't get that just because of business rules, that's quite a nice position for it to be in. Anything you would do here to improve or enhance?

Brock: Yeah. I think I would put my product manager hat on and wonder: should my second lead form be in that position? Do I need to move the price history above, then the lead form, then the ad? Or put the ad between the two? Because this is counter to what they're doing on desktop. And that might be because they have a much more thought-out desktop and mobile strategy than I'm perceiving as I'm flicking through the site.

Ben: But it is strange to me that if you scroll up a little bit — a form — go a bit further down — yeah, they've now got their native recirculation-type product, another native, and then a 320x50. So we've got two odds kind of sandwiched at the bottom here as well. This looks to me like it might just be a legacy website management CMS thing.

Brock: Yeah, super easy to fix — which is the best part really.

Ben: The only thing I'm thinking is there could be some data-driven reason we haven't thought through — or they could have decided in actual fact that the 300x250 is performing better than the insurance tab they have, or your payments on the vehicle or getting a loan or financing.

Brock: Yeah. But it's not the core call-out or CTA on page. I think in summary, you've got all of these little widgets — a few of them are ads, some of them are more standard sized, some of them are native — and then of course you've got your contact dealer buttons and your forms. So you could really run a nice multivariate test and see what goes where.

Ben: Correct, yeah. What's actually going to produce more revenue? As you mentioned before, the ad might be doing pretty well — and then where are we getting the actual interaction, which is the most important thing when you're trying to sell cars?

Why is outstream video a big opportunity for Cars.com?

Ben: Yeah, definitely. Anything else to call out? I know it's not coming up on mine, but they had a really nice outstream piece as well. Do you want to explain a little bit about outstream and how that was being used?

Brock: Yeah, I think outstream for this website is absolutely perfect. It goes back to the previous point around full-screen mobile video. This is not necessarily full screen but it is video — and car manufacturers or OEMs spend a metric ton on marketing. They do video really well. They've always had TV budgets. What we're trying to do for this kind of website is take more YouTube budgets. So what's quite easy to buy on YouTube or in other high-frequency video platforms — we want to get the high-spending OEMs, the high-spending car manufacturers, to build beautiful video creative and put that into the key positions on Cars.com. Because they just look so good from an ad point of view. The experience is nice, it's clickable, it's interactive, there's sound — they'll pay a much higher CPM for Cars.com. And the idea with outstream is you can get that video content and video advertising without having video content yourself. You don't need to be a content producer. You don't need to be on YouTube building that kind of content. You can stay as a publisher and true to your roots, but still get that video quality and that high CPM.

Definition — Outstream video: video ads that play within article or page content — not inside a video player — allowing publishers to run video advertising without producing their own video content.

Ben: Pretty crazy upside. And if Cars.com classified all of the cars on the site — they knew the make and model — and they went to the manufacturers and said, "Look, we're going to have a video in the gallery." You'll know exactly what make and model it is. In the first generation it might be exactly what you manufacture — but then again, you can sell competitive spaces to competition, which might be very interesting. So if I'm flicking through, say, a Mercedes-AMG, and at the end I get a Mercedes-AMG video, that might be quite an immersive experience that really uses the video placement to drive the lead.

Brock: Right. So we are happy in this kind of organisation to use media as a proxy for pushing people towards actually spending money.

Ben: Definitely.

Summary: What are the biggest wins for Cars.com's ad strategy?

Ben: I think that's probably a nice place to wrap up. To run through what we've talked through here — the whole strategy is about facilitating what already exists. We're trying to make sure that we're not impacting the core business, but maybe enhancing it. We're driving revenue through the traffic we have and trying to drive that a bit better. In terms of what we saw on the desktop side, there was a bit of opportunity on the search section around maybe having better formats, enhancing the lazy loading to improve and drive viewability. We moved into the details section where we've got stricter business rules to make sure we're still pushing those leads to our dealers — but we still saw some opportunities, particularly around the gallery, to use that a bit better when we know users are going to be really engaged there. There was a quick win at the bottom of the page just making that a bigger unit. And then across mobile, the big wins we looked through were really video — maybe getting some stronger formats in there — and maybe taking a bit more advantage of the space. We guessed on the fact that we're getting maybe less buying intent on mobile — we don't know that for sure — but if that was the case, maybe we can take a little bit more advantage of the space, get some 300x250s in as well. Bits to do — I think we could get a bit of uplift there from the site. But overall, they've fulfilled on the strategy you'd expect from this kind of publisher.

Ben: I think that'll end and recap episode 3 of Layout Teardown — with myself Ben and Brock. And if you learned something, enjoyed it, or have some feedback for us, leave a comment. We'll see you here next time.

This is an edited transcript of Layout Teardown, Episode 3. The words are Ben and Brock's own — lightly edited for readability (filler words, false starts, typos, punctuation). No claims have been rewritten or generated by AI.

‍

Publift Scoop

Publift Scoop

Sign up to the Publift newsletter for all the latest news, updates and insights from across the industry.
Sign up
Play video
Play icon
Play video

Layout Teardown: Episode 3 — Cars.com's Sophisticated Ad Strategy

Layout Teardown: Episode 3 — Cars.com's Sophisticated Ad Strategy

Contents
Contents

Layout Teardown: Episode 3 — Cars.com's Sophisticated Ad Strategy

Layout Teardown is a video series where Ben and Brock breakdown the ad strategy and website layout of popular publishers. In this week's episode, we're taking a look at Cars.com, one of the biggest auto marketplaces in the US. Ben and Brock speak about the detailed insights and suggestions for optimising cars.com’s ad revenue. From tweaking the desktop layout to maximising mobile monetisation with various ad units, stay tuned to learn valuable tips to develop your ad strategy and improve overall performance.

Key Takeaways

  • Ads come second to leads, and that is the right call. Cars.com makes its real money when a buyer fills out a form and gets passed to a dealership. A single lead is worth far more than an ad click, so the ad strategy stays conservative.
  • The hero banner is being wasted. The big desktop billboard is a premium slot that car brands would pay well for, but eager buyers scroll straight past it before it has finished loading. Viewability sits around 40% to 50% as a result. Keeping it pinned for just the first 100 pixels of scroll would fix that almost immediately.
  • The image gallery is a huge missed opportunity. Buyers spend a lot of time flipping through vehicle photos, which makes it one of the most attentive moments on the entire site. Currently, that space is ad-free. Slipping an ad into the gallery rotation every ten slides would deliver near-perfect viewability without meaningfully disrupting the experience.
  • Car brands have video budgets burning a hole in their pocket. Manufacturers spend enormous amounts on video advertising, and Cars.com is sitting on the perfect placement to capture some of that spend. An outstream video unit inside the photo gallery, matched to the exact make and model the user is already looking at, would be a compelling pitch to any automotive brand.
  • Mobile is being underleveraged. There is no sticky footer, no 300x250 within the listings, and no real effort to monetise the casual browser who is just doing research on their phone. That audience may never fill out a lead form, but they are still valuable and currently generating almost nothing.
  • The map is creating a lot of noise. Every time a user drags or zooms, the page reloads the listings and fires off a fresh round of ad requests. Most of those go nowhere. People interact with maps constantly, so this adds up quickly and needs to be cleaned up before it distorts performance data.

What makes Cars.com interesting from an advertising perspective?

Ben: Hey and welcome to episode 3 of Layout Teardown, where we bring you through the layouts and programmatic strategy of some of the biggest publishers globally. Today we're going to walk through Cars.com, which is one of the biggest auto marketplaces in the US. This is an interesting one because it's a different type of publisher — they have a different business model. They're generating most of their revenue through basically leads: people buying cars online. And really advertising — or "media" as they call it — is normally seen as a supplementary business. Now we know from a lot of our bigger publishers that we work with that it's a growing area within their businesses, but they need to make sure that those ads aren't impacting what's really driving the revenue. So pretty unique from that perspective — also just in terms of the user type, the data that they have, and the auto segment having so much revenue and money in it in terms of advertising. There are just a lot of ways that they can make money. So it's an interesting type of website to look at from that perspective. My name is Ben.

Brock: Hey, I'm Brock, and we will be jumping into this one. We'll kick it off by looking through the site on desktop and mobile, showing you guys what we think is good, where we see some improvements, and generally giving you guys some tips as we go.

What does Cars.com's traffic look like, and what does the revenue mix between leads and media look like?

Ben: Okay, so we are now going to jump in and have a look at the data. Obviously we haven't seen any of the real data, but we're using a few Chrome extensions and tools to give us an insight on what the traffic looks like. We're pulling up SimilarWeb within here, so we can see they've got a really significant amount of monthly page views coming through on site and a really strong visit duration — this is kind of what you would expect on a marketplace like this. You've got people browsing and looking through sites a lot. In terms of geography, Brock — we've got almost all traffic really coming from the US.

Brock: Yeah, it makes sense to me. I would imagine these guys, who were born and bred in the United States, have got a US-centric strategy. I think that will spill into their media strategy as well. When you think about why Cars.com existed in the beginning, they would have been all about just getting the user to actually buy that car. So they have — I'm guessing — a very substantial lead product that goes out to all the dealers around the States. And now what you're seeing is a blended media strategy that's very much focused on the US. Plus I guess for users, they want to go to the place that has the most cars, the most choice. So it makes sense that you'd get that wheel of growth — the more cars you have, the more users you get, that kind of stuff. That's easier to do in one country.

Ben: The conversions as well I would add to that — if they've got a nice wheel that's churning out users into these dealerships, or even an online form, I'm sure a lead cost form on the dealer side would be very, very lucrative.

Brock: Definitely. And then just on traffic sources, it's pretty split between search and direct. So that's interesting — they've obviously got a pretty good search strategy in place, and I guess there are people coming to the site anyway, but it seems pretty heavily on the search side from what I would have anticipated.

Ben: Yeah. "Cars.com" — that's the complete URL — I feel like that's going to be feeding into their search strategy.

Brock: Definitely. And I'm sure with that kind of brand recognition they're going to have a lot of people coming back, and that builds up your direct channel over time. So in terms of the makeup — what does the revenue breakdown look like between their actual lead product and then within media, what's that like direct versus programmatic?

Ben: Yeah, I would imagine it follows a very standard marketplace model, even outside of the automotive space. They would have been born directly looking to get that lead revenue and would have probably been 100% automotive revenue driven — as in sales. And then over time some media starts to creep its way in: put an ad somewhere, put an ad next to the gallery or whatever they do. And then very, very quickly the media part becomes like a massive piece of their P&L. I would think now it's probably getting closer to 50/50, although that's fairly much guesswork.

Brock: Yeah. But when a company like that has to go through such a transformation, it becomes really important to have a nice blended strategy and a blended team — and that's not that easy to actually execute. Really any website where you've got such close buyer intent, where people are literally searching and in the idea of making a purchase — advertisers want to be across this. And unlike, say, property sites, which are similar, you don't have ads for one house, whereas with cars you've got brands and you've got specific vehicles. So there's a lot of revenue and money put behind these kinds of ads.

Ben: So there's a lot of revenue available. As we know, the product guys don't particularly like the ads because they think it's pulling away from the end objective. But we've seen some really slick strategies that allow you to keep ads relatively simple but use all that really rich buyer intent data to drive revenue — not just open auction, but selling it in a variety of ways.

Brock: Yeah, I would be very surprised if Cars.com haven't really spent millions of dollars developing out their audience strategy, and that is then fuelling their media strategy. If they're able to tap into that buyer intent and collect it in a privacy-safe way and make it available to buyers, I have no doubt they'd be able to turn a tidy profit out of that.

Ben: Yeah. If I could tell Volvo that this person was interested in a Volkswagen — that's some super valuable information, and I'm sure you could charge a pretty penny for it as well.

What are the two most important page types for ad revenue on Cars.com?

Ben: Okay, so we will now jump into looking through the layout specifically. Really on any of these types of sites, there are two core areas that drive 70 to 80% of the revenue. We've got our listing section — or search section — where we've searched for a type of vehicle that we want. And then we've got the details page: once we've gone through the listings page and we're looking at more information. These types of sites will have a lot of other content, but really that's just about bolstering traffic, SEO, experience — those kinds of things. The bread and butter, the revenue, is coming from these two page types. And really from the search listings from a media perspective — the reason being, once you get to a details page, you're at an even higher phase of intent, and so the model of getting leads is even more important at that point. So it's normally a bit more protected.

Brock: Yeah. I think conceptually it's probably basic enough marketing funnel stuff. You've gone from search/purchase consideration into buying intent into action behaviour, and hopefully for these guys — purchase.

Ben: Exactly.

What does the desktop search/listings page look like, and what are the optimisation opportunities?

Ben: So we'll start off on the search page. First thing we can see is we do have a header bidding auction running — they've got a bit of demand in there. From your analysis, they've actually done their own Prebid build here, by the looks of it.

Brock: That's what it looks like. I also noticed that you must have a VPN on because obviously we're in Australia and I didn't see any Prebid. So it looks like they have a localised strategy — that might just be because of some of the requirements with some of their OEM deals that they have to manage, and of course privacy around the world is not as easy and simple as it can be.

Definition — OEM: Original Equipment Manufacturer; in this context, major car brands like Ford, Toyota, or BMW that advertise on automotive platforms.

Ben: That's true. We've seen a lot of publishers just cut off bothering with European ads just because of the effort involved. Plus we've got 93% coming from the US and Canada, so maybe they've just decided the risk isn't worth the return.

Brock: Yeah. The other thing is they could really use that almost 10% of non-US traffic as just a test bed. Looking at SimilarWeb, they're getting some Australian traffic — they're going to get decent CPMs, they'll have enough volume to actually try something. That might be something they look at in the future.

Ben: Yeah. First thing we see obviously is this huge billboard — a 970x250 — really nice position. I would guess this is sold out most of the time direct. That's kind of a pretty premium spot.

Brock: Yeah, I think so. I think this is your centrepiece. I would imagine they have a homepage solution as well that they might be shipping out for direct. But you know, search is where the action happens. It's above the fold, so it inherently has a bit of extra value. It's a 970x250, and if you've got OEMs looking to put in a beautiful, nice rich media ad, you can charge a pretty penny for that for sure.

Ben: Any optimisation opportunities? The major thing will be speed in this case. If you want to refresh that one there — they've done a pretty good job at optimising the page, and as you can see, you're also loading a well-designed and probably heavy creative. You might have a bit of user banner blindness from that one — off they go, they start scrolling, they look at cars. There was a bit of a delay there. I guess the one tweak we could do is try to make this unit a bit more valuable and a bit more viewable. From similar units that we've seen, particularly at that kind of loading speed, you're at kind of a 40 to 50% viewability rate. So it looks like it should be a lot higher than that.

Brock: What we've seen other car marketplaces do is just have a small sticky function — so that unit sticks for 100 pixels — so we're just a little bit more likely to get that user's eyeballs on it. It rolls away straight away and we're back to the same content. I think that would probably be a small piece we could do there, but it could have a big impact. Even on the direct side, if they're serving roadblocks across a search result or key category, they could have that headline billboard and then like a 728x90 sticky that runs for the first say five or six images or links — that might be a way to keep that nice premium-looking top-of-page experience and then of course get a little bit of extra cash from a nice viewable unit.

Ben: Next we've got our within-listing units. It looks like a lot of these are being lazy loaded as well, so we've only got certain ones actually loading as we scroll. Lazy loading actually looks pretty good — it seems to be catching them. We've got this super nice native unit — presumably direct?

Brock: It looks like it, yeah. You can count how many actual listing positions there are — so in four. They've got a nice car image, all the text matches, same colour for the call to action, and it's served through GAM as well. Keeps it nice and centralised, easy for the media teams to manage — looks pretty slick.

Definition — GAM (Google Ad Manager): Google's ad server used by publishers to manage, serve, and report on ad inventory.

Ben: And then we've got our sidebar sticky on the left-hand side — a 300x600. Really nice. We've got this unit that looks like a shoppable format, gallery-type unit — maybe custom?

Brock: I'd be direct only on that, and I'd get a very nice CPM from it. It's good that they are allowing something interactive in there as well — not only does it break up the content, it actually gives the brand the ability to be quite creative and try and drive some sort of return on investment on what they're actually buying from Cars.com.

Ben: And clearly they've bought out the whole section here. Any optimisations you can see at the moment?

Brock: I like this placement, by the way — just at the kind of page transition section. We often see this is like the second most valuable location on these types of pages, just because users very often are changing page. Normally the further down an ad is, the less likely it is to be viewed — but for this unit it's actually got pretty high viewability and also high engagement. You've got users clicking and making some kind of action here, and they're more inclined to click on or engage with that unit at that point. Anything else you would change, do, or add differently?

Ben: Yeah, it's very hard to say when you've got a nice roadblock there. But the real question would be: how often are buyers going through the direct channel and then taking everything? So we want to be looking at what our sell-through rates are from a direct point of view and making sure that we give at least very valuable programmatic buyers an opportunity to participate. Some methods could be having standalone programmatic units — you could leverage first look through the ad server.

Brock: Sure. I think in this case it might be null and void because you've obviously got a nice direct roadblock there. But under, of course, approved situations where there might just be a standard line item — basically not what you see on the screen right now — you'll allow Google to take a piece of inventory that would have gone to standard line items at a pre-conceived rate. So we might get to a point where we think: okay, we're going to get $50 CPM out of this standard line item, but we'd be happy to give that to Google at $25. What that does is break up your reserved and unreserved programmatic stack. What we want to do is give Google an opportunity to diversify — let them pay when they're going to pay — but not go in and actually ruin the serving outcomes of your direct campaigns. So basically: normally if you have a direct campaign, it gets to serve above everything and that's always going to happen. But with first look, we're saying to GAM: if we get such a high CPM, we'll let it fill for these impressions. It still lets your direct campaigns fill as planned, so your advertisers are happy — but you're also getting a high CPM through if it can compete at that level.

Definition — First look: a setup in GAM that allows a demand source like Google to bid on reserved inventory at a pre-set price floor, giving high-value programmatic demand access to premium placements without disrupting direct campaign delivery.

Ben: That's right. And the difference is essentially the open auction value and the price that you put in at the rule level. Easy way to squeeze a few extra dollars at scale — we've seen that work super well for direct-heavy publishers. That can be a really quick win.

Brock: The other thing I'm not seeing is refresh. You often don't want to refresh your direct ads because advertisers don't particularly want that to happen. But a lot of the time we see premium enough publishers just get a huge amount more inventory — both for direct and programmatic — through strategic uses of refresh. Maybe not refreshing sponsorship, but having an option to refresh direct. You just get 2x the inventory to sell, so it's a really quick way of increasing what your direct team can sell but also what you can sell programmatically as well.

Ben: Often that's a pretty big issue you get with a premium brand — they just run out of things to sell. And it always comes back down to what your sell-through rate is. There are ways around that: if you know who's going to be buying an impression, particularly if it's going direct, a simple key-value plus refresh solution will maintain the likelihood of the direct campaign serving and not being refreshed away for something that's less valuable.

Brock: Yeah. Last thing is maybe just A/B testing the number of ads we have within the listings, their sequence, those kinds of things. We've been able to get some gains in A/B testing these things — potentially looking at the lazy loading that we actually have on page and trying to make that a bit more scroll-speed-based. Often we find you'll get one user on this type of site who is flicking through the cars very quickly, and another who is analysing every thumbnail to the nth degree. We want to be able to get both. So that we've got high viewability rates — adjusting our lazy loading depth: normally we just pick a pixel point above the unit to begin loading it. If we massively increase that for someone scrolling quickly and shrink it for someone scrolling slowly, we're more likely to get high viewability for both.

Ben: Yeah. The final idea is probably not going to be liked by the guys at Cars.com, but — AdSense for Search.

Brock: It's not well liked amongst non-media product teams, but it does pump. It's very clickable — it's basically linked ads, very much tied to the context — and Google will pay plenty for those clicks. You'd stick it right by the pagination at the bottom of the page. It'd be a perfect A/B test candidate — particularly if you've got the purple link colour, maybe an off-white background — you might be able to pick up a few people.

Definition — AdSense for Search (AFS): a Google product that places contextually relevant search ads on a publisher's site, typically at the bottom of a page near pagination or search results, earning revenue on a cost-per-click basis.

Ben: Of course, the thing you really have to measure is how much that's going to cost you by sending a user somewhere else. So what we've found with any of these types of sites is there's always hesitation to do any of these optimisations because anything that's going to impact people moving from search to details page is a negative. So we want to be able to A/B test anything we're doing — show that there's an uplift, but also show there's no impact to the wider business. Check: are we making more revenue, or are we impacting what we hypothesise would change? And are we impacting other areas — is there a change in pages per visit, or in the number of people moving into the next part of the funnel? If we can show there's no real change there but we've got our AdSense for Search revenue — then it's a win.

Brock: This is the most sensitive area of the website for ads. This is where buyer intent is super high and we really want to drive them to that lead.

Ben: What are people paying for these leads, do you know? How much is the pricing here?

Brock: I have no idea, but I bet it's more than the ads.

Ben: Yeah, it's more than the ads. I think it can be upwards of $200 — we might need to check that. But have a look at the price of this car — how much are they willing to pay? And what's their rate of conversion?

Brock: Pretty high, definitely.

What does the desktop details page look like, and where is the biggest ad opportunity?

Ben: Then in terms of ads that we actually have on the details page — we've got our top unit here, which is serving a 728x90. It's got a few other sizes in there — a 970x90 slightly longer version — and then we've got a mobile size there as well. As we scroll down, we've got the gallery. This is often a pretty hot area of the page for users, and we've got a really nice position there. My guess is view time on that is pretty high — viewability is going to be close to 100% as well. What's interesting is on the previous page, before you put the gallery into motion, you can still cycle through the images quite easily, and in this case they've very cleverly decided to lazy load that ad — it isn't loaded until you've actually opened up the page.

Brock: That's right. What I've seen across any of the verticals — particularly real estate — they want you to look at the images. And that might be a bit overblown in the real estate vertical because you're going to buy something you have to look at every day, right? You want to actually have a nice full screen. They've got massive refresh opportunity — the question would be how many people are actually going to the semi-full-screen mode, and what's their refresh opportunity on that?

Ben: Definitely. And obviously they're not refreshing — but we often think that would be a pretty big opportunity. So the idea would be: if we have X percent of people just using it as a slideshow, we could actually have some of these images be ads essentially as well, couldn't we?

Brock: That's right. And that could be standard programmatic, or of course they can wire that up to be part of their roadblocks — so one in every ten within the gallery, they surface an ad. We've seen really strong viewability and CPMs on those kinds of units. You're catching people in flow state — they want to see all the images and they'll click consistently, bang bang bang, and there you go.

Ben: Next up we've got a hardcoded direct unit, a few nice features for the user, and then we've got two units at the bottom — both lazy loaded. I don't know how much sizing we have on these; I think they're fairly fixed. This one can do a 300x250, and then our bottom one has the same sizes as the top one. There are a few opportunities with those too. I mean they're not going to be your money makers.

Brock: Yeah, but that being said, you can get away with a lot more. Having a latent 970x250 on this page is always going to be more valuable than not. There's no real harm in this unit being a billboard — being significantly bigger real estate. It's already at the bottom, it's already basically by the T&Cs. So that's a pretty quick win — we always suggest for our team and our customers, if you can get a 970x250 in even if it's in a lower position, then that's worth doing. The difference in performance is kind of night and day.

Ben: Cool. So that's the page. What would we look to enhance? The dream would be to be able to get some kind of placement in here where we've got user attention — users are going to go from the gallery down to the details to make their decision. This is where all of the engagement is. But at the same time, this is also where they're making their purchase decision.

Brock: Yeah. This is really challenging. This contact seller form is the gold nugget — this is where they want all their users to go. They want people to put their names in, collect that data, charge for that lead. Trying to get an ad around there is always going to be tricky internally. But that's where the performance is going to be — may test it, see it, we'd always suggest that.

Ben: Of course. The other thing is they've got this really nice navigation bar — very easy for you to flick to the next in the series. The real question will be: how many people are using that? Do they have telemetry on that usage? And if so, could they make it more prominent and move the ad down a little bit? It's nice how it pulls you back up to the top of the page — I don't know if this is on purpose, but you get a view of that ad every single time. I'm here, I don't particularly need to be pulled back up to the top again, but I do — and I get to see that ad again. It's nice for our ad viewability. I wonder from a user point of view — particularly if you're maybe not looking to really buy and you're just looking at cars for a hobby or you're at the start of the funnel — this might be something that they transfer to a SPA or single-page application type model where you can flick without having the full reload.

Brock: I mean, that's an investment — that's actually quite a massive investment, and I would never suggest that without a real need and use case. But for comfortable usage it might be a benefit.

Ben: Yeah, definitely. Quick wins: have that unit stick for 100 pixels like we said for the last one. Apart from that, it's going to be hard to pull any ad units up — potentially this unit could be closer to where the content is going to be, but this is going to be a difficult place to actually get to optimise. Blowing that out a bit, making it a bit bigger, if we could A/B test getting a small unit within content here — and then obviously our gallery optimisation. I think that's where the magic happens.

Brock: Yeah. That's where the rubber meets the road.

What does the mobile layout look like on the search page, and what's missing?

Ben: All right, we will jump into the mobile side of things next. First things first — we've got our top unit here, it's a standard 320x50. Pretty normal mobile banner. As we flick through the site, then we've got our native direct products running through GAM as well — look really nice, very clickable. As we continue, it's pretty light on ads and we've got these really small 320x50s coming through within the listings. They've kept it very simple — there are two, they're lazy loaded.

Brock: Yeah, there's another one there and then we're at the bottom unit. We've got a 300x250 there at the pagination section.

Ben: It feels super light to me.

Brock: I think you're right. It's counter to what I would have thought. I would have definitely thought that a desktop user is just more likely to transact. There's a human hang-up on buying something worth $120,000 through your phone, perhaps. So I would have thought mobile gives you a bit more flexibility in that sense.

Ben: On the flip side, it could be that that's their growing audience. So you're saying mobile users are less likely to convert — so maybe we've got more option to monetise them in a different way.

Brock: That's what I would have guessed. Particularly for a car site. I think that that's less true for a fast-moving consumer good or something like that, but for a car — particularly a house — I would have thought you'd be on your computer.

Ben: Yeah, but I guess in this case you're probably just filling in a lead form. And that's why we put that disclaimer at the start. The other piece is why they wouldn't use 300x250s in content — which is interesting, because the actual site is built around almost those card-image proportions. I would have thought it actually blends in quite well and I'd recommend pumping them out to a 300x250 pretty quickly. I think it would have a massive impact on performance.

Brock: There's also the notable omission of the mobile sticky footer. It's a volume gainer — it's all about just getting impressions. It gives them a little bit more to sell, particularly for their direct strategy. There's refresh opportunity when you do that, of course. Highly viewable, decent CPM — it's more or less a standard across the internet these days. And it drives up overall viewability across the site as well, so you're really helping everything perform a little bit better.

Definition — Mobile sticky footer: a 320x50 ad unit fixed to the bottom of the screen on mobile, remaining in view as the user scrolls. Widely used to boost impression volume and overall page viewability.

Ben: Yeah. I mean, how valuable that is is arguable, but when you're trying to tell a brand a story — particularly these car brands — you want to be telling them a story with data. Viewability is one of those key data points that you want to make sure keeps going up.

Brock: What else have we seen perform well on this type of site? The real question comes down to what the user is doing. If there's healthy scroll, I think you could probably get away with even direct or programmatic — some sort of interscroller unit. What I've seen work quite well in the direct sense is full-screen video. So when I say full screen — remember we're on a mobile device, scrolling down between two car listing pages. You'll have a nice immersive video, and these OEMs are going to really splash the cash at that kind of position.

Definition — Interscroller: a high-impact mobile ad format that reveals as the user scrolls, typically occupying the full viewport briefly before the user continues scrolling past.

Ben: So I scroll down within the listings and basically get a full-page video from a car brand in my face. Industrial kind of job.

Brock: 100%, yeah. It's going to be intense, there's no doubt about it. But you're charging for that. And on the flip side, you trade off accordingly — you'd serve this video unit and you wouldn't have other in-content units, but you would have a sticky footer as well. And we can use that position programmatically as well — you can do programmatic interscrollers. You'll be controlling the direct creative mainly from the ad server, and perhaps you might need to load in some JavaScript, but these days most of it just runs out of a standard requested size anyway.

Ben: Anything else you're testing or adding? Again, it's about getting that 300x250 demand. If we're wanting to really boost what's left over — remnant, or what comes to us in the open auction — we want to be getting the best sizes in there, and for mobile that's 300x250. Once we've got that space we can keep the 320x50 running anyway, keep that competition in there. Key thing we'd probably do there is try and fix that container and just get smaller units scrolling within that position so we get a little bit more viewability — our product is called the mini sticky. You're getting about an 8% viewability bump from that. Looks a little bit nicer.

Brock: Anything else? Yeah, I think those are the main three that I'd push. Might A/B test the ad sequence as well — that kind of stuff.

Ben: Of course. And that's a whole other area really. What the real question is: what is the media team trying to achieve with the site? If it's volume, there are ways to get around that with new units, and that's going to spill over into the direct stuff. If it's just CPM viability, then that's the path you probably take.

What does the mobile details page look like, and what would you change?

Ben: Okay, so any last optimisations that you can see here, Brock?

Brock: Yeah, when I was playing around with the site before, I noticed that even on the search page you can actually flick through images. So it's interesting — I'm hoping that they have some sort of measurement on that, maybe some sort of event telemetry system that allows them to see if that's being used. If it is, you could sneak an MREC in there perhaps.

Definition — MREC: a 300x250 medium rectangle ad unit, one of the most common and high-performing display ad sizes.

Ben: Yeah, definitely. It looks like you only get a few, but when you click through it brings you to the details page. So that's probably a nice transition — let's have a look through this section of the site. Similar enough to what we were looking at on desktop. We've got a top unit — standard mobile banner size — and then as we scroll down we've got a 300x250, and I think there's one more at the bottom. We also got a 320x50 filling in there. Let's just check the sizes — we also got a native right next to us, yes. I don't think that is actually serving through GAM as well — I think that might be hardcoded. So we do have multiple sizes coming through here.

Brock: The first thing that jumps out to me is you've got two ad units down the bottom that are next to each other. If I remember correctly on desktop, they had: big image, form, written content, form. And now on mobile they've got obviously the gallery or big image, initial form at the top, or you click the button — and they've kind of mixed up the order. That kind of makes me think maybe what we said before about mobile being less important from a conversion point of view. But what it probably is, is the template they're using for maintaining their website from a layout point of view is just sort of shifting all the stuff that was on the right-hand side and pushing it down — just wherever it goes.

Ben: Yeah, you know, maybe there's some optimisation they can do there. What are you changing or adding here? Let's have a look at the gallery to begin with.

Brock: I mean, this is the centrepiece of the page. You're able to scroll horizontally, plus we can click in — so we've got two options, same as desktop. As we're scrolling, we have no ads, but within here I actually really like how you flick down pretty quickly and we do have ads lazy loaded within here. It's loading, or else I've maybe hit a floor here. Assuming you haven't hit the floor and your internet is reliable, then maybe there's some lazy loading buffer optimisation they could do, or even just have the first unit render immediately.

Ben: But this is a long page — we've got quite a bit of scroll depth here. I think a mobile sticky probably wouldn't hurt too much in terms of the actual conversion rate. We've got the user not quite beside the contact form anymore — we've got that viewability and view time. I guess, had we been able to expand the image to full screen — I'm not getting that here — or at least maybe a View button. Either way, as a user, how am I getting the full immersive experience? I'm not really sure. So I guess we're saying: one, we could probably harness this section a little bit better — having one of these be an ad unit maybe every five or ten images, test that. Two — we click into the actual gallery, we've got scroll ability. I guess users are probably in here for quite a period of time. So we're taking advantage of that with a mobile sticky unit and potentially some lazy loaded units within there as well.

Brock: What they also could do is embed their lead form as a native style or some sort of creative template that takes those positions — so they can at least see if they can capture people at the end of the image cycle.

Ben: Yeah, that's a really good point. We've then got our next important unit here — a 300x250 — right by the contact seller button, which actually seems like a pretty good position. You'd want it higher up the page, but if you can't get that just because of business rules, that's quite a nice position for it to be in. Anything you would do here to improve or enhance?

Brock: Yeah. I think I would put my product manager hat on and wonder: should my second lead form be in that position? Do I need to move the price history above, then the lead form, then the ad? Or put the ad between the two? Because this is counter to what they're doing on desktop. And that might be because they have a much more thought-out desktop and mobile strategy than I'm perceiving as I'm flicking through the site.

Ben: But it is strange to me that if you scroll up a little bit — a form — go a bit further down — yeah, they've now got their native recirculation-type product, another native, and then a 320x50. So we've got two odds kind of sandwiched at the bottom here as well. This looks to me like it might just be a legacy website management CMS thing.

Brock: Yeah, super easy to fix — which is the best part really.

Ben: The only thing I'm thinking is there could be some data-driven reason we haven't thought through — or they could have decided in actual fact that the 300x250 is performing better than the insurance tab they have, or your payments on the vehicle or getting a loan or financing.

Brock: Yeah. But it's not the core call-out or CTA on page. I think in summary, you've got all of these little widgets — a few of them are ads, some of them are more standard sized, some of them are native — and then of course you've got your contact dealer buttons and your forms. So you could really run a nice multivariate test and see what goes where.

Ben: Correct, yeah. What's actually going to produce more revenue? As you mentioned before, the ad might be doing pretty well — and then where are we getting the actual interaction, which is the most important thing when you're trying to sell cars?

Why is outstream video a big opportunity for Cars.com?

Ben: Yeah, definitely. Anything else to call out? I know it's not coming up on mine, but they had a really nice outstream piece as well. Do you want to explain a little bit about outstream and how that was being used?

Brock: Yeah, I think outstream for this website is absolutely perfect. It goes back to the previous point around full-screen mobile video. This is not necessarily full screen but it is video — and car manufacturers or OEMs spend a metric ton on marketing. They do video really well. They've always had TV budgets. What we're trying to do for this kind of website is take more YouTube budgets. So what's quite easy to buy on YouTube or in other high-frequency video platforms — we want to get the high-spending OEMs, the high-spending car manufacturers, to build beautiful video creative and put that into the key positions on Cars.com. Because they just look so good from an ad point of view. The experience is nice, it's clickable, it's interactive, there's sound — they'll pay a much higher CPM for Cars.com. And the idea with outstream is you can get that video content and video advertising without having video content yourself. You don't need to be a content producer. You don't need to be on YouTube building that kind of content. You can stay as a publisher and true to your roots, but still get that video quality and that high CPM.

Definition — Outstream video: video ads that play within article or page content — not inside a video player — allowing publishers to run video advertising without producing their own video content.

Ben: Pretty crazy upside. And if Cars.com classified all of the cars on the site — they knew the make and model — and they went to the manufacturers and said, "Look, we're going to have a video in the gallery." You'll know exactly what make and model it is. In the first generation it might be exactly what you manufacture — but then again, you can sell competitive spaces to competition, which might be very interesting. So if I'm flicking through, say, a Mercedes-AMG, and at the end I get a Mercedes-AMG video, that might be quite an immersive experience that really uses the video placement to drive the lead.

Brock: Right. So we are happy in this kind of organisation to use media as a proxy for pushing people towards actually spending money.

Ben: Definitely.

Summary: What are the biggest wins for Cars.com's ad strategy?

Ben: I think that's probably a nice place to wrap up. To run through what we've talked through here — the whole strategy is about facilitating what already exists. We're trying to make sure that we're not impacting the core business, but maybe enhancing it. We're driving revenue through the traffic we have and trying to drive that a bit better. In terms of what we saw on the desktop side, there was a bit of opportunity on the search section around maybe having better formats, enhancing the lazy loading to improve and drive viewability. We moved into the details section where we've got stricter business rules to make sure we're still pushing those leads to our dealers — but we still saw some opportunities, particularly around the gallery, to use that a bit better when we know users are going to be really engaged there. There was a quick win at the bottom of the page just making that a bigger unit. And then across mobile, the big wins we looked through were really video — maybe getting some stronger formats in there — and maybe taking a bit more advantage of the space. We guessed on the fact that we're getting maybe less buying intent on mobile — we don't know that for sure — but if that was the case, maybe we can take a little bit more advantage of the space, get some 300x250s in as well. Bits to do — I think we could get a bit of uplift there from the site. But overall, they've fulfilled on the strategy you'd expect from this kind of publisher.

Ben: I think that'll end and recap episode 3 of Layout Teardown — with myself Ben and Brock. And if you learned something, enjoyed it, or have some feedback for us, leave a comment. We'll see you here next time.

This is an edited transcript of Layout Teardown, Episode 3. The words are Ben and Brock's own — lightly edited for readability (filler words, false starts, typos, punctuation). No claims have been rewritten or generated by AI.

‍

Publift Scoop

Publift Scoop

Sign up to the Publift newsletter for all the latest news, updates and insights from across the industry.
Sign up