Video Ad Spend: Why 78% of Marketers Fail in 2025

In 2025, 78% of marketers reported a significant increase in video ad spend, yet only 34% felt confident in their ability to accurately attribute conversions back to specific video campaigns, according to a recent IAB report. This growing disparity highlights a critical challenge: how do we precisely measure the impact of video conversion within the complex, AI-driven attribution models of today?

The 78% Disconnect: Why Marketers Struggle with Video Attribution

The statistic from the IAB, indicating a substantial increase in video ad investment juxtaposed with low confidence in attribution, reveals a fundamental problem. Marketers are pouring resources into video, recognizing its power to engage, but many operate with a blind spot regarding its true return on investment. The issue isn’t a lack of data. It’s often a lack of appropriate frameworks for interpreting that data, particularly when video’s role is frequently at the top or middle of the funnel. A video conversion isn’t always an immediate click-through to purchase. It can be a brand lift, a recall improvement, or an educational touchpoint that influences a later conversion via another channel. The conventional last-click model, still prevalent in many organizations, simply misses this nuance. We need to acknowledge that video often plants the seed, and the harvest comes later, through different means.

AI Attribution Models: The Promise and The Peril

The rise of AI attribution promises to untangle these complexities, yet it introduces its own set of challenges. AI models, by their nature, learn from historical data. If that data predominantly reflects last-click conversions or over-weights easily quantifiable direct interactions, the AI will perpetuate those biases, understating video’s contribution. A recent eMarketer study found that AI-driven attribution platforms, while generally more sophisticated, still struggle with accurately assigning credit to upper-funnel content like video if not properly configured. This isn’t a failing of AI itself, but a reminder that AI models are only as good as the data they consume and the parameters we set. For instance, if your AI model for a retail client prioritizes direct response channels like paid search and ignores the brand-building power of a compelling video campaign on YouTube Ads, you’re looking at a skewed reality. You must actively educate the AI, feeding it diverse data sets and adjusting weighting parameters to reflect video’s less direct, but no less powerful, influence.

Beyond the Click: The Importance of View-Through Conversions (VTCs)

One of the most significant shifts in understanding video’s impact has been the growing acceptance and sophistication of view-through conversion tracking. A Nielsen report from late 2025 indicated that for certain industries, particularly automotive and CPG, VTCs accounted for over 30% of total attributed conversions that involved video exposure. This statistic directly challenges the traditional obsession with clicks. A user might watch a 30-second ad for a new car model, not click on it, but then days later, influenced by that exposure, visit the dealership’s website directly or even walk into a showroom. Without VTC tracking, that initial video touchpoint would be entirely lost. On platforms like Meta Business, defining appropriate look-back windows for VTCs becomes paramount. Setting too short a window (e.g., 24 hours) risks missing the delayed impact, while too long a window (e.g., 90 days) can lead to over-attribution. My professional experience suggests a 7-day to 14-day VTC window often strikes the right balance for most video campaigns, though this demands rigorous A/B testing to validate for specific audiences and product cycles.

Integrating Offline Data for a Complete Picture

The digital area only tells part of the story, especially for businesses with a significant offline presence. Imagine a local restaurant running video ads targeting specific neighborhoods in Atlanta, like those around Ponce City Market or Midtown. A user sees an ad for a new menu item, doesn’t click, but later visits the restaurant. How do you attribute that? This is where the integration of offline conversion data becomes critical. According to HubSpot research, marketers who effectively integrate their CRM data with digital ad platforms see an average of 20% higher attribution accuracy. This involves uploading customer lists, transaction data, or even loyalty program sign-ups to platforms like Google Ads for enhanced matching. For example, a customer who saw a video ad and later made an in-store purchase using a loyalty card could be matched, providing a powerful signal back to the AI attribution model. This requires careful data hygiene and privacy compliance, but the insights gained are invaluable for truly understanding video’s full funnel impact. Without this important step, businesses are essentially operating with one eye closed, unable to connect digital impressions to real-world revenue.

The Conventional Wisdom Misses the Mark on Brand Building

Many conventional attribution models, even advanced ones, still tend to undervalue brand-building video campaigns. The prevailing wisdom often dictates that if a campaign doesn’t drive immediate clicks or direct purchases, its value is nebulous, difficult to quantify. I strongly disagree with this perspective. Video’s unique ability to convey emotion, tell a story, and build trust has a deep, albeit often indirect, impact on conversion rates further down the line. Consider a B2B company launching a new software product with a series of animated explainer videos. These videos might not generate direct sign-ups immediately, but they significantly increase brand awareness and educate potential clients, making them more receptive when a sales representative reaches out or when they encounter a retargeting ad. The mistake is trying to force these top-of-funnel activities into a bottom-of-funnel measurement framework. We need to measure video’s brand impact through metrics like brand lift studies, search query volume increases for branded terms, and even qualitative surveys about ad recall and sentiment, rather than solely relying on direct conversion events. These softer metrics, when fed into an AI attribution model alongside hard conversion data, create a much more accurate and equitable distribution of credit. Effective video conversion tracking in the AI era demands a shift from simplistic last-click models to sophisticated, multi-touch attribution that integrates diverse data sources and acknowledges video’s unique, often indirect, influence on the customer journey.