If you’ve ever sat across from a media vendor during a CTV campaign pitch, you’ve probably heard some version of this promise: “We guarantee 98% delivery on your connected TV buy.” It sounds impressive. It sounds reliable. And for many media buyers, it sounds like exactly what they need to justify signing a contract.
But here’s the uncomfortable truth — CTV delivery benchmarks, especially that suspiciously clean 98% figure, are one of the most misunderstood and frequently manipulated metrics in modern media buying. Understanding what’s really behind these numbers can save your brand thousands of dollars, protect campaign performance, and expose the proposal lies that vendors hope you’ll never notice. In this article, we’re pulling back the curtain on how CTV delivery metrics actually work, why 98% is often a misleading benchmark, and the three most common lies buried inside media proposals that use this number to close deals.
What Is CTV Delivery and Why Does It Matter?
CTV delivery refers to the percentage of contracted impressions or ad spots that are actually fulfilled during a connected television campaign. When a media buyer purchases a specific number of impressions across CTV platforms — think Roku, Amazon Fire TV, Apple TV, or smart TV apps — delivery measures how many of those impressions were actually served to real viewers.
At first glance, this seems like a simple metric. You buy 1 million impressions, and the vendor delivers 1 million impressions. But the reality is far more complicated. The CTV ecosystem is fragmented, technically complex, and filled with nuances that make delivery figures easy to manipulate.
Why does delivery matter so much? Because your media budget is allocated based on the assumption that impressions will be delivered. Under-delivery means wasted budget, missed reach goals, and campaigns that fail to perform as planned. In a world where advertising accountability is everything, delivery is the foundation of every CTV campaign.
And yet, it’s also one of the most frequently gamed metrics in the industry. Here’s why.
The 98% CTV Delivery Benchmark: Where Does It Come From?
The 98% figure has become something of an industry standard talking point. Vendors toss it around in proposals as if it were a verified, universal truth. But where does it actually come from?
In reality, there is no single, universally agreed-upon standard for CTV delivery benchmarks. Different measurement partners, DSPs, and direct publishers use different methodologies. The 98% figure is often self-reported by vendors, based on internal metrics that may not align with third-party verification standards.
Here’s how the benchmark gets constructed:
- Vendors measure delivery using their own ad servers, which may count impressions differently than a buyer’s measurement tool.
- They often calculate delivery across their entire network, masking poor performance on specific placements or devices.
- The 98% figure may include impressions served to low-quality environments that technically “delivered” but provided zero real value.
- Discrepancies between vendor-reported and third-party-verified delivery can range from 5% to 20% or more.
The bottom line? When a vendor says “98% delivery,” they are giving you a number that is built to impress, not to inform. And that’s just the beginning of the problem.
Proposal Lie #1: “98% Delivery” Doesn’t Mean What You Think It Means
This is the most fundamental lie embedded in CTV proposals, and it works because the language sounds precise. When most media buyers hear “98% delivery,” they assume it means 98% of their paid impressions reached a real human viewer on a connected TV screen. That assumption is almost always wrong.
What Vendors Actually Count as “Delivered”
Most vendors define a “delivered” impression as an ad that was served by their ad server — not necessarily an ad that was seen, completed, or verified by a third party. Here’s the critical distinction: – 10 Pitfalls to Sidestep in CTV Advertising
- Ad server delivery: The vendor’s system records that an ad was sent to a device. This counts as delivered even if the device never rendered it.
- Rendered delivery: The ad actually appeared on the screen. This is a higher standard that not all vendors use.
- Verified delivery: A third-party measurement tool confirms the ad appeared on a connected TV device in a valid environment. This is the gold standard — and the least commonly reported.
The gap between these definitions is enormous. A vendor can claim 98% delivery based on server-side logs while a third-party tool shows only 80% verified delivery on actual CTV screens. The proposal doesn’t lie outright — it just uses a definition of “delivery” that favors the vendor’s numbers.
The Role of Invalid Traffic in CTV Delivery Claims
Invalid traffic (IVT) is another hidden factor. CTV is increasingly targeted by sophisticated bot traffic and spoofed device IDs. When a fraudulent impression is served, many vendor ad servers still count it as “delivered.” This means a portion of that celebrated 98% delivery rate may include impressions that were never seen by a human being.
Industry estimates suggest that CTV ad fraud can account for anywhere from 5% to 17% of all programmatic CTV impressions, depending on the inventory source. A delivery benchmark that doesn’t filter for invalid traffic is a benchmark built on sand.
Proposal Lie #2: Delivery Rate Is Not the Same as Viewability or Completion Rate
Here’s where media buyers often get confused, and vendors are happy to let that confusion persist. Delivery rate, viewability, and completion rate are three entirely different metrics — yet proposals frequently use them interchangeably or present delivery data in a way that implies all three are equally strong. (Learn more about ctv delivery)
Breaking Down the Three Metrics
- Delivery Rate: What percentage of contracted impressions were served by the ad server. This says nothing about quality.
- Viewability: Whether the ad was actually viewable on screen according to MRC standards. In CTV, viewability standards are still evolving.
- Completion Rate: What percentage of served ads were watched to completion (typically 100% for CTV, since most CTV ads are non-skippable).
A campaign can have 98% delivery, 70% viewability, and 60% completion rate — and the vendor will lead with the 98% delivery headline in every performance report. The other numbers get buried in appendices or footnotes.
Why This Matters for Campaign Outcomes
If you’re buying CTV advertising to drive brand awareness, reach, or direct response, completion rate and viewability are far more important than raw delivery. An ad that was “delivered” but never seen provides zero value to your campaign.
The best media plans set explicit KPIs for viewability and completion rate alongside delivery. If a vendor’s proposal only highlights delivery benchmarks, that’s a red flag. Ask directly: “What is your average completion rate and viewability rate for this inventory?” If they struggle to answer, the 98% delivery claim is doing a lot of heavy lifting to distract from weaker performance elsewhere.
Proposal Lie #3: The Inventory Behind That 98% May Not Be Premium CTV
This is perhaps the sneakiest of the three proposal lies, because it requires understanding the complex and often murky world of CTV inventory sourcing. Not all CTV inventory is created equal — and the inventory that hits that 98% delivery benchmark may not be the premium, brand-safe, living-room CTV environment you think you’re buying.
The CTV Inventory Spectrum
CTV inventory ranges widely in quality:
- Premium SVOD/AVOD apps: Hulu, Peacock, Tubi, Pluto TV — high-quality, brand-safe environments with real viewership data.
- Mid-tier streaming apps: Niche or regional streaming services with decent but less-verified audiences.
- Long-tail CTV inventory: Thousands of obscure streaming apps with minimal oversight, limited viewership, and higher fraud risk.
- OTT gray areas: Inventory that is technically served on a connected device but in environments that blur the line between CTV and web video.
Many vendors pool all of this inventory together to hit their delivery benchmarks. A campaign that promises premium CTV delivery may actually be fulfilling a significant portion of impressions on low-quality, long-tail inventory — but the aggregate 98% delivery number looks clean.
The Problem with Bundled Inventory
Bundled inventory proposals obscure the quality breakdown of where your ads are actually running. You might be paying CPMs that reflect premium CTV rates while a substantial portion of your delivery happens on inventory worth a fraction of that price.
Always demand a detailed inventory breakdown. Ask for the specific publishers, apps, and platforms that will carry your campaign. Insist on transparency about what percentage of delivery will come from each tier of inventory. If a vendor is unwilling to provide this, the 98% delivery promise is likely covering for inventory quality issues. – Home
How Vendors Manipulate CTV Delivery Benchmarks in Proposals
Understanding the tactics vendors use to inflate or mislead with delivery benchmarks helps media buyers become sharper negotiators. Here are the most common manipulation tactics in play:
- Aggregate reporting: Combining high-performing and low-performing inventory to present a clean average that hides problem placements.
- Favorable time windows: Reporting delivery benchmarks from periods of high inventory availability, which don’t reflect performance during competitive buying windows like Q4.
- Excluding under-delivery from metrics: Some proposals report delivery only on campaigns that ran to completion, excluding contracts where delivery fell short.
- Server-side vs. client-side discrepancies: Using server-side delivery counts that consistently outperform client-side or third-party verified numbers.
- Defining impressions loosely: Counting partial views, background playback, or out-of-stream video as CTV impressions to pad delivery numbers.
Each of these tactics is technically defensible but fundamentally misleading. A well-crafted proposal can present 98% delivery figures that are completely accurate by the vendor’s own definitions while being almost meaningless by the standards that actually matter to a media buyer.
Critical Questions Every Media Buyer Should Ask Before Signing a CTV Proposal
The best defense against misleading delivery benchmarks is a rigorous set of questions asked before a contract is signed. Here’s what every media buyer should demand answers to:
- How do you define a “delivered” impression? Is it server-side, client-side, or third-party verified?
- What is your average completion rate and viewability rate for this specific inventory?
- What third-party measurement partners do you support? (IAS, DoubleVerify, Nielsen, Comscore, etc.)
- Can you provide a breakdown of inventory by publisher and app tier?
- What percentage of your CTV inventory is pre-bid IVT filtered?
- How do you handle under-delivery, and what is your make-good policy?
- Can I access campaign-level delivery reports through a third-party dashboard?
- What are your delivery benchmarks by device type (smart TV vs. streaming stick vs. gaming console)?
- Are delivery guarantees tied to specific inventory tiers, or to your full network?
- What is your historical discrepancy rate between your ad server and third-party measurement?
A vendor who can answer all of these questions clearly and confidently is a vendor worth trusting. A vendor who deflects, provides vague answers, or becomes defensive when asked these questions is one to approach with extreme caution. (Learn more about ctv delivery)
What Good CTV Delivery Actually Looks Like: Setting Real Benchmarks
So if 98% delivery is often a misleading benchmark, what should media buyers actually be looking for? Good CTV delivery is multi-dimensional — it goes beyond a single percentage to encompass quality, verification, and accountability.
Realistic and Meaningful CTV Delivery Standards
- Third-party verified delivery of 90%+: A realistic and strong benchmark when measured by an independent verification partner, not just the vendor’s own ad server.
- Completion rates of 90%+: CTV ads are largely non-skippable, so high completion rates should be standard, not exceptional.
- IVT rates below 3%: Industry best practice is to filter for invalid traffic pre-bid and post-bid, with clear reporting on IVT levels.
- Brand safety compliance of 99%+: Every impression should be served in a brand-safe environment, with clear categorization of content adjacency.
- Transparent inventory reporting: Granular breakdowns by publisher, app, device type, and daypart.
The Importance of Make-Good Policies
No campaign delivers perfectly. What separates good vendors from bad ones is what they do when delivery falls short. A robust make-good policy should include:
- Clear definitions of what constitutes under-delivery (typically delivery below 95% of contracted volume).
- Automatic credit or additional impressions to compensate for under-delivery.
- Restrictions on make-good inventory quality — make-goods should come from equivalent inventory, not low-tier fill.
- Transparent reporting that shows when make-goods were activated and fulfilled.
How to Protect Your Media Budget from CTV Delivery Manipulation
Beyond asking the right questions and setting better benchmarks, there are structural steps media buyers can take to protect their budgets and hold vendors accountable for real CTV delivery performance.
Leverage Third-Party Measurement from Day One
Never rely solely on vendor-provided delivery data. Integrate a third-party measurement partner — DoubleVerify, Integral Ad Science (IAS), Nielsen, or Comscore — into every CTV campaign from the start. Set up independent tracking pixels and demand access to real-time campaign dashboards that pull data directly from the measurement partner, not the vendor.
Build Delivery Guarantees into Contracts
Work with your legal and procurement teams to include explicit delivery guarantees in vendor contracts. These should specify:
- The minimum third-party verified delivery percentage required.
- The specific inventory tiers where delivery must occur.
- Completion rate and IVT thresholds that must be maintained.
- Financial penalties or automatic make-goods triggered by under-delivery.
Run Pilot Campaigns Before Committing Large Budgets
Before committing to a large CTV buy, run a smaller pilot campaign with full third-party measurement active. This gives you real performance data — not vendor benchmarks — to evaluate delivery quality, completion rates, and inventory composition. Use this data as leverage in negotiations for the full campaign buy.
Audit Inventory Regularly During Campaign Flight
CTV campaigns are not “set it and forget it.” Audit inventory reports weekly during campaign flight. Look for shifts in delivery sources, unexpected inventory from unfamiliar publishers, or sudden drops in completion rate. Early detection of delivery issues saves budget and prevents end-of-campaign scrambles to meet KPIs.
Conclusion: The Truth Behind the Numbers
The 98% CTV delivery benchmark is one of the most powerful — and most misleading — numbers in media buying today. It sounds authoritative, it appears in nearly every CTV proposal, and it gives buyers a false sense of security about campaign performance. But as we’ve explored in this article, that number hides a multitude of sins: vague definitions, quality gaps, inventory issues, and measurement discrepancies that can fundamentally undermine your campaign’s effectiveness.
The three proposal lies — that delivery means what you think it means, that delivery equals viewability and completion, and that all CTV inventory is created equal — are not random mistakes. They are structural features of how many vendors present CTV campaigns to maximize their chances of closing a deal and minimize their accountability for real performance.
As a media buyer, your job is to see through these lies. That means asking harder questions, demanding third-party verification, insisting on inventory transparency, and building accountability into every contract. CTV is one of the fastest-growing and most powerful advertising channels available today — but only if you hold vendors to standards that actually reflect the value they’re claiming to deliver.
The next time a vendor opens a proposal with “We guarantee 98% CTV delivery,” you’ll know exactly what to ask next. And that knowledge is worth far more than any delivery benchmark they’ll ever put in a proposal.
