How to Measure Deduplicated Reach Across CTV and Online Video
How to report cross-channel video reach honestly by separating observed, modeled, and unavailable duplication between CTV and online video.
The short answer
When a brand runs CTV and online video together, the question is how many unique households or people were reached in total. Adding each channel's reach double counts everyone who saw both. Deduplication removes that double counting. The difficulty is that duplication is often only partly observable, and reports rarely say how much of the deduplication is measured versus modeled.
Why deduplication is hard
CTV reach is usually measured at the household level. Online video reach is often measured by device or cookie. Linking a household's TV to its phones and laptops requires an identity graph, which may be incomplete or inaccurate. Different partners use different graphs. So the overlap between channels is partly observed, partly modeled, and partly unknown.
Report three kinds of duplication
| Type | What it means | Confidence |
|---|---|---|
| Observed | Overlap directly measured through shared identifiers | Higher |
| Modeled | Overlap estimated statistically, often from panel data | Medium |
| Unavailable | Overlap that cannot be measured or estimated | Unknown |
A good report says how much of the total reach estimate rests on each. If most deduplication is modeled, the combined reach number is an estimate, not a measurement.
A simple worked example
Illustrative figures:
- CTV reached 1.2 million households.
- Online video reached 1.5 million households.
- Observed overlap: 300,000 households.
- Modeled additional overlap: 200,000 households.
Combined reach: 1.2m + 1.5m minus 0.3m observed minus 0.2m modeled = 2.2 million households. Of the 500,000 total overlap, 60% was observed and 40% was modeled. If the modeled share is wrong, combined reach could be meaningfully higher or lower.
Ask partners for the method
- What identity graph links CTV households to other devices?
- What share of overlap is directly observed?
- What model estimates the rest, and how was it validated?
- Is reach counted in households, devices, or people?
The MRC standards distinguish measured and modeled audiences, which is a useful frame for these conversations.
Use a single source where possible
Reach deduplication is more reliable when one measurement provider sees both channels. When each partner reports its own reach, deduplication across them is usually rough. Consider a third-party cross-channel measurement provider if cross-channel reach is a key decision input.
Why it matters for frequency
Duplicated reach means some households are seeing more ads than either channel reports. This affects frequency management and can hide over-exposure. See how to interpret reach, frequency, and duplication together.
Practical takeaway
Require every cross-channel reach report to split overlap into observed, modeled, and unavailable. Treat combined reach as a range when the modeled share is large.
Sources
External guidance and platform documentation change. Links were current at publication; check them again before relying on them for a decision.
Editorial note. Analysis and frameworks are the author's own and do not represent Acxiom or any current or former employer, client, or named platform. Examples labeled hypothetical or illustrative are not results from real campaigns. Nothing here is legal, regulatory, or medical advice.
New pharma programmatic breakdowns, occasionally
One email when I publish something worth reading. Benchmarks, measurement teardowns, and case studies with the caveats attached. No cadence promises, no reselling your address.
Unsubscribe any time. See the privacy policy.
Working through this decision on a real plan?
I work on health and pharma data, identity, and activation, after five years running HCP and DTC programmatic agency-side. Happy to talk through how this applies to your situation.