Your 2027 CTV budget needs evidence now

Most B2B teams will wrap up their 2027 plans between now and Thanksgiving. If you don’t have your own CTV data by then, you’ll either drop the budget line or include a number you can’t back up.

CTV often lacks a solid answer. Not because it’s weak, but because no one in the room has tested it against their own accounts. The number usually comes from a vendor presentation or a conference, and when plan cuts happen, that’s the first line to go.

It plays out this way every planning cycle. The marketers aren’t wrong about CTV. They just don’t have the data they need in the meeting.

You’ve got a quarter, give or take. If your plan closes in November, that’s eight to 10 weeks of usable runway once you account for getting a campaign live. If it closes in September, you’re reading this for the mid-year revision, and that’s fine.

Either way, it’s enough time to generate real data about CTV against your account list, and it isn’t enough time to prove CTV works. Those are two different claims, and the gap between them matters more than most vendors will tell you.

What 90 days can get you

Run a test now with the specific goal of producing planning inputs rather than a verdict, and you can answer four things before the plan closes.

1. How much of your target account list you can reach in streaming

Most teams guess at this, and the guesses are usually generous. A vendor’s addressable universe isn’t your reach. Your list is. This is an important callout because most vendors are built for B2C and may only attempt to carve out B2B.

2. What it costs to deliver against your ICP versus a general audience

Broad CTV delivery is cheap and mostly beside the point.

The number worth carrying into a planning meeting is the cost to get in front of the people you actually want, and doing it often enough to register. That’s the number that decides whether next year’s version is affordable at the scale you’d want.

3. Whether your creative survives a television

Most B2B videos were cut for a feed, where they’re fighting against constant scrolling. On CTV, it runs full-screen, unskippable, in front of someone who sat down to watch something.

Plenty of teams find out in week three that their library doesn’t hold up in that setting. Better to learn that now than in February with the money already committed.

4. Whether targeted accounts start turning up on your website

Website visit tracking is the first honest signal CTV gives you. It’s directional, not conclusive.

But if traffic from the accounts you targeted moves while traffic from the ones you left alone stays flat, something is happening, and it’s happening well before pipeline could have caught up.

Those four answers turn a 2027 CTV line from a belief into an estimate with your own inputs behind it.

What a quarter won’t buy you

Incrementality. Payback. Neither one.

If you don’t already know, here’s my guide to running a CTV incrementality test and why most fail. A test you can defend needs account-level randomization, a holdout suppressed across every line item — not just the CTV campaign, and enough conversion volume to reach significance.

In B2B, with long cycles and a finite account list, it takes longer than a quarter. Usually a lot longer. If someone tells you 90 days definitively proves incrementality, push back.

Don’t make claims. Set expectations with your CFO about what you’re looking to understand: reach, cost, creative viability, and early site signal from target accounts. Then tell them the incrementality work is for a 2027 project that this test can make possible.

Finance people are far more comfortable with “Here’s what we don’t know yet” than marketers expect. Hazy expectations are what cause plans to get nixed.

Buy narrow, not wide

There’s a market reason to keep this test tight, and it comes down to what’s happened to supply.

Ad loads are climbing. Prime Video doubled its ad load after launching its ad business in early 2024, and as eMarketer noted in February, other platforms are following.

Lighter ad loads were a big part of what justified CTV’s premium pricing. As those loads increase, streaming CPMs are flattening as supply grows. More inventory at roughly similar prices means a cheap, wide buy today sweeps in a lot of media that has nothing to do with your buyers.

Buyers can feel it. The IAB’s 2026 Digital Video Ad Spend & Strategy report has CTV spend projected to grow 11% this year. In the same research, 43% of buyers expressed “somewhat to no confidence” in the quality of the inventory they’re buying.

Targeting and audience reach now rank as equally important as business outcomes. IAB’s Chris Bruderle framed the buyer’s question as “Where is my ad actually running, where does the inventory come from, and how much of it is invalid traffic?” More money going in, less trust going with it.

For a test on a deadline, that settles the design question. A wide buy across a long tail of apps gives you a noisy result that takes months to read, and you don’t have months. A tight buy on a couple of premium apps or one trusted source of inventory, where you can get reporting on where your ads ran, reads in weeks.

Keep the design boring

You’re not trying to run a sophisticated test. More robust testing is exactly what you’re trying to set up for next year.

Instead, pick one account segment or a specific market rather than the whole list — ideally, the tier your sales team is already working in — so the results fit into a conversation that’s already happening. Concentrate delivery with specific inventory sources, because trying to run across too many won’t produce readable results.

Set frequency high enough to reach a buying committee and then leave it alone — a quarter is too short to also test frequency strategy. Run one or two creative variants instead of a full matrix.

Map out the tracking plan way ahead of the first impression. Don’t try to measure the entire buyer journey. Settle on which accounts count as targeted and which don’t. This last point may sound trivial, but I’ve seen teams go back and forth after the campaign ended.

Put the review date on the calendar now, in the same week your planning numbers are due. A readout that lands after your numbers are due is a readout for next year’s argument.

Be willing to walk away from CTV

The test can come back and tell you no. If the reach against your list is thin, or delivery against your ICP costs more than the channel could return, leave CTV out of the 2027 plan and instead put the money where it’s already working. That answer is worth the quarter, too.

Weak creative is a separate problem and a fixable one, so treat it as a reason to rebuild the spots rather than to skip the channel.

Funding the line on faith is the worst outcome. A number nobody can source gets defended in every budget conversation for a year and then quietly cut anyway, which costs a planning cycle and leaves you knowing exactly what you know today.

To recap:

• Pick the segment this week.
• Sort out the tracking.
• Get a campaign live within two weeks.
• Put the readout on the calendar for the week your planning numbers are due.

That’s it — four decisions, made early enough to be useful. Otherwise, you’ll be having this same conversation next September.

The post Your 2027 CTV budget needs evidence now appeared first on MarTech.