Things are moving fast in CTV so it’s critical to stop, evaluate and speak to your partners to ensure you are building right and prepared for the next horizon. I have been hearing a lot about performance buyers moving into CTV, but what metrics are transferable and how does inventory not get commoditised? Is agentic adoption being embraced by publishers and buyers with equal vigour? And what about linear TV, while it’s declining, can it still offer value to digital buyers? I recently had the chance to spend a few valuable days in Bali with broadcasters, streaming platforms and buyers from across APAC at our event and many of these topics were discussed at length. India, Australia, Southeast Asia, China and Japan were all in the room and it was an incredibly valuable chance to hear directly from the people planning, buying and selling premium video in this region. The theme was The Performance Era of CTV: how buyers and sellers prove premium video is working.
A year or two ago, most conversations at events like this were about whether CTV could compete with walled gardens on their terms: speed, measurement, ease of buying. This year, the conversation was sharper. It was about proving performance the way buyers now expect it, without opening the door to being valued the same way as cheap, commodity video. Every broadcaster and streamer in that room has watched that happen to other inventory before. Nobody wants CTV to be next.
eMarketer’s latest forecast has Asia-Pacific as one of the slower-growing ad markets in the world this year, up 6.4% overall, even as digital ad spend within the region keeps outpacing the broader market at 8.3%. (Asia-Pacific Ad Spending Forecast 2026, eMarketer) We’re seeing the same pattern in our own business: PubMatic’s global CTV revenue grew 13% year over year last quarter and represented approximately 20% of total revenue. Growth is slowing in absolute terms, but digital, and CTV specifically, keeps taking share, and buyers are changing what they expect from not only us but also the broadcasters and platforms that they work with. Below are my key takeaways and observations.
Performance Measurement Fragmentation
CTV performance is real, but it needs a definition that fits the objective. Unique reach, ROAS, VCR, foot traffic, incrementality and full-funnel influence all came up as valid measures of “performance” — depending on who’s asking. Performance is not just an outcome however – it exists everywhere, right through the supply chain to make CTV competitive with walled gardens on speed, transparency and proven outcomes.
The problem begins when those different outcomes are treated as directly comparable, particularly when they are set against platform-reported results built with different identity signals, attribution windows and measurement methodologies. It is not necessarily a question of bad faith. The market has evolved through a mix of broadcast, digital-video, social and commerce-media systems that were not designed to report on the same basis. IAB Europe notes that inconsistent standards across broadcasters, FAST services, OEMs and CTV apps still make reach, frequency and outcome comparisons difficult. (IAB Europe CTV working-group Q&A)
The fix is agreeing upfront: the business outcome being pursued, the measurement method that will test it, and the rules under which CTV is being compared with other channels. For brand-building, that may mean incremental reach, attention and brand lift; for performance activity, it may mean a pre-agreed combination of conversion measurement, incrementality testing and media-mix modelling.
IAB’s 2026 Digital Video Ad Spend & Strategy report stated that 67% of US buyers reported somewhat-to-no confidence in CTV inventory quality for open exchange and the IAB’s 2026 Outlook Study found cross-platform measurement was a focus for 72% of buyers, up from 64% the prior year. (IAB Digital Video report, IAB 2026 Outlook Study)
The figures are US-specific, but the underlying challenge is recognisable across markets: as CTV moves closer to digital buying and optimisation, the industry needs consistent ways to validate quality, deduplicate reach and assess outcomes across platforms.
Agentic Demand/Supply Gap
Discussion at the event made it clear that demand-side agentic adoption is moving faster than supply-side right now, and I see that as an opportunity rather than a problem. Buyers are already working with automated tools day to day. The supply side has the chance to build something that matches that ease and speed, while still giving publishers the control and transparency they need over price and approval.
Earlier this year, ThinkResult ran the first agentic CTV campaign in APAC. Setup, QA and activation took 30 minutes, up to a 90% reduction from legacy platforms, and the campaign reached over a million households on premium CTV devices. As ThinkResult’s co-founder Jignesh Shah put it, “that time back is what lets our team focus on audience strategy and campaign optimisation.” Agentic simplifies discovery, setup, optimisation, troubleshooting and reporting. The test I’d apply to any agentic tool isn’t whether it’s new but whether it increases efficiency.
PubMatic has 20+ agents available today such as Seller Agent, Deal Management, Demand Insights and Tag Management that can vastly improve internal workflow efficiency and engagement with the demand-side agents. Adoption of these is a priority for our CTV publishers to capture that value that agentic offers.
Linear-to-Digital Bridge
Linear TV remains critically important across APAC, and in some of our biggest markets, the gap with CTV is huge. In India, WPP Media’s latest forecast puts CTV at just 17% of TV and professional video ad spend in 2026, meaning linear still holds roughly 83% of that budget. (WPP Media TYNY, via BestMediaInfo) That’s a bigger opportunity than migrating budget to CTV: building a genuine pathway for digital buyers to reach that linear audience, on measurement, and on audience planning, because it’s a far larger pool of budget than a CTV-only conversation reaches.
One of the clearest examples of that pathway came from a broadcast partner in Japan, one of the biggest broadcast TV markets in the world, where linear TV and CTV are still bought and measured largely separately. Their answer was to open broadcast inventory to digital-style buying: purchasable through web-based tools and, increasingly, programmatically, with reporting that mirrors digital metrics and lands within hours of a show airing. Ads can be instructed as little as three seconds before broadcast. That’s the bridge: a way for digital buyers to reach linear audiences, while linear keeps doing what it already does well.
This plays out differently market to market, which is exactly why a single playbook doesn’t work across APAC. In Japan, broadcasters and buyers are already well into that convergence. In Thailand and the rest of Southeast Asia, the more urgent conversation is different: pulling budget back from a near-total tilt toward social channels and into higher-quality formats, CTV among them, where attention and brand safety are easier to stand behind.
The Deal Setup Bottleneck
Deal setup and maintenance is sadly still the most practical gap in programmatic CTV, and I think it gets underestimated because it doesn’t sound as interesting as agentic AI or measurement standards. In the room, we heard that deal setup can take days and still end in no spend, and one platform’s own ad operations lead called it one of the biggest challenges they face. Publishers often don’t have visibility into a buyer’s budget, targeting, frequency caps or KPIs until well into the process, and sales teams frequently don’t see the level of detail that their own ad ops or the trading teams see and may not be aware there are issues.
That last point matters more than it sounds in my opinion. Publishers have spent years building sales teams who understand relationships, packaging and pricing. If automation looks like something that routes around those people rather than arms them, they won’t trust it, and they’d be right not to.
Proving Performance
None of what I’ve described above is one company’s job to finish alone, and it won’t happen on the timeline of one event. Getting there means the industry agreeing on what performance actually means before comparing it, sell-side tools that move as fast as buyers already do, real pathways into the linear budgets that still dominate several of our biggest markets, and deal processes that get faster without cutting out the people who run them.
But the room in Bali showed something the agenda didn’t plan for: broadcasters, streamers and buyers who normally only meet across a negotiating table, comparing notes with each other instead. One of our own team put it well at the very end of this event: the goal is for publishers to capture an “unfair share” of digital revenue, without ever slipping into the commoditisation that would cost them the premium they’ve built. That’s probably where it really gets built, one campaign, one resolved deal at a time.
