A Southwest Dealer Cohort Put 20% of One Week's Budget Into CTV. Here Is What the Numbers Said Back.
The question isn't whether Connected TV belongs in a dealer's media plan. That debate is over. The question that matters now is whether you can run it at scale inside a single week, alongside seven other channels, without watching your cost per lead climb past the point where the math breaks.
A Southwest cohort of 11 to 20 dealers just answered that question. One week in July 2026. Eight channels, $40K to $50K in total spend, and a blended cost per lead that held under $30. CTV took 20% of the budget. The CPL held anyway.
Here is what actually happened, and why the channel mix that produced it is not an accident.
What Does $40K to $50K Across Eight Channels Actually Look Like?
The cohort ran eight paid channels simultaneously during the week of July 11 through 18, 2026: Google Search, Streaming CTV, Google Performance Max, Google Demand Gen, Meta Automotive Inventory Ads, Microsoft Search, Meta Traffic, and Meta Leads. That is not a list of channels a human media planner manages by hand. Each channel has its own auction dynamics, its own creative formats, its own conversion event definitions. Running eight of them in parallel, under a single weekly budget, requires something other than a spreadsheet and a weekly agency call.
Total spend across the week was $40K to $50K, reaching approximately 1.6 million impressions and driving more than 52,000 clicks. The click-through rate implies users who were interested before they arrived. The CPC stayed around $1 across the whole portfolio, which, for a market where the average new vehicle sale generates thousands in gross, is effectively free acquisition cost at the top of the funnel.
The cohort produced over 1,400 conversions at a blended CPL between $20 and $30, with a conversion rate in the 2% to 4% range. To put that in context: a single mid-tier Google Search campaign for a franchise dealer, running without the surrounding channels pulling demand forward, would typically struggle to hold that CPL on its own. This cohort held it across eight.
Why Did the Budget Split Allocate 20% to CTV?
CTV is not a performance channel in the traditional sense. It doesn't capture demand. It creates it. The rationale for a 20% allocation inside a single week is that it seeds the in-market signal the other channels harvest. Someone who sees a vehicle ad on a streaming platform is more likely to click the Google Search ad they encounter two days later. The CTV impression is the first touch; the search click is the close.
The math on this has been visible for a while. A prior Southwest cohort quarter running CTV at 15% of a $300K to $400K total spend held a blended CPL under $30 across 10 channels over 90 days. That was a quarterly result. This is a weekly one, with a heavier CTV weight and the same CPL ceiling. The channel's contribution to the overall portfolio efficiency compounds faster than most dealers expect, because the channels it feeds, primarily Google Search and Performance Max, are the ones that convert.
What's different about a 20% CTV weight versus a 15% weight is margin. At 20%, CTV is no longer the experimental line item. It is the second-largest channel in the plan, behind only Google Search at 35%. That means the budget engine had to generate enough conversion volume from the remaining channels to justify the share that CTV was holding without converting directly. It did.
How Does a Channel That Doesn't Convert Justify One-Fifth of the Budget?
This is the question every GM and every CFO asks about CTV, and it's the right question. The honest answer is that CTV's contribution is visible in what the other channels do after it runs.
Google Search took 35% of the spend and served as the primary conversion engine. Performance Max and Demand Gen split another 20%, collectively extending reach across Google's inventory while feeding PMax's audience signals with cross-channel engagement data. Meta's three channels, AIA, Traffic, and Leads, each held 5%, operating as retargeting and catalog layers that caught users who had already interacted with the brand via either the Google properties or the CTV placements.
Microsoft Search at 5% is worth noting. It is a small allocation, but it captures a specific buyer cohort that doesn't overlap cleanly with Google Search. Some dealer groups already understand that certain channels serve different buyer profiles, and the hybrid model of automating some channels while keeping others is a real operational pattern. At 5%, Microsoft represents the tail of the demand curve that Google doesn't capture. At a $1 CPC, it earns its allocation every week.
The point is that none of these channels operated in isolation. The CTV impression reached the buyer first. Google Search caught the buyer when they acted on it. PMax and Demand Gen extended that reach across Google's network. Meta held the retargeting layer. Microsoft caught the remainder. The 20% CTV allocation didn't drain performance from the others. It fed them.
What Does the Conversion Volume Tell You About the Quality of the Traffic?
A 2% to 4% conversion rate on more than 52,000 clicks means the traffic arriving at the dealer's site was qualified before it landed. That doesn't happen by accident. It happens when the targeting layers upstream, the audiences built on first-party pixel data, the OEM offer signals embedded in the creative, and the inventory feed accuracy that determines whether the vehicle in the ad is actually available, are all current at the moment of impression.
Feed currency matters here more than most dealers realize. If the vehicle in the Performance Max ad was sold yesterday and the feed hasn't refreshed, the click converts into a dead end. The buyer calls or submits a form on a unit that's gone, and the CPL calculation just counted a conversion that will never become a deal. Over a week with 52,000 clicks, even a 5% dead-end rate on conversions is meaningful waste.
The cohort held its conversion rate in the 2% to 4% range, which implies the inventory behind the ads was current. That's not a media-buying achievement. It's an infrastructure one.
What's the Measurement Implication of Running Eight Channels in One Week?
Eight channels, one week, and a blended CPL figure. That CPL number only means something if every conversion across every channel is being counted once, consistently, regardless of what browser fired the event or which platform's pixel caught it.
That is harder than it sounds. Safari's Intelligent Tracking Prevention drops client-side pixels on a significant share of iOS traffic. A dealer running Meta AIA and Google Search simultaneously, without a server-side conversion path, will undercount conversions on both platforms. The Safari problem has been documented: the pixel drops, the conversion happened, and the platform never sees it. Most dealer stacks still don't have the second sentence: a server-side path that catches what the browser blocked. When the measurement plane is incomplete, the CPL figure is optimistic in a bad way: it looks better than it is because it's missing conversions, and the budget engine optimizes toward the channels where measurement is working rather than the channels where performance is actually best.
A blended CPL that is credible requires every channel's conversion event to be consistently captured. That is a data infrastructure requirement before it is a media-buying requirement.
How AUTONOMi Runs a Portfolio Like This
What the aggregate describes is not a media plan that a human team assembled, monitored, and rebalanced across eight channels over seven days. The operational surface is too wide. The decision cadence is too fast. The inventory refresh cycle, the creative rebuild when a featured model sells, the budget reallocation when one channel's CPL drifts above threshold, these are not weekly agency tasks. They are daily infrastructure operations.
AUTONOMi's AEGIS engine makes a single daily budget allocation decision across every paid channel the dealer has active, in one reasoning pass rather than per-channel piecemeal calls. That single-pass shape is what allows a 20% CTV allocation and a 35% Google Search allocation to be held simultaneously, recalibrated against each other daily, without one channel's spend drift pulling another channel below its efficient floor. The difference between a budget that is locked and a budget that is merely frozen matters: locked allocations protect the dealer's intentional holds while leaving everything else fluid for AEGIS to optimize. The CTV allocation in this cohort was held at 20% because the budget engine respected it as a deliberate position, not because it was frozen and therefore invisible to optimization.
The inventory rebuild cascade runs daily, diffing each dealer's live website VIN by VIN and rebuilding only the affected ad groups across Google Search, Performance Max, Demand Gen, Microsoft, and TikTok. When a unit sells, the ad groups pointing to it are updated in the same cycle. That is what keeps a 2% to 4% conversion rate honest: the traffic isn't landing on dead inventory.
On the measurement side, AUTONOMi deploys both client-side pixels via Google Tag Manager and, for dealers with server-side tagging active, a first-party server-to-server path forwarding events to Meta's Conversions API, Google Ads, TikTok's Events API, and Microsoft's UET Conversions API, deduplicated by a shared event ID. The server-side path catches what Safari's ITP drops on the client-side path. The blended CPL this cohort reported is a number the measurement plane can actually support.
Eight channels. One week. One budget engine. The CPL held under $30 because the infrastructure underneath it was built to hold it there.
The Southwest Is Already Running at This Altitude. The Rest of the Country Is Catching Up.
The Southwest cohort data keeps returning to the same structural result: multi-channel portfolios, run under a single autonomous budget engine, hold CPL ceilings that individual channel campaigns cannot match. A prior week from the same cohort produced over 1,100 conversions at a comparable CPL ceiling from the same $40K to $50K spend range. This week added approximately 300 more conversions inside the same budget envelope, with a heavier CTV allocation.
The trajectory isn't subtle. More channels, better coordinated, produces more output from the same dollar. The question is not whether to run eight channels. The question is whether your current setup can actually coordinate them, or whether each channel is effectively its own island with its own agency login, its own reporting, and its own optimization logic that has never read what the other channels are doing.
If you're running two or three channels and wondering why your CPL keeps drifting, the answer may be that the channels you're not running are the ones that create the demand the channels you are running need to convert. If you want to see what the full portfolio looks like for your rooftops, model your dealer group's spend and find out what a coordinated eight-channel week actually costs at your volume.
