CTV Sells Like TV and Measures Like Display. Buy It Like Neither.

Connected TV is sold with the language of television. Big screen, lean-back attention, premium content, the living room. It is then reported with the language of display: impressions, viewability, click-through on a screen without a mouse. The sales pitch borrows TV's prestige and the reporting borrows digital's precision, and a buyer who accepts both halves is being flattered twice.

Start with what CTV genuinely inherits from television. The ad is full screen. It cannot be scrolled past, and in most inventory it cannot be skipped. Sound is on. Those are real advantages over almost everything else in digital, and they explain the CPMs, which run three to eight times open-web video. What CTV does not inherit is television's audience guarantee. Nielsen-era TV sold you a rating with decades of methodology behind it. CTV sells you a device graph's opinion about a household, and the room may contain one person, five people, or a dog and an empty couch. The ad plays either way, and the impression counts either way.

Now the display half. CTV reporting arrives with digital's confidence: exact impression counts, completion rates in the high nineties, sometimes even attributed conversions. Look closely at the attribution and you will usually find IP matching, which means the phone that converted shared a router with the TV that played the ad. In a household of four, on a metric like that, your campaign gets credit for a purchase made by someone who was in another room. Completion rate is even less informative; on unskippable inventory, a 97 percent completion rate measures the player, not the person. The one honest thing in the stack is frequency data, and it is the thing fewest buyers look at.

The waste in CTV concentrates in three places. Frequency mismanagement is first: buying the same household through four DSPs and six publishers with no shared cap, so your heaviest-exposed households see the spot forty times a month while your reach curve flatlines. Content adjacency is second: "premium CTV" packages routinely blend broadcaster apps with screensaver-grade FAST channels, and the blend prices at the average while delivering the bottom. Attribution theater is third, and it is where the budget defense gets written, which is how it survives.

Here is how I buy it instead. Cap frequency at the household level, two to three per week, enforced in as few buying seats as possible, because a cap you split across platforms is not a cap. Demand app-level transparency and build an inclusion list; if a seller cannot tell you which apps ran your spot, they are telling you something. Treat CTV as a reach and priming instrument and measure it the way reach deserves: matched-market holdouts and branded search lift, not IP-matched conversions. I covered why finance teams only trust that class of evidence in the number your CFO actually believes, and CTV is the channel where the gap between the dashboard and the holdout is widest.

One number worth carrying into your next CTV negotiation: ask for the reach curve, not the impression count. On a typical uncapped multi-seat buy I audited last year, 11 percent of households received 42 percent of the impressions. The plan was reported as forty million impressions of premium video. It was actually a modest reach buy plus a very expensive harassment campaign against a few hundred thousand living rooms, and the fix, one seat and a hard cap, bought back a third of the budget without losing a household of reach.

Priced honestly, against attention rather than against television nostalgia, CTV earns a place on most consideration-stage plans. It is the only digital channel where the creative canvas still resembles the thirty-second spot, and for brands whose story needs sound and time, that is worth paying for. Just buy the actual product: a full-screen impression in a probably-occupied room, with frequency you must police yourself and measurement you must build outside the platform. The romance and the precision are both extra.