Live sport is one of the few things on a screen that still gathers a large audience at the same moment. The Economist's The World Ahead 2026 made the point plainly: as streaming splits audiences into niches, live sport is what still draws the concurrent crowds advertisers pay for. Rights money follows. S&P Global Market Intelligence expects the annual value of US TV and streaming sports rights to nearly double between 2017 and 2027 (cited in EMARKETER's Advertising Trends to Watch in 2025).

Yet in most media plans, sports CTV still sits in the brand column. You buy the match to be seen next to a big moment, not to hit a cost per acquisition. That is changing, and the reason is less about sport than about measurement.

Why it stayed in the brand budget

Nobody clicks a television. Most digital attribution is built on an observable click or a tracked user, and a living room screen gives you neither. A household sees your spot during the second half on Saturday, searches your brand on Monday and buys on Wednesday. Last-click attribution hands the sale to paid search. The match never shows up in the report.

Add fragmentation. The same fixture runs on a broadcaster, a pay-TV app and a streaming service, each with its own currency. Deduplicated reach is hard to calculate, and the conversion path breaks the moment the viewer picks up a phone.

So large advertisers with brand budgets absorbed the uncertainty, and everyone else stayed out. A performance team cannot defend a line in the plan whose only proof is "trust the reach numbers". The trust problem is still real: a survey of marketers reported in The Drum found that only a third fully trust the CTV performance numbers platforms report, and more than 70% said they would spend more if the proof were stronger.

What changed in measurement

Two methods that do not need a click are back in use, and both suit live sport.

Media mix modelling (MMM). MMM fell out of fashion when every user event could be tracked. It is back because user-level tracking breaks down across walled gardens, in privacy-restricted environments and on CTV, where identity across devices is patchy. MMM does not need to identify anyone. It relates weekly or daily spend by channel to business outcomes over time, and estimates what each channel adds. To include sports CTV, the model needs spend and impressions by platform and week, the creative in rotation, and outcomes at the same level of geography and time.

Holdouts. A holdout shows what would have happened without the ad, which view-through reporting never does. View-through tells you a household saw the spot and later converted. It cannot tell you whether that household would have converted anyway. A worked example published on LinkedIn by Manmohan Singh puts numbers on the gap: on the same USD 504,000 CTV campaign, view-through attribution reported USD 23 per conversion and an incrementality holdout reported USD 112. Budgets are usually set on the first number.

A geo holdout is the version a lean team can run. Pick matched markets, run sports CTV in some and not in others, hold everything else steady, and compare outcomes over the test period. A tournament gives the test a natural start and end. Say a travel brand runs a 4-week holdout across a regional football tournament: half of its test cities see the campaign, half do not, and branded search and bookings are read city by city.

The two methods work best together. The holdout gives a clean causal read on one flight. The MMM carries that read forward and gets better with each season you feed it.

How sports inventory is sold

Sports inventory does not come from one place, and the route sets the price and what you can measure.

  • Rights holders and their streaming partners. Leagues and federations license rights to broadcasters and streaming services, which sell the ad breaks. The biggest packages still go in upfront commitments. Amazon told investors on its Q3 2025 call that it exceeded its own expectations for live sports upfront commitments for 2025-2026.
  • Live event breaks. In-game breaks carry the premium. Pre-match, half-time and post-match windows reach much of the same audience, and rates vary widely by market, sport and stage of the tournament.
  • Sponsorship. Title sponsorship of a broadcast, a segment such as the match highlights, or a recurring element like the line-up graphic. These are sold direct and priced as packages.
  • In-content formats. Overlays, squeeze-backs and branded moments that sit inside the stream rather than in a break. Programmatic buying struggles with these: an AdExchanger report on agent-to-agent CTV buying noted that buying through agents lets an advertiser ask for the first or last slot of every break, or a full-screen moment at the final whistle, which standard auction specs do not describe well. Our guide to in-content CTV formats covers how they are specified.

Private marketplace deals have opened part of this inventory to buyers without an upfront. The sponsorships and in-content formats, which are often the most memorable placements, still sit mostly in direct deals.

Where APAC buyers start

Start small and measured, not big and hopeful.

  1. Pick the event and the window. A tournament with a clear start and end. Decide whether you need live in-game breaks or whether the windows around the match do the job.
  2. Set the test before the flight. Choose matched markets, decide which run the campaign, and write down the outcome you will read: sign-ups, bookings, branded search. The person who owns the budget signs off on this design before any money moves.
  3. Ask the publisher for delivery data you can model. Impressions by market and by day, matched to your test cells. If a seller cannot report at that grain, you cannot run the test, whatever the CPM.
  4. Run the numbers. Open-source MMM packages such as Google's Meridian and Meta's Robyn let an in-house team build a first model without a commissioned study. The output is approximate, but it gives a defensible basis for the next budget conversation.
  5. Feed it back. If the test shows lift, add sports CTV to your regular MMM inputs so every season improves the next estimate.

APAC teams have one practical edge. Many regional streaming services that carry sport are still building their ad sales, so a mid-sized brand can often get a direct conversation, a test structure and market-level reporting that a global upfront buyer takes for granted.

Where agents help

The hard part of a measured sports test is often the paperwork: the same brief sent to several publishers in several markets, answers in different shapes, delivery reported late. With Skopa, the buying agent from Oderra, a buyer sends one brief to every matching selling agent, compares the plans side by side and confirms the one to book; delivery is then read against that plan, shortfall first, so the test cells stay clean.