Streaming now accounts for nearly half of US TV viewing. Advertisers are rethinking how they buy CTV ads as fraud fears and pricing mismatches persist. Quality controls, not just deal type, are shaping campaign results.
Streaming has become a dominant force in US television, with connected TV (CTV) now making up 47.5% of total viewing time and programmatic ad spend. As CTV’s share grows, advertisers are scrutinizing how their budgets are allocated and whether their campaigns reach real viewers. The industry’s reliance on private marketplaces (PMPs) for CTV transactions is nearly total, with the Association of National Advertisers (ANA) reporting that 99.2% of programmatic CTV spend flows through PMPs. This preference is driven by the promise of curated supply, set pricing floors, and perceived quality, but it has not eliminated concerns about fraud and inventory value.
Advertisers remain wary of fraudulent impressions and invalid traffic. According to industry data, 63% of marketers struggle to confirm their CTV ads reach actual viewers, and 57% believe a significant portion of spend is lost to fraud. Despite these concerns, 72% say CTV outperforms their campaign baselines. The real divide is not between PMPs and the open market, but in the level of protection and quality controls applied. Without safeguards, more than a quarter of CTV impressions may not meet minimum quality standards. However, when pre-bid filtering, app-level verification, and supply-chain validation are in place, open market performance can improve substantially.
Pricing misalignments also complicate CTV buying. Programmatic systems often treat vastly different content environments as equivalent, applying similar CPM floors regardless of audience engagement or content quality. This can result in low-value placements-such as display banners on static content-absorbing a disproportionate share of campaign budgets, even when fraud controls are enforced. Buyers may find themselves paying premium prices for inventory that does not match their expectations for CTV, prompting a need for more granular control over where their ads appear.
PMPs were initially seen as the solution to supply chain opacity, offering direct access to known publishers. But as transparency has improved, buyers can now reach premium inventory through the open market with similar directness. Meanwhile, some PMP deals have started blending high-quality and lower-quality supply to meet volume commitments, reducing the clarity and control buyers once expected. Curation, the next evolution of the PMP model, promises performance but often limits buyer visibility and control even further. Unless buyers’ own data is used to construct curated packages, these deals can become supply-side bets with little transparency into actual inventory.
Effective CTV buying now depends less on deal type and more on the buyer’s ability to enforce quality controls. ANA’s Q1 benchmark report found that top-performing advertisers converted 54% of programmatic spend into qualified impressions, compared to just 32.1% for lower performers. The difference was not in transaction costs or deal types, but in how actively buyers managed waste and enforced quality standards. The most durable solution is a supply intelligence layer that classifies inventory before bidding, allowing buyers to direct spend toward premium, legitimate, or excluded apps as needed. Curation and PMPs remain useful, but only when built on systematic classification and transparency. As the industry continues to evolve, the central question is whether structured buying solves problems that better tooling and controls could address directly.
For additional perspective on how advertisers are shifting their focus from reach to measurable outcomes in digital video, see this analysis of the open internet’s progress against walled gardens: how open internet ad spend is closing the gap with major platforms.
