Fox didn’t agree to pay $22 billion for Roku because it wanted to get into the hardware business. On June 15, 2026, Fox Corporation announced a definitive agreement to acquire Roku in a cash-and-stock deal valuing the streaming platform at roughly $22 billion in enterprise value, pending regulatory approval and a shareholder vote, with closing expected in the first half of 2027. Once it closes, Fox will control access to Roku’s platform, which reaches over 100 million households globally and more than half of all U.S. broadband households.
That’s not a hardware play. That’s the fifth major CTV walled garden (following Amazon, Netflix, Disney, and YouTube) and performance marketers now have to navigate all of them without any shared identity layer connecting them.
Each of these platforms holds proprietary audience data tied to its own login ecosystem, measurement methodology, and terms for what it’ll share back with advertisers. For performance marketers planning Q3 and Q4 CTV allocations, that consolidation creates a real measurement problem. How do you evaluate incrementality across platforms that have no reason to talk to each other?
There’s an answer, and it’s been sitting in your mail file the whole time. The physical household address is the one targeting and measurement signal that works across every walled garden, because every streaming household has one. And because it’s also the identity key that’s been driving programmatic direct mail measurement for years, with no platform consolidation able to touch it.
A Lot of Walled Gardens, Not Much Interoperable Identity
Here’s the landscape heading into the back half of 2026. Amazon DSP targets against its retail purchase graph. Netflix uses its own authenticated subscriber data. Disney combines Hulu, Disney+, and ESPN+ viewing into a proprietary audience graph. YouTube targets against Google’s logged-in identity. And once the Fox-Roku deal closes, Fox will be selling against Roku’s platform data, with Roku’s relationship spanning more than 100 million households worldwide.
Every one of these platforms will happily take your CTV budget. None of them will let you deduplicate audiences across platforms, run unified holdout tests, or compare incrementality on a shared identity key. You end up buying reach inside five separate black boxes, each with its own dashboard, each measuring “conversions” by its own definition.
For a performance marketer accountable to a blended CPA or total-portfolio ROAS, that’s not a minor inconvenience. It’s a structural gap in your measurement infrastructure. You can’t optimize a cross-platform CTV buy if you can’t resolve who you actually reached across platforms back to a single identity, and none of these platforms have any incentive to solve that for you. Their walled gardens are the product they’re selling.
The Household Address Persists Where Every Other Identifier Breaks
Digital identity has spent the better part of a decade in entropy, and the trend isn’t reversing. Third-party cookies are largely gone. Mobile ad IDs require opt-in, and consent rates on iOS have stayed low since Apple’s ATT framework shipped in 2021 — current global opt-in averages sit around 25%, meaning roughly three out of four iOS users are still blocking cross-app tracking. Probabilistic device graphs introduce match rates that swing wildly depending on platform and geography. Unified ID frameworks depend on publisher adoption that remains inconsistent across CTV environments specifically.
The household address, by comparison, has stayed exactly what it’s always been: a verified, persistent, deterministic identifier tied to real economic activity. Streaming subscriptions, cable bills, and home internet accounts all resolve to a physical address one way or another, because someone has to pay for the service and someone has to receive the hardware.
That’s the point worth sitting with. The household address isn’t a workaround for a broken identity graph. It’s the identity layer that already exists underneath every one of these platforms, whether or not the platform chooses to expose it to advertisers.
This is also exactly why programmatic direct mail measurement has never gone through the identity crisis digital advertising cycles through every couple of years. Matchback attribution resolves every mail piece to a specific household from the start. Holdout groups isolate lift at the household level, not a probabilistic device level. And because the household address connects cleanly to first-party CRM data, third-party demographic and behavioral files, and purchase history, it tends to carry a richer signal than any cookie or device ID ever managed to.
What This Means for Your Q3/Q4 CTV Strategy
If you’re putting real budget into CTV this half, you need a cross-channel identity framework before you need a media plan. In practice, that looks like three things.
Build your CTV and direct mail audiences from the same household-resolved CRM. When your CRM data is resolved to household addresses and enriched with third-party behavioral and demographic data, you can target the same households across both CTV and programmatic direct mail. That shared identity key is what lets you design a real incrementality test — expose one household segment to CTV only, another to direct mail only, a third to both, and hold out a control group entirely. This isn’t a hypothetical setup. It’s how Postie structures every campaign, with native holdout group creation and matchback attribution built into the process rather than bolted on afterward.
Use the household as your cross-platform deduplication layer. You can’t deduplicate audiences inside Amazon’s and Disney’s walled gardens directly — they’re not going to let you. But you can resolve your own conversion data back to household addresses and measure which homes converted, regardless of which CTV platform happened to serve the impression. Matchback attribution against a household-resolved file gives you a cross-platform view that no single CTV dashboard is ever going to hand you voluntarily.
Measure CTV incrementality against a channel you can already prove. Tracking conversions isn’t the hard part of CTV measurement. The hard part is proving those conversions wouldn’t have happened anyway. If you’re already running direct mail with holdout-based incrementality measurement, you have a working baseline for what household-level targeting produces on its own. Layer CTV on top of that, measure the combined lift against your existing direct mail holdout, and you get a real incrementality read on what CTV actually added — not a platform-reported ROAS figure generated inside a walled garden’s own attribution model, by a system with every incentive to make itself look good.
A Lot of Money Is Chasing the Same Identity Key You Already Have
Every major media company right now is spending heavily to reorganize around the household as the foundational unit of audience identity. Fox’s pending acquisition of Roku is the latest and largest example, but it’s not an isolated one. Amazon, Disney, and Netflix have all built their advertising businesses on top of authenticated household-level subscriptions in one form or another.
Programmatic direct mail has operated on exactly this identity framework since before CTV existed as a category. Every mail piece is addressed to a household. Every response gets measured at the household level through deterministic matchback attribution. Every audience is built from CRM and third-party data resolved to physical addresses, with lookalike modeling extending reach to households that look like your best existing converters.
The identity key that the biggest names in media are now spending tens of billions of dollars to consolidate around is the same one that’s been printed on every piece of mail you’ve ever sent.
The opportunity here isn’t really about direct mail replacing CTV. It’s that performance marketers now have a real reason to use the household address as connective tissue across their entire media mix, not just as a direct mail targeting unit. It will be used as the shared measurement layer that lets you actually compare what CTV is doing against a channel that’s been delivering measurable, holdout-tested ROAS all along, regardless of how many walls the streaming platforms keep building.
