CTV for HCPs: Can Streaming TV Reach Physicians?

The pitch for HCP CTV is easy to like. Physicians watch TV. Their professional inboxes and medical sites are crowded. Reaching them on the couch with a high-quality video spot sounds like a fresh channel. The reality depends on a chain of identity matches, each of which drops people, and on the fact that a living room TV is not a personal device.

How NPI to household matching works for CTV

A CTV ad is served to a device, and the device sits in a household. To get from an NPI to that device, a provider usually runs four steps:

1. NPI to person. The NPI is linked to a professional profile (name, specialty, practice address). The limits of what an NPI tells you already apply here: practice addresses are often wrong or shared.
2. Person to home. The professional profile is matched to a consumer identity record with a home address or personal email. This is the weakest link, because the NPPES registry lists practice locations, not homes.
3. Home to household graph. The home address is linked to the IP addresses and devices seen at that address.
4. Household to TV. The graph identifies which of those devices is a TV or streaming device that can receive ads.

Each step has its own match rate, and they multiply. A hypothetical example: if 80 percent of NPIs resolve to a person, 60 percent of those to a home, 75 percent of those to a household graph record, and 80 percent of those to an addressable TV, the end-to-end match is 0.8 × 0.6 × 0.75 × 0.8, or about 29 percent. Then you still have to win the auction or get the impression served during the flight. The broader mechanics are in how an NPI becomes an addressable audience, and the household graph article covers how the last two steps are built.

Scale: how many physicians can CTV actually reach?

For a target list of 20,000 primary care physicians, a 29 percent match leaves roughly 5,800 households. If the flight reaches 60 percent of those (a reasonable hypothetical), that is about 3,500 physicians. For a list of 1,500 subspecialists, the same math gives about 260. At that point a direct mail piece or a rep visit can be more precise and cheaper.

Two things make the real numbers harder to read. First, providers report "matched" counts that may include households linked by probabilistic methods, which carry more error. Second, some providers expand to "lookalike" households when the matched set is too small to deliver, which quietly turns an HCP buy into a consumer buy. Ask whether expansion is on, and ask for reporting that separates deterministic and probabilistic matches. The article on deterministic and probabilistic HCP identity covers why that split matters.

Spillover: who else is watching

A physician household usually has more than one person. Spouses, children, parents, and guests watch the same TV. Some share is unavoidable, and spillover is not a scandal. It does change two things.

The first is compliance. If a non-HCP is likely to see the spot, your regulatory team will probably want the creative to work for a general audience, which rules out some professional-only claims and data. Plan for that from the start rather than trying to retrofit an HCP banner message into a TV spot.

The second is measurement. If you count every reached household as a reached HCP, your HCP reach is overstated. If the physician was not home or not watching, the impression was a consumer impression. No CTV measurement method can confirm who was in the room unless it uses a panel with people meters, and panels have very few physicians in them.

What message fits HCP CTV

Given small scale and shared viewing, the strongest uses are narrow:

• Reinforcement. A physician who has seen your professional display and email, and maybe a rep detail, sees a polished video spot at home. CTV adds frequency and a different context, not first awareness.
• Disease awareness. Unbranded content about an under-diagnosed condition or a testing gap can work for physicians and the family watching alongside them.
• Launch moments. A short, high-visibility burst around a launch or a new indication, when the brand wants physicians to notice something changed.

Detailed clinical data, dosing, and efficacy charts do not belong here. A physician on the couch is not going to read a forest plot on a TV. Sequencing matters more than the spot. See how to sequence HCP media across display, video, and professional channels.

How to measure HCP CTV

Measurement for HCP CTV is harder than for other HCP channels, because there is no click and no NPI-level engagement signal from the TV itself. A practical approach uses three layers:

The denominator issue is the one I would push on first. Report reach against the full submitted list, not the matched subset. The same discipline appears in measuring HCP target list reach without hiding denominator problems. For lift, hold out a random share of matched NPIs before the flight. With a few thousand reached physicians, an Rx lift study will rarely have the power to detect a modest effect, so set expectations before the results arrive.

Questions to ask a provider before buying

1. What share of my NPI list matched to a home, and how was the home address sourced?
2. How many matches are deterministic versus probabilistic?
3. Is audience expansion or lookalike modeling on by default?
4. Will you report reach against the full submitted list?
5. Can you support a randomized NPI holdout and share exposure at the NPI level for an Rx study?
6. Which publishers and apps does the inventory come from?

Practical takeaway

Before approving an HCP CTV test, ask the provider to run your actual NPI list and return the step-by-step match funnel, from NPIs submitted to addressable TV households. If the end number is too small to matter next to your professional channels, put the budget into frequency on the channels that already reach those physicians.

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.