CTV Attribution Limits Explained: Attribution links actions to CTV ads but doesn't prove causation.; Conversion windows vary—longer ones include more actions, possibly unrelated.; Providers exclude some inventory and actions; check what's missing in reports.
Image: Streaming Advertising Guide

Incrementality

Part of CTV campaign measurement

Understanding the limits of CTV conversion attribution

Learn how CTV conversions are matched and credited, what coverage and windows change, and why attribution alone cannot prove sales impact.

Measurement Desk · · 3 min read

CTV conversion attribution credits a later recorded action to an eligible television ad exposure under a provider's matching and credit rules. It does not establish that the ad caused the action. Before using an attributed-conversion figure, find out which exposures and actions the method could connect, over what period, and what it excluded.

Check the measurable path

A person may encounter an ad on a shared TV and later act on a phone or computer. A cross-device attribution system needs an eligible exposure record, a recorded conversion event and a supported way to associate them.

That capability can be limited by product, region and inventory type. Check the provider's current documentation for the relevant market and inventory before relying on a forecast or report.

A household match cannot identify which resident watched and which resident acted. Actions that the provider cannot connect to an eligible exposure will not appear as attributed actions under that method. Keep the attributed figure separate from total business actions.

Read the window and credit rule

A conversion window determines how long after an eligible exposure or click an action may receive credit. Changing it can change the reported count without changing customer behaviour. A longer window can include more later actions, including some that would have occurred anyway; a shorter one may omit delayed responses.

Floodlight conversion windows are one documented example of a configured window. Other reporting models can use different rules. Ask which model and windows apply to the actual CTV result, including any change between campaigns.

CTV Conversion Attribution: Key Factors and Their Impact

  • Short Conversion Window (e.g., 1-day)May miss delayed actions; underestimates attributed conversions
  • Long Conversion Window (e.g., 30-day)Includes more later actions, including those likely to occur regardless of ad exposure
  • Impact on Reported CountChanging the window alters results without reflecting actual customer behaviour

Audit the attributed figure

Ask for five definitions with the result:

  1. Exposure:Which CTV publishers, devices and dates could enter the match?
  2. Action:Which site or app event counted, and how were repeats or cancelled transactions handled?
  3. Match:What linked TV exposure to the later action, and what eligible delivery was outside that method?
  4. Credit:What post-view and post-click windows and attribution model applied alongside other campaign touchpoints?
  5. Coverage:Which inventory or business outcomes were excluded or unknown?

Audit the Attributed CTV Conversion Figure

  • 1. Exposure: Which CTV publishers, devices and dates were eligible?Only exposures from supported publishers and devices within the matching window are included
  • 2. Action: Which site or app event was counted?Only recorded conversion events (e.g., purchases, sign-ups) with repeat/cancel handling defined
  • 3. Match: What linked TV exposure to the action?Cross-device match via household ID or device graph; excludes unmatched actions
  • 4. Credit: What attribution model and windows applied?Post-view/post-click windows and model (e.g., last-touch, linear) influence credit distribution
  • 5. Coverage: What inventory or outcomes were excluded?Non-supported inventory types (e.g., certain OTT platforms) or business actions may be missing

Match the conclusion to the evidence

An attribution report can support a statement that the platform credited recorded actions to eligible CTV exposures within its stated rules and coverage. It cannot, by itself, say how many actions would not have happened without CTV. That stronger question needs a suitable comparison or experiment.

Key Limitations in CTV Conversion Attribution

Attribution does not prove causation
Credits actions to ads but cannot confirm they were caused by them
Household matches lack individual tracking
Cannot identify which resident watched or acted
Excluded inventory types
Some OTT platforms and non-standard CTV inventory are not measurable

More from Incrementality