For years, “performance” measurement was basically credit assignment. Last-click doesn’t tell you the full story of what caused a conversion; it tells you where the conversion showed up. That was tolerable when channels were siloed. It’s not tolerable in 2026, when CTV is increasingly bought with growth expectations and stitched into cross-device funnels.
The market is also hearing a more candid admission from the biggest platforms: last-click can materially overstate impact – by more than 30% in some analyses – because it rewards whoever happens to be nearest to the moment of conversion.
In CTV, where exposure is often view-through, and actions happen on other screens, that distortion can turn “good ROAS” into a mirage. That’s why incrementality is becoming a necessary baseline. In practice, it means fewer arguments about attribution models and more routine experimentation baked into media plans.
Standards are forming
As CTV becomes more competitive – with more buyers, higher CPMs, and more scrutiny – standards have risen. Performance teams can usually make CTV work in tests. The hard part is defending scale. Finance and procurement want to know whether the spend created new outcomes or simply intercepted demand that would have converted anyway.
As that posture spreads, “show me attribution” won’t be enough; “show me lift” becomes the cost of entry. Expect this to show up first in managed-service deals, then as the norm in larger self-serve commitments.
This won’t kill click-through measurement. Most growth advertisers will run incrementality in parallel with attribution, using attribution for fast optimization and incrementality as the validation for scaling budgets.
Why parallel measurement works
To be clear, attribution is still useful, for speed and capturing a moment in time in particular: identifying tracking discrepancies,…