By TGM Team · Fri Sep 25 2026 · 5 min read
Streaming officially overtook linear TV in 2025 — but CTV monetization and connected TV ad revenue haven't caught up to that shift yet. Here's what programmatic ad-tech actually has to solve, and what publishers should build first.
The streaming industry in May 2025, did something so unprecedented that we never expected before, it out-watched broadcast and cable combined.
As per the monthly Gauge report of Nielsen, streaming was at 44.8% of all US TV viewing that month, edging past broadcast (20.1%) and cable (24.1%) combined at 44.2%, and free ad-supported streaming channels alone accounted for 5.7% of total TV viewing, more than any single broadcast network. That wasn't a blip. It was a structural handover of attention, and it's held since.
Now here's the question that matters more than the milestone itself: has connected TV ad revenue moved with it? Not entirely, and not automatically. Industry forecasts going into 2027 have flagged CTV as genuinely under-monetized relative to the time people actually spend watching it as a lean-back, highly engaged environment where ad spend still lags viewing share.
Closing that gap isn't a matter of waiting for more advertiser interest; the leftover months of 2026's own upfront numbers already shows the budget moving toward streaming. It's a matter of what's actually plumbed in behind the screen.
That's what CTV monetization means in practice, and it's worth being precise about what the ad-tech underneath it actually has to do - not just define the term, but explain the mechanism.
In simple terms, CTV monetization is the process of converting connected-TV ad inventory into revenue through programmatic infrastructure: supply-side platforms, real-time bidding, header bidding, private marketplace deals and dynamic ad insertion, rather than relying only on manually sold, direct-booked campaigns.
Though the mechanics are recognizably programmatic, the environment isn't: a smart TV or streaming stick doesn't behave like a browser, which is exactly why CTV monetization has become its own specialized discipline rather than a simple extension of display or web video ad tech.
(For the fuller breakdown of what separates CTV from OTT and linear, see Connected TV Advertising: What CTV Is & How It Works? - this piece picks up from there, on the monetization mechanics specifically.)
Ad insertion has to happen without a browser. On the open web, an ad slot loads inside a page. On a streaming device, an ad has to be stitched into a video stream in real time, through server-side ad insertion (SSAI) on most modern platforms, or client-side insertion on some apps, before real-time bidding can even happen.
SSAI in particular has to manage ad pod sequencing, frequency capping and a seamless handoff back into content, all without the viewer noticing a stitch point.
Getting this layer wrong doesn't just cost efficiency; it's the difference between an impression that's actually addressable and programmatically sellable versus one that isn't.
Multi-hop resale paths make this harder, not easier. Every additional hand-off between the original insertion point and the final buyer is another place for the ad break to be mistimed, duplicated or mismatched to content.
Verification matters more here than almost anywhere else in digital. CTV has a fraud problem that's easy to underestimate: roughly 19% of global programmatic CTV traffic was flagged as invalid traffic in late 2025, much of it from bot farms simulating television viewing to collect ad dollars for impressions no one watched.
On top of that, the industry-standard bar for what counts as a legitimately viewable ad that is set by the Media Rating Council, requires at least 50% of a display ad's pixels in view for one continuous second, or 50% of a video ad's pixels for two continuous seconds.
CTV inventory that can't be verified against fraud and viewability standards isn't really monetizable inventory yet; it's unverified reach that sophisticated buyers will discount or avoid.
Pricing infrastructure has to be more deliberate than a single floor. Reported 2026 US CTV CPMs typically range from about $20 to $40, clustering around $25, but that range reflects real differences in content tier, verification quality and directness of the supply path, not noise.
A single flat price floor across an entire CTV footprint either underprices premium inventory or prices out lower tiers of it entirely. And the supply chain itself affects what a publisher actually keeps: OpenX's "TV by OpenX" initiative, built around removing resellers and non-TV content from its CTV inventory pool and offering direct publisher integrations across more than 110 million measurable CTV devices, is a concrete example of an industry move toward directness specifically because shorter, more transparent supply paths keep more of the CPM with the publisher instead of the chain between them.
(TGM's own take on why standardized inventory alone won't fix this is in an earlier published blog "Programmatic CTV Monetization: Why Standardised Inventory Alone Won't Solve FAST's Growth Problem" this section is the operational, floor-by-floor version of that argument.)
Content classification has to be buyer-legible, not just accurate. A FAST channel aggregating third-party or licensed content, or a CTV app blending originals with syndicated programming, needs consistent, machine-readable classification of what a given ad break is actually running against, genre, brand-safety tier, live versus on-demand.
Buyers increasingly won't bid confidently on inventory they can't classify, regardless of how large the audience behind it is. This is the same transparency problem curation is built to solve on the demand side; on the publisher side, it starts with classification hygiene that has nothing to do with how good the content actually is.
CTV used to be sold and bought largely as a brand-awareness channel, reach and frequency, not outcomes. That's changed faster than most monetization roadmaps have caught up to.
Horizon Media SVP of Media Strategy Priya Mehta has argued publicly that the old brand-versus-performance framing in CTV is now a false choice: ROAS targets that would have seemed unrealistic on CTV budgets two years ago are now standard practice, made possible by measurement infrastructure that's finally caught up to the format.
That shift matters directly for monetization, because it changes what a publisher has to prove to a buyer. Reach and completion rate alone are no longer a sufficient sales pitch, attribution, verified delivery and audience-quality signals are increasingly what determines which inventory actually gets bid on, and at what price.
A publisher still selling CTV inventory purely on reach is negotiating against a measurement bar that's already moved past them, regardless of how much audience they've actually built.
Swift TV, which is the FAST app that TGM's own infrastructure powers, running directly on connected-TV devices, sits inside exactly this stack. The same insertion, verification, pricing and classification problems described above apply to inventory TGM operates itself, not just inventory it sells on behalf of other publishers.
That's a design constraint more than a marketing point: infrastructure built to monetize FAST inventory has to actually hold up on FAST inventory TGM runs directly before it's credible to offer anyone else.
Treat ad insertion as revenue infrastructure, not a technical afterthought. Whether SSAI or client-side, the insertion layer determines whether an impression can be programmatically sold at all; it's worth evaluating on latency, addressability and how cleanly it supports real-time bidding, not just on whether it works.
Budget for verification, not just demand. With roughly a fifth of global CTV programmatic traffic flagged as invalid, a publisher without active fraud and viewability verification is asking buyers to take their inventory on faith, and sophisticated buyers increasingly won't.
Segment floors by tier, not by channel. A single CTV floor price ignores real differences between premium, mid-tier and lower-cost inventory; segmented floors protect premium yield while keeping lower tiers competitive in the auction instead of priced out of it.
Shorten the supply path before adding to it. More demand partners isn't automatically more revenue if each additional hop adds cost and opacity. Fewer, more direct, better-verified paths, the direction OpenX and others are visibly moving CTV inventory toward, tend to keep more of the CPM on the publisher's side of the ledger.
Keep content classification current, not just correct at launch. Genre, brand-safety tier and live/on-demand status all change as a FAST channel's programming rotates; classification metadata that goes stale quietly degrades buyer confidence in inventory that hasn't actually gotten worse.
The attention has already moved durably, not temporarily, and Nielsen's data says so plainly. The publishers who close the gap between that attention and actual connected TV ad revenue won't get there by waiting for more advertiser demand to show up, and they won't get there by treating CTV as a bigger version of web video monetization either.
They'll get there by building the parts of the stack that make CTV inventory genuinely trustworthy to buy: clean insertion, real verification, deliberate pricing and legible classification.
That's less a marketing message than an engineering one which is four separate, unglamorous infrastructure problems, not one big strategic pivot, and it's exactly the layer TGM's monetization work, including Swift TV's own inventory, is built around.
Q1. What's the difference between CTV monetization and video ad monetization generally?
Video ad monetization is the broader category; CTV monetization specifically deals with the technical and commercial mechanics unique to connected-TV devices — server-side ad insertion, device-level verification and CTV-specific demand — rather than in-browser or in-app video.
Q2. Is CTV inventory actually scarce, given how much streaming has grown?
Not necessarily — the constraint is less about supply volume and more about how much of that supply is verified, addressable and trusted enough for programmatic buyers to bid on confidently.
Q3. Do higher CTV floor prices always mean higher revenue?
No. A floor set too high across the board can suppress bid density and cost a publisher more in lost auctions than it gains in per-impression price — segmentation by inventory tier matters more than a uniformly higher number.
Q4. Is FAST inventory monetized the same way as premium CTV apps?
The mechanics are similar, but FAST inventory typically commands lower CPMs than premium ad-supported tiers, which is exactly why tiered floor pricing and directness matter — treating all CTV inventory as interchangeable leaves yield on the table in both directions.
