Connected TV is the fastest-growing programmatic channel and the least rigorously measured for fraud. Those two facts are not a coincidence — they are cause and effect. Fraud follows money, and it follows money fastest into channels where the verification infrastructure hasn’t caught up with the spending.

The numbers from this quarter make the gap impossible to ignore. DoubleVerify’s May 2026 global study documented a 140% year-over-year surge in CTV fraud schemes in Q1 2026, a tenfold increase in fraudulent CTV apps year over year, and approximately $1.8 million lost per billion unprotected impressions. IAS has published parallel CTV fraud incidence data pointing in the same direction. And yet — this is the part that should bother every buyer reading this — fewer than one in five advertisers measure invalid traffic or fraud as a CTV KPI.

With Q2 upfront buying windows open and the MRC’s CTV-specific viewability and invalid traffic guidelines in active revision, this is the quarter to take CTV supply chain quality seriously. Here is the practitioner briefing on what the fraud actually looks like, why the measurement gap persists, and what buyers can operationalize now.

Why CTV Is Structurally Easier to Defraud Than Display

The display ecosystem spent a decade building fraud defenses the hard way. The 2018 takedown of the 3ve and Methbot operations — documented in detail by the Department of Justice when it indicted eight defendants for causing tens of millions of dollars in losses — forced the industry to develop browser-level detection, ads.txt authorization, sellers.json transparency, and supply path optimization as standard practice. None of that maturity transferred cleanly to CTV.

The structural weaknesses are specific and worth cataloging:

Server-side ad insertion breaks client-side verification. Most CTV advertising is delivered through server-side ad insertion (SSAI), where the ad is stitched into the content stream by a server rather than rendered by a client device. Verification vendors cannot run JavaScript in an SSAI environment the way they can in a browser. The verification signal arrives via server-to-server reporting from the SSAI vendor itself — which means the entity reporting on the impression’s validity is frequently part of the supply chain being verified. Fraudulent SSAI endpoints can fabricate entire streams of plausible-looking impressions, complete with spoofed device signatures, without any real device ever requesting an ad.

Device and app spoofing is cheap and scalable. A CTV bid request declares its device type, app identity, and content context — and in much of the open CTV exchange ecosystem, those declarations are unverified. Fraud operations spoof premium device types (declaring a data center server to be a living-room smart TV) and premium app identities (declaring a junk app to be a recognized streaming service). The tenfold increase in fraudulent CTV apps that DoubleVerify documented is the industrialization of this technique: spin up apps whose only function is to generate ad requests, get them into long-tail app stores, and monetize the bid stream until detection catches up.

The app store layer has no equivalent of domain reputation. On the web, a fraudulent domain accumulates a detectable history. CTV app stores across Roku, Fire TV, Samsung, LG, Vizio, and Android TV operate with inconsistent review standards, and an app removed from one store reappears in another under a new identity. App-ads.txt adoption — the CTV and mobile descendant of the web standard we covered when the IAB first released the ads.txt specification — remains incomplete across exactly the long-tail inventory where fraud concentrates.

CPMs justify the effort. CTV CPMs run five to ten times display CPMs. A fraud operation that would earn $2 per thousand fake display impressions earns $20-40 per thousand fake CTV impressions. The economics pull every sophisticated fraud operation toward CTV, and the verification gap lets them stay.

The Measurement Gap Is a Choice, Not an Accident

The “fewer than 1 in 5 advertisers measure IVT as a CTV KPI” finding deserves more scrutiny than it usually gets, because it describes a decision, not an oversight.

CTV buying still carries the halo of television. Media planners categorize it alongside linear TV — a premium, brand-safe, fraud-free channel — rather than alongside programmatic display, where fraud measurement is table stakes. The upfront buying ritual reinforces this: deals negotiated directly with recognized streaming publishers feel like buying television, and direct-sold premium inventory genuinely is lower-risk. But the growth in CTV spending is increasingly flowing through programmatic open exchange and resold paths, where the supply chain between the buyer and the actual device can include multiple SSAI vendors, resellers, and intermediaries — any of which can be the point where fabricated inventory enters the stream.

There is also an uncomfortable incentive alignment problem. Agencies reporting strong CTV performance to clients have limited motivation to introduce a fraud metric that could reclassify a meaningful share of delivered impressions as invalid. Publishers and intermediaries earning revenue on every transacted impression have less than no motivation. The parties with the clearest incentive to quantify the problem — advertisers — are the ones least equipped to do so without making verification a contractual requirement.

This is the same accountability pattern the industry has seen in every programmatic transparency study: measurement happens when buyers demand it, and not before.

What the MRC Revision Will and Won’t Fix

The Media Rating Council’s CTV-specific viewability and sophisticated invalid traffic (SIVT) guidelines are in active revision, and the update matters because the existing accreditation framework was built for browser environments. The MRC’s published standards and accreditation process remain the industry’s only independent audit mechanism for verification vendor methodology — when a vendor claims its SSAI-environment fraud detection works, MRC accreditation is the closest thing to proof the market has.

A revised CTV SIVT standard should force convergence on what counts as a valid CTV impression: how device authenticity must be established, what SSAI vendors must expose for independent verification, and how viewability translates to an environment where the ad nominally always plays full-screen. Buyers should expect verification vendors to recertify against the revised standard and should ask their vendors directly about accreditation status for CTV-specific (not general digital) detection.

What the standard will not fix is adoption. A rigorous SIVT definition that fewer than 20% of advertisers apply to their CTV spend changes nothing at the market level. The standard creates the measuring stick; only buying pressure creates measurement.

The Practitioner Checklist for Q2 Upfronts

The upfront window is the moment of maximum buyer leverage. The practical moves:

Make IVT a contractual CTV KPI now. Add invalid traffic rates, measured by an MRC-accredited third party, to every CTV insertion order and programmatic guaranteed deal. Set a maximum IVT threshold with make-good obligations. The vendors and publishers who resist this clause are telling you something.

Buy authenticated supply paths. Prefer direct publisher deals, programmatic guaranteed, and curated marketplaces where the path from buyer to device is short and documented. The OS-level transparency argument behind The Trade Desk’s Ventura operating system and its first publisher integrations is fundamentally a supply chain integrity argument: the fewer opaque intermediaries between the bid and the screen, the fewer places fabricated inventory can enter.

Enforce app-ads.txt and sellers.json. Configure DSP buying to require valid app-ads.txt authorization for CTV app inventory and to exclude undisclosed resellers in sellers.json. This is the cheapest available filter and it remains under-deployed.

Demand SSAI transparency. Ask which SSAI vendors sit in your supply paths and whether your verification vendor has server-side integration with them — not just post-hoc log analysis. Verification that the SSAI vendor self-reports is not verification.

Reconcile device-level reality. Where you have access to ACR data, panel measurement, or outcome data (site visits, conversions) tied to CTV exposure, reconcile it against reported delivery. Fabricated impressions produce no downstream behavior; a CTV line item with strong reported delivery and zero measurable outcome signal deserves a fraud investigation before a creative one.

The CTV fraud problem is quantifiable — DoubleVerify and IAS have quantified it. What the industry refuses to do is price it in. The buyers who make supply chain quality a contractual requirement this upfront season will pay effective CPMs meaningfully lower than the buyers who keep treating CTV like television. The $1.8 million per billion unprotected impressions is not an abstraction; it is the current market rate for not asking.

FAQ

Q: How big is CTV ad fraud compared to display and mobile fraud? Measured fraud rates on protected CTV inventory appear comparable to other channels, but that framing is misleading — the documented losses concentrate in unprotected inventory, where DoubleVerify’s May 2026 study put the cost at roughly $1.8 million per billion impressions. Because CTV CPMs run five to ten times display CPMs, each fraudulent CTV impression costs the buyer far more, and the Q1 2026 data showed fraud scheme volume growing 140% year over year — faster than in any other channel.

Q: Does buying CTV through programmatic guaranteed or direct deals eliminate fraud risk? It substantially reduces it but does not eliminate it. Direct deals with recognized streaming publishers shorten the supply chain and remove most opportunities for spoofed inventory to enter. Residual risk remains in SSAI-layer manipulation and in cases where a publisher’s own monetization stack resells or extends inventory through third parties. Contractual IVT thresholds with third-party measurement remain worthwhile even on direct buys.

Q: What is the difference between GIVT and SIVT in CTV measurement? General invalid traffic (GIVT) covers detectable-by-list invalid activity — known data center IP ranges, declared crawlers, obvious non-human patterns. Sophisticated invalid traffic (SIVT) covers deliberate fraud designed to evade detection: spoofed devices, fabricated SSAI streams, hijacked residential IPs, and fraudulent apps misrepresenting their identity. CTV fraud is overwhelmingly an SIVT problem, which is why the MRC’s CTV-specific SIVT guideline revision matters — list-based GIVT filtering catches very little of it.

Q: If fraud is this significant, why do CTV campaign metrics still look strong? Because the standard CTV reporting stack measures delivery, not validity. Completion rates on fabricated SSAI impressions are typically excellent — the fraud operation controls the reporting. Unless an advertiser layers in accredited third-party IVT measurement or reconciles delivery against independent outcome data, fabricated impressions are indistinguishable from real ones in a campaign dashboard. Strong reported metrics are exactly what successful CTV fraud looks like.

Further Reading from Authoritative Sources

  • U.S. Department of Justice: 3ve and Methbot ad fraud operation takedown and indictments — The DOJ’s Eastern District of New York prosecution remains the most thoroughly documented public record of how industrial-scale ad fraud operations are structured, and the spoofing techniques described in the indictment are the direct ancestors of today’s CTV device and app spoofing schemes.
  • Media Rating Council standards and accreditation — The MRC is the independent audit body for measurement and verification methodology; its in-revision CTV viewability and invalid traffic guidelines will define what accredited CTV fraud detection must demonstrate, making it the primary reference for buyers evaluating verification vendor claims.