Why Creative is the Missing Layer in CTV?

There's a paradox at the heart of modern CTV advertising. We've built incredibly sophisticated infrastructure to answer questions about who sees an ad (audience data), where it runs (inventory quality), and when it's shown (contextual signals). But the actual thing the viewer experiences, namely the creative itself remains largely unexamined, unstructured, and unoptimized.

This isn't a minor oversight. According to Nielsen's comprehensive analysis of 500 CPG campaigns, creative quality accounts for 49% of a brand's sales lift from advertising—far more than targeting (9%), reach (22%), or recency (5%). Google's research puts the creative contribution even higher, at 70% of campaign success. Yet while we've invested billions in optimizing the other variables, we've left the majority driver to intuition, tribal knowledge, and hope.

It's time for that to change.

The Unstructured Data Problem

Consider how your data stack works today. Audience data arrives structured along demographic segments, behavioral cohorts, purchase intent signals, all queryable and activatable in milliseconds. Contextual data is categorized along content verticals, brand safety classifications, attention signals, all feeding real-time decisioning. Inventory data is graded along viewability scores, fraud detection, environment quality, all informing bid logic.

Now consider creative. What data structure represents a 30-second CTV spot? In most organizations, it's a filename. Maybe a folder structure. Perhaps some manual tags applied by an overworked trafficking coordinator. The actual content, namely the visual elements, the audio cues, the messaging hierarchy, the emotional tone, the brand signals, the compliance status exists only as pixels and waveforms, opaque to every system that touches it.

This is the unstructured data problem, and it creates cascading failures across the entire creative lifecycle.

The Seven Broken Loops

When creative remains unstructured, seven critical feedback loops break down:

1. Strategy to Performance: Creative strategists can't see what's actually working in market because performance data exists at the creative ID level, not the attribute level. They're making decisions in the dark.

2. Brief to Execution: There's no systematic way to verify whether the finished creative actually contains the elements the brief specified. Alignment is assumed, not measured.

3. Compliance to Scale: QC and validation happen manually, creating bottlenecks that slow launches and inconsistencies that create risk. Every human reviewer catches different things.

4. Trafficking to Intelligence: As creative moves through the ad tech stack, what little metadata exists gets stripped away. By the time an ad serves, the systems that could use creative intelligence have no creative intelligence to use.

5. Serving to Relevance: Ad servers make decisions about which creative to show with no understanding of what's in the creative. They can't match creative attributes to audience preferences or content environments.

6. Measurement to Action: Post-campaign analysis tells you which creative won, but not why. Without attribute-level performance data, every insight is anecdotal and every recommendation is a guess.

7. Learning to Compounding: Institutional knowledge lives in people's heads and leaves when they do. Each campaign starts from scratch, rediscovering truths that were learned and lost before.

Why This Problem Hasn't Been Solved: A Deep Dive

If creative drives nearly half of advertising effectiveness and marketers increasingly recognize this, then why hasn't someone built the creative intelligence layer? The answer lies in a convergence of technological limitations, market timing, structural incentives, and measurement challenges that are only now being overcome.

1. The Technology Simply Wasn't Ready

Computer vision and AI have followed a remarkable but specific trajectory. While the field began in the 1960s with Larry Roberts' work on deriving 3D information from 2D photographs, practical applications remained limited for decades. The real inflection point came in 2012, when Google Brain built a neural network of 16,000 computer processors that could recognize pictures of cats using deep learning. This was legitmate a breakthrough that seems quaint now but represented a paradigm shift.

The 2010s saw transformative advances with Convolutional Neural Networks (CNNs), particularly AlexNet in 2012, which dramatically outperformed previous methods in image classification. Google's TensorFlow launched in 2015, democratizing machine learning. Apple's iPhone X brought facial recognition to consumers in 2017. But these advances focused primarily on image analysis, specifically static frames, faces, objects.

Analyzing video creative at scale requires something more sophisticated. It requires a combination of significantly sophisticated, simultaneously learning mehtods and multitued of data points. This capability has only matured in the last few years. What would have required armies of manual labelers in 2015 can now happen in seconds with modern multimodal AI models.

Source: GlobalData, "History of Computer Vision Timeline" (2020); Kotwel, "Exploring the History & Revolution of Computer Vision" (2024)

2. CTV Hadn't Reached Critical Mass

The streaming revolution took time. While early adopters embraced Netflix and Hulu, the advertising opportunity remained limited. CTV ad spend in the U.S. was just $6 billion in 2019. By 2024, it had grown to over $30 billion which represented a 5x increase in five years. The IAB projects CTV ad spending will reach $33.35 billion in 2025, with continued double-digit growth pushing toward $47 billion by 2028.

This growth was accelerated by several factors: the 2024 end of the writers' and actors' strikes enabled content production to resume, major sports events moved to streaming platforms, and self-serve programmatic tools made CTV accessible to brands of all sizes. CTV is now ranked as the #1 "must buy" channel among advertisers, according to IAB's 2025 Digital Video Ad Spend Report, with 68% of advertisers listing it as essential to their media plan.

At $5-10 billion, you can muddle through with manual processes. At $30+ billion and growing, systematic creative intelligence becomes essential. The scale now demands infrastructure that didn't make economic sense five years ago.

Source: IAB "2025 Digital Video Ad Spend & Strategy Report" (April 2025); Statista, "Connected TV Advertising in the U.S." (2024); eMarketer/MNTN Research (January 2025)

3. Signal Loss Has Made Creative More Critical Than Ever

For years, digital advertisers could compensate for mediocre creative with sophisticated targeting. Third-party cookies enabled granular audience segmentation, cross-site tracking, and precise attribution. If your creative was just "okay," you could still drive results by showing it to exactly the right people at exactly the right time.

That safety net is fraying. The IAB estimates that signal loss from existing cookie deprecation in Safari and Firefox, plus Apple's App Tracking Transparency (ATT) framework, has already limited advertisers' ability to target and track 50-60% of internet users. Apple's ATT saw only 16% initial opt-in rates when it launched in 2021. Google's Chrome cookie changes, though repeatedly delayed and now evolving into a "user choice" model—will likely see opt-in rates well under 10%.

A February 2024 survey by Basis Technologies found that 98% of marketers are concerned about signal loss, yet 49% don't feel their organization is prepared to succeed in a cookieless world. More than one-third say they won't be able to confidently reach target audiences.

"As various technical and policy limitations further restrict the ability for digital marketing to have cross-channel, high-resolution identity solutions, the subsequent loss in efficiency must be counterbalanced by a gain in effectiveness, specifically, creative effectiveness."

— ANA Creative Effectiveness Research

When you can't out-target your competition, you have to out-create them. Creative is the one variable you fully control and it's becoming more important precisely as other levers become less reliable.

Source: IAB Signal Loss Estimates (2024); Basis Technologies "Identity vs. Privacy" Report (January 2024); Epsilon "Future of Third-Party Cookies" (2024); ANA Creative Effectiveness Research (2024)

4. The Adjacent Players Don't Want to Solve It

The ad tech ecosystem is fragmented by design, with each player optimizing for their piece of the puzzle:

Ad servers optimize delivery and frequency, but have no commercial incentive to understand what's inside the creative file they're serving. A creative is just a payload to deliver.

Measurement vendors measure outcomes, completions, conversions, brand lift but not the creative inputs that caused them. They can tell you what happened, but not why.

Creative tools focus on production like editing, versioning, collaboration but the asset stops being their problem once it's exported.

Brand safety vendors analyze the content environment, not the ad. They tell you if your ad ran next to problematic content, but not if your ad itself has issues.

DSPs and SSPs care about impression volume and bid optimization, not creative quality. They'll happily serve a terrible ad millions of times if the bid clears.

Each player addresses their piece; no one owns the whole. Creative intelligence falls into the gap between everyone's responsibilities.

5. Marketers Have Vastly Underestimated Creative's Impact

Perhaps the most striking barrier has been a fundamental misperception. According to Advertiser Perceptions' February 2024 study, brands and agencies estimate that creative represents only 19% of total sales effect. NCSolutions' actual measurement reveals creative generates 49% of incremental sales—more than double what marketers believe.

This perception gap has persisted for years. As System1's Chief Customer Officer Jon Evans notes: "Creative is the number one factor in explaining the performance of your advertising and yet most marketers still don't realize it. That means that those who focus on getting the creative right have a huge competitive advantage."

Why the disconnect? Partly because targeting feels more "scientific" and controllable. Partly because creative is traditionally seen as art, not data. And partly because, until now, there was no systematic way to connect creative attributes to outcomes, so the relationship remained invisible.

Source: Advertiser Perceptions/NCSolutions (February 2024); Westwood One "Marketers Vastly Understate Sales Effect of Creative" (March 2025); System1 Research

6. Traditional Creative Measurement Doesn't Work

Even when marketers recognized creative's importance, the available measurement approaches have been inadequate:

Pre-testing gathers consumer panels to evaluate ads before launch, measuring recall, recognition, and likeability. But as Gain Theory's Senior Partner James Dodge notes, "Most pre-testing does not provide an accurate prediction of how the ad will directly impact sales because it captures immediate consumer reaction, not in-market response—in other words, a sale."

Gut instinct remains surprisingly common. Research published in the Journal of Marketing Behavior found that intuitive predictions made by marketers were "correct no more often than random chance."

Awards are often used as a proxy for creative quality. But a study by Peter Field and the IPA found that while award-winning work tends to be more effective, the correlation is far from perfect, and awards measure industry opinion, not business outcomes.

Marketing mix models (MMMs) typically don't account for creative performance at all. As Recast's analysis notes: "One dollar of ad spend on Facebook is treated the same as any other dollar spent on that channel, regardless of which ad was running."

The result? According to Marketing Week's 2024 Language of Effectiveness survey, while 80.5% of marketers believe creative effectiveness is crucial to campaign success, only 57.3% have analysis in place to measure it—and that number is actually down from 58.5% the previous year.

Source: Gain Theory/Performance Marketing World (June 2024); Journal of Marketing Behavior; Marketing Week "Language of Effectiveness" Survey (2024); Recast "Creative Performance" Analysis (February 2024); LIONS/WARC "State of Creative Effectiveness" (2023)

The Contextual Intelligence Gap

Some might argue that existing contextual intelligence solutions address this problem. They don't, and understanding why is crucial.

Contextual intelligence, as practiced by companies like Mobian, IAS, and DoubleVerify, focuses on analyzing the content environment where ads appear. This is valuable work. Knowing whether your ad runs adjacent to brand-safe content, understanding the emotional tone of the programming, detecting sentiment and themes are variables and data signals which help advertisers make better placement decisions.

But contextual intelligence answers a fundamentally different question than creative intelligence. Contextual asks: "Is this a good environment for my ad?" Creative intelligence asks: "What's actually in my ad, and how does that affect performance?"

Consider the difference:

Contextual intelligence can tell you that your ad ran during an uplifting, family-friendly cooking show.

Creative intelligence can tell you that your ad features fast-paced editing, prominent talent, a blue color palette, an emotional arc that peaks at 18 seconds, a CTA that appears for 2.3 seconds, and brand elements that match your guidelines and can correlate all of those attributes to performance outcomes.

The Compounding Effect

The real power of creative intelligence isn't any single capability. It's what happens when the loops close.

Imagine a system where measurement data automatically feeds back into ideation. Where every campaign's learnings are captured, structured, and available for the next brief. Where "what worked" isn't an anecdote from a wrap-up presentation but a queryable knowledge base that gets smarter over time.

Kantar's research, matching 450 ads with ROMI profit data from WARC's database, found that the most creative and effective ads generate more than four times as much profit as average ads. This isn't because creative teams had better intuition. It's because they found winning patterns and repeated them systematically.

This is the compounding effect. Every campaign makes the next one better. Institutional knowledge doesn't walk out the door, it lives in the system. The gap between data-rich organizations and data-poor ones widens into an insurmountable competitive advantage.

This is what creative intelligence makes possible. Not just better decisions today, but systematically better decisions forever.

The Path Forward

The CTV industry stands at an inflection point. The scale has arrived with over $30 billion in ad spend and growing. The technology has matured—multimodal AI can now analyze video at scale. The need has become urgent as signal loss is making creative the most important lever advertisers have. And the perception gap is closing as marketers are waking up to creative's true impact.

At Bodhi, we believe creative intelligence isn't a nice-to-have feature or a point solution. It's the missing infrastructure that the entire ecosystem needs. We're building it, the same way Bloomberg built financial data infrastructure, the same way companies built audience data infrastructure, the same way the industry built programmatic infrastructure.

Because creative is too important to remain a black box. Because learnings are too valuable to lose. Because the $30 billion blind spot has been blind long enough.

The creative intelligence era is here. The only question is whether you'll lead it or follow it.