CTV 能给效果营销人员的度量带来哪些启示

As performance marketers add CTV to their media plans, they need to be careful when it comes to measurement says Alex Yip, Director of Product Strategy at AppsFlyer. While CTV certainly can boost outcomes, attribution tools on other platforms may actually encourage marketers to shift budgets away from CTV.

In his quarterly column for VideoWeek, Yip explores how performance advertisers can avoid this trap, arguing that as CTV brings incremental audiences, it requires incremental measurement.

Marketers who have spent years saturating search and social eventually run out of room, and what room remains is expensive. CTV is often one of the first places they look for incremental reach beyond their existing stack. Streaming offers access to additional audiences, and buying it increasingly feels familiar to someone who has spent most of their career inside an ads manager. That familiarity extends to the data: CTV metrics can now be joined with the rest of their data in one place.

A performance marketer adds CTV to a budget built on search and social. A few weeks later, search conversions are up while streaming looks expensive against the acquisition target. The obvious move is to shift money back to search. But if CTV helped create the demand that search is now capturing, that move could undo the improvement.

This is the question CTV forces. Performance CTV providers have made it easier to buy against acquisition goals and to measure incrementality, but familiar buying does not mean familiar measurement. CTV has no click, and that changes how its value shows up. The harder part is understanding what that measurement does and doesn’t tell you, and applying the same discipline across the whole budget. Otherwise, CTV becomes another dashboard to check and another set of results to reconcile at the next planning session.

Why clicks undercount CTV

Timing alone cannot establish that CTV drove the search lift, but the mechanism is easy to see. Search and app store channels are natural steps in the user journey. A viewer who sees a streaming ad and gets interested will naturally go search for the brand on Google or the App Store, but that search is a stop along the path, not what drove the interest, and in a last-touch world it takes the credit. The marketer needs to investigate what changed because CTV entered the mix, including purchases that other channels received credit for.

Click-based reporting follows a direct interaction, and because CTV has no click, it will minimise the impact measured for the channel. View-through attribution can associate a streaming impression with a later conversion on another device, and both depend on rules about which interactions qualify for credit. Neither, by itself, establishes whether the customer would have purchased anyway. Understanding the user journey across screens is critical to accounting for omni-channel marketing, and recognising these patterns is what should trigger an incrementality test to validate what each channel is truly delivering.

There are a few ways to run that incrementality test:

• An audience holdout withholds the campaign from a random slice of the eligible audience and compares its outcomes with the exposed group. 
  • It needs clean separation between the groups and enough conversions to tell lift from ordinary variation.
• A geographic experiment runs CTV in selected markets and withholds it from comparable ones, measuring total purchases or installs so effects that surface in search and social are captured. 
  • It depends on stable marketing activity elsewhere and on accounting for spillover between markets.
• Modeled baselines estimate the activity expected without the campaign and compare it with what was observed, giving an ongoing read without a live holdout on every campaign. 
  • The estimate is only as good as the model’s assumptions and how it handles other drivers of demand.

Marketers don’t need to become statisticians to use this evidence. They do need to understand what was compared, which outcomes were counted, and how confidently the result supports a spending decision. A lift figure becomes much more useful when the team knows what sits behind it.

The twist: your organics may be your CTV

Consider a brand that has run CTV for years without measuring it for performance. The campaigns were live, but the installs they drove landed in the unattributed, “organic” column.

Then the brand starts measuring CTV properly. Organic installs drop, while CTV-attributed installs rise. A marketer looking at the dashboard may conclude that CTV is cannibalising organic and pull back.

That hypothesis is probably backward. Nothing about the business changed except the accounting. Those installs were never organic; they should have been attributed to CTV all along. Cutting CTV in response would remove what was driving them.

The check is straightforward. If total installs held steady while the mix shifted from organic to CTV, the shift is a relabeling. If the team wants to rule out real cannibalisation, a holdout or geo test will show it: withhold CTV in a region and see whether organic actually rises or falls.

Taking that discipline across the budget

The core message is that CTV brings incremental audiences and needs to be measured incrementally. This is an adjustment for teams accustomed to optimising within individual platforms. The routine is familiar: watch the acquisition cost, make changes, increase spending where results look strongest. As more channels enter the mix, overlapping attribution claims make those comparisons harder.

That matters most for omni-channel brands. If teams are living in their own silos, with TV in one group and digital in another, each sees its own numbers and the numbers will not add up; staying in a last-touch world will not do those channels justice. A CTV campaign can generate incremental sales while leaving open whether the next dollar belongs there. Answering that requires a view across the alternatives, including how campaigns affect one another. The people managing TV, search and social need to be part of the discussion, particularly when the effect being investigated appears in their results.

Neutrality matters here. A media provider has a commercial interest in demonstrating the value of its inventory. Its measurement can be rigorous and useful, while the buyer still needs an evaluation that serves the whole budget. Neutral third-party measurement and advertiser-controlled experiments help keep that broader decision in the advertiser’s hands. A neutral third party can also dedupe conversions across channels, so credit is not claimed twice and the totals add up.

That becomes especially relevant as performance CTV capabilities move inside larger media businesses. Buyers benefit from connecting streaming with broader audience and commerce capabilities, while continuing to need a consistent view across the companies competing for their spending.

For CTV, this is also the route from an experimental allocation to a permanent place in the media mix. Incremental reach gives marketers a reason to try streaming. Credible evidence of incremental outcomes gives them a reason to fund it again and build it into their growth expectations. Knowing what the business gains from CTV, and what it would lose without it, makes the budget considerably easier to defend.

Follow VideoWeek on LinkedIn.