AppsFlyer CTV Attribution: 2026 Marketer Wins

In 2026, getting CTV ad attribution right is still a huge problem for marketers, leading to wasted ad spend and a lot of confusion about what’s actually working. Without being able to measure it properly, your big investment in CTV advertising becomes a black box, hiding which ads are actually getting users to install your app or buy something. Marketers have to find a way to definitively connect a CTV ad view to a later app install or purchase.

My team was staring down this exact problem eighteen months ago as we were ramping up our CTV spend. We had campaigns running across a bunch of publishers and saw a general lift in app installs, but we had no idea which CTV placements or ads were the winners. Our first attempt, just looking at impression and click data from our DSPs, gave us zero insight into what people did after seeing the ad. We couldn’t optimize anything, and we couldn’t justify the budget. We knew we needed a real ad attribution solution.

Our first mistake was trying to force old mobile ad attribution tactics onto CTV. We thought about impression-based tracking with a lookback window, but that was totally unreliable because of the insane volume of impressions on CTV and the messy, multi-device user journey. A user sees an ad on their TV, then grabs their phone a few hours later to download the app. How do you link those two things without a common ID? We learned that probabilistic matching, which can work for some mobile stuff, just wasn’t precise enough for the fragmented CTV world. Our other failed idea was just trusting the metrics from publishers. They’re okay for a first look at reach, but they’re completely siloed, so you can’t do any cross-channel comparison or figure out how CTV fits into your bigger marketing picture. We had to get a unified, independent measurement platform.

We ended up going all-in on an AppsFlyer integration, tailored specifically for CTV. This was a structured, multi-phase project, not some quick plug-and-play fix. The whole point was to connect the dots between someone seeing an ad on their TV and then doing something in our mobile app, using a mix of server-to-server (S2S) postbacks and solid data modeling.

The first thing we did was configure our CTV campaigns to fire impression and click data straight to AppsFlyer with S2S postbacks. This is absolutely necessary because CTV platforms don’t have the simple device identifiers you find in mobile. Instead, DSPs give you aggregated data or device IDs that you can’t just plug into a mobile attribution platform. We had to work directly with our DSPs to make sure every single impression and click on our CTV ads triggered a secure, real-time postback to AppsFlyer, packed with parameters like campaign ID, creative ID, publisher, and timestamp. For instance, when we ran a campaign on a CTV publisher through The Trade Desk, we put the postback URL right into The Trade Desk’s system so the impression and click data went straight to our AppsFlyer account. This gave AppsFlyer the raw CTV engagement data right from the source.

Then we had to get the data models aligned. AppsFlyer’s CTV solution is built to handle these cross-device journeys. We used deterministic matching whenever we could, like when we had logged-in user IDs across both the CTV app and the mobile device, but that’s pretty rare. For most users, we leaned on AppsFlyer’s probabilistic modeling. It looks at a bunch of data points, IP address, location, device type, time of the ad view, to make an educated guess that connects a CTV ad view to a later app install or in-app event. This probabilistic matching is more accurate than just using last-click attribution on mobile. We set our view-through attribution window to 72 hours, which seemed to be the sweet spot for catching the delayed effect of CTV ads after we analyzed our own user behavior and saw most CTV-influenced installs happened within three days.

We also had to establish a rigid naming convention for all our CTV campaigns and creatives. It sounds small, but if you don’t do this, the attribution data is basically unreadable. Every campaign, ad set, and creative got a standard name including the DSP, publisher, creative type, and audience segment, like “DSP_PublisherName_CreativeType_AudienceSegment_Q3_2026.” This careful labeling let AppsFlyer automatically parse and categorize all the incoming S2S data, which gave us a clean view of exactly which parts of our CTV campaigns were working.

Once the data started flowing, the real analysis could start. The AppsFlyer dashboard was our mission control for CTV performance. We zeroed in on a few key metrics: view-through installs (VTIs), click-through installs (CTIs), and especially in-app event completion rates. This let us focus on actual business outcomes instead of just impressions and clicks. We could see that one ad on a specific CTV app was driving a much higher VTI rate and a lower cost per install (CPI). Having that kind of granular insight let us shift budget to the better-performing campaigns in near real-time, maximizing our return on ad spend (ROAS). An early 2026 eMarketer report predicted CTV ad spend would hit over $30 billion globally, so having this kind of precise attribution is obviously needed to justify that kind of investment.

One of the more advanced things we did was send custom in-app events from our app back to AppsFlyer. This meant we could attribute more than just installs. We could attribute deep-funnel actions like “subscription initiated” or “first purchase completed” right back to the original CTV campaign. So if a user saw a CTV ad, installed our app, and then subscribed within that 72-hour window, AppsFlyer would credit that subscription to the CTV impression. That level of detail gave us a much clearer picture of the real value our CTV ads were generating.

The payoff was fast and clear. In the first quarter after we got the full AppsFlyer CTV setup running, we saw a 22% increase in our CTV campaign’s return on ad spend (ROAS). This wasn’t from some random spike in installs. It was because we were spending smarter. We found the underperforming publishers and creatives and moved that money to the top 15% of our CTV inventory that was bringing in high-value users. In that same period, our cost per qualified lead (CPQL) from CTV campaigns dropped by 18%. This clarity gave us the confidence to scale our CTV budget because we knew we had a reliable way to measure its impact, and our experience matched what a recent IAB NewFronts 2026 report found: advertisers who can effectively measure CTV are 3x more likely to increase their budgets. This clarity changed everything for us, turning CTV from an experiment into a data-driven performance channel.

The numbers were great, but we also saw other big wins. Our media buying team finally got a real feel for what our audience wanted to see on CTV. We learned that short, punchy ads did great on ad-supported video on demand (AVOD) platforms, but longer, story-driven ads worked better on subscription services. This kind of insight helped us sharpen our creative strategy and make more effective ads. It also got our creative and media teams working together better because they were both looking at the same performance data.

One challenge we ran into, and this is a big one, is that privacy rules and the death of identifiers are always changing. AppsFlyer’s probabilistic models are good, but they’re not magic. We have to constantly watch what’s happening in the industry and work with AppsFlyer to adjust our attribution methods. That means keeping up with new privacy rules and making sure our data practices are compliant. It’s not something you set up once and forget. It’s a constant process of tweaking and adapting to keep your measurement accurate.

In the end, using AppsFlyer for CTV attribution gave us the transparency we needed to turn our CTV ad budget from a line-item expense into a measurable investment. This kind of approach enables intelligent optimization.

To really measure CTV ad performance, marketers have to build a solid server-to-server attribution framework using a platform like AppsFlyer. The focus must be on tracking deep-funnel events to connect ad views to real business results.