Performance TV and the Search Budget Rethink in 2026

Philadelphia Cream Cheese has stopped buying search ads against prospective new customers on Google. Not paused. Stopped. The reason, as Digiday reports, is disarmingly obvious in hindsight: people don’t Google cream cheese.

At the same time, AdExchanger has been pulling apart the emerging concept of “performance TV” — the idea that connected CTV inventory can now be held accountable to the same outcome-based metrics that search and social have owned for a decade. Put these two stories next to each other and you have the clearest signal yet that the channel hierarchy marketers have relied on since roughly 2012 is getting a long-overdue audit.

When Search Doesn’t Fit the Category, Stop Buying It

The Philadelphia move sounds obvious. But the real question is how many brands are still running search budgets against categories that don’t have active search intent — simply because search has always been the default “performance” channel and nobody stopped to interrogate the brief.

Search captures demand. It doesn’t create it. For a product like cream cheese, purchase decisions happen at the shelf or are triggered by recipe context on social and video — not by someone typing a category query into Google. Kraft Heinz, Philadelphia’s parent, reallocated that budget toward channels that actually match how the category is purchased. According to Digiday, the brand is doubling down on social and influencer formats instead.

The strategic implication is precise: channel selection should follow category purchase behavior, not industry convention. In SEA markets, where Shopee and Lazada run their own closed search ecosystems with strong in-platform discovery mechanics, this is especially pointed. Brands pouring budget into Google search for low-intent FMCG categories may be optimising a metric that has nothing to do with how their customers actually shop.

Performance TV Is Real — But It’s Not a Search Replacement

AdExchanger’s reporting on “performance TV” is careful to draw an important distinction. CTV can now be bought and measured against lower-funnel outcomes — website visits, app installs, product purchases — thanks to improving identity resolution and clean room infrastructure. Advertisers are treating streaming inventory less like a branding vehicle and more like a targetable, attributable channel.

But the same reporting is explicit: streaming TV is not becoming a pure-play performance channel the way search or display are. The viewing context is fundamentally different. Someone watching a show on Netflix is not in a purchase mindset in the same way someone searching for “best running shoes under $100” is. The creative requirements are different, the cadence of action is different, and the attribution window needs to be longer.

The practical upshot for media teams: performance TV is a legitimate addition to the outcome-focused toolkit, but it needs its own measurement model. Forcing CTV into the same last-click or short-window attribution framework you use for search will undervalue it structurally and lead to budget decisions based on flawed signals.

---

What This Means for Your Attribution Stack

These two stories converge on the same underlying problem: most mid-to-large brands in 2026 are still running attribution models that were designed for a two-channel world of search and social. The infrastructure hasn’t kept pace with how channels actually behave.

Clean rooms and identity resolution tools — the current frontier of adtech investment — are supposed to solve exactly this. By allowing brands to match their first-party data against publisher and platform data in a privacy-safe environment, clean rooms can in theory give you a cleaner read on how CTV exposure influences downstream purchase without relying on deterministic click paths. But the honest reality is that clean room adoption in SEA remains patchy. The major platforms support it; the mid-tier publisher ecosystem largely doesn’t yet.

For brands operating across markets like Thailand, Indonesia, or Vietnam — where the open web coexists with heavy LINE, TikTok, and Shopee usage — the attribution problem is compounded by fragmented identity signals across platforms that don’t interoperate. The answer isn’t a single magic attribution model. It’s building a measurement architecture that treats different channels with different methodologies: geo-based incrementality testing for CTV, search query share analysis for intent-driven categories, and platform-native analytics for closed ecosystems.

The Broader Signal: Channels Are Being Re-Qualified

What’s happening beneath both these stories is a broader re-qualification of channels against actual category dynamics. The playbook of “search for performance, TV for awareness” is being replaced by something more granular and more honest — and adtech infrastructure is finally sophisticated enough to support that shift, at least partially.

For SEA brands specifically, this moment matters because the channel landscape here never fully mirrored the Western model to begin with. CTV penetration is growing — particularly in markets like Singapore and Malaysia where smart TV and streaming adoption is accelerating — but it sits alongside very strong social commerce and in-app discovery behavior that has no real Western equivalent at scale. The brands that will win the next planning cycle are those willing to stress-test every channel assumption against actual category purchase behavior, not inherited media convention.

The question worth sitting with: if you ran an honest audit of your current search and CTV spend against verified purchase-intent signals in your specific category, how much of your “performance” budget is actually performing?