Measuring Reach and Frequency Across Linear TV, CTV, and Online Video

A typical pharma DTC plan runs a national linear schedule, a CTV line across a few streaming partners, and online video on YouTube and the open web. Each source reports its own reach. Adding them overstates total reach, because the heaviest TV households show up in every report. The goal is a single view of how many households were reached at least once, how often, and where the overlap sits.

Why measuring reach and frequency across linear TV, CTV, and online video is hard

The three channels are measured in different units by different systems.

Linear is planned in ratings points, CTV in impressions, and online video in views or impressions. A gross rating point and a CTV impression are not the same unit. The first job is converting everything into the same unit, usually households reached and impressions per household.

How cross-platform currencies work

A currency is the measurement source buyers and sellers agree to transact on. For decades that was panel-based ratings. Now several providers offer cross-platform measurement that combines large TV datasets with digital impression data and a panel for calibration. Different sellers accept different currencies, so many advertisers end up using one for national linear and another for streaming guarantees.

That is workable as long as you keep a single source for your own reporting. Let sellers transact on whatever currency the deal uses. Report your plan results on one consistent source so quarter-to-quarter comparisons mean something. How each provider builds its numbers is covered in how CTV measurement providers differ on reach, frequency, and device deduplication. MRC accreditation status is a useful check, as long as you confirm which metrics are covered.

Deduplication methods across the three channels

There are three layers of overlap to remove.

Linear and CTV. If the measurement source sees the TV screen itself (ACR on smart TVs, or set-top box return path data), it can observe that the same household saw your linear spot and your CTV ad. Overlap here is mostly observed, within the source's coverage.

CTV and online video. A household's TV and its phones share a home IP some of the time. A household graph links them. Much of this linking is probabilistic.

Online video across platforms. YouTube and open web video reach some of the same people. Walled platforms often limit what impression-level data leaves, so this overlap may be estimated from aggregate data or a panel.

A hypothetical example: linear reaches 9.0 million households, CTV 3.0 million, and online video 4.0 million. Summed, that is 16.0 million. Observed linear and CTV overlap is 1.5 million. Modeled CTV and online video overlap is 0.8 million. Modeled linear and online video overlap is 1.4 million. Households in all three, counted three times in the sum and removed twice through pairwise overlaps, need to be added back once. Say that triple overlap is 0.4 million. Deduplicated reach is 16.0 minus 1.5 minus 0.8 minus 1.4 plus 0.4, or 12.7 million households. Of the 3.7 million in pairwise overlap, 2.2 million was modeled. If the modeled overlap is off by 25 percent either way, total reach moves by roughly 0.55 million. The existing deduplicated reach guide recommends splitting overlap into observed, modeled, and unavailable for exactly this reason.

Frequency across channels

Average frequency hides the problem. Heavy linear viewers are often heavy streaming viewers too, and they absorb a large share of impressions. A plan can show an average of 6 across channels while a fifth of households are above 20. That costs money and annoys viewers with a 75 second spot full of side effects.

Ask the provider for the cross-platform frequency distribution, then look at the share of impressions going to households above your effective frequency ceiling. On CTV and online video you can cap by household. On linear you cannot, so the fix is usually to reduce CTV delivery to heavy linear households. Some platforms can suppress households already exposed on linear using ACR data. CTV frequency management for pharma and reading reach, frequency, and duplication together go further.

A practical approach for a pharma brand

1. Choose one reporting source for cross-platform reach before the plan is finalized, and get every partner to send exposure data to it.
2. Agree on the unit. Households for planning and reporting, unless the brand has a reason to report persons.
3. Define the universe. All US TV households, or your target audience estimate. Use the same one in every report.
4. Get exposure files from CTV and open web partners with timestamps and IDs the provider can match. Check what walled platforms will share.
5. Report overlap by type: observed, modeled, and not measured.
6. Set a cross-platform frequency target and a ceiling, and review the distribution after two to three weeks.
7. Use the incremental reach of each channel (households reached only on that channel) to decide allocation. A channel with high reach but little unique reach is mostly adding frequency.

Point 7 is the useful decision metric. If CTV adds 2 million households that linear did not reach, that is the case for CTV. If it adds 300,000, CTV is a frequency channel and should be priced that way. CTV vs. online video for pharma covers how to assign roles once you see the unique reach.

Where this fits in the bigger plan

Cross-platform reach tells you how many households had the chance to see the message. It does not tell you whether the message changed anything. Pair it with an outcome read, ideally a controlled test. Tools for measuring CTV incrementality compares the options, and the series guide on CTV and OTT healthcare advertising shows how the layers fit.

Practical takeaway

For the next flight, ask your cross-platform provider for one table: unique reach by channel, reach shared by each pair of channels, and the share of each overlap that is observed versus modeled. Use the unique reach column to decide whether CTV is buying new households or extra frequency.

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.