Daniel Friscia
October 13, 2025
19
minutes read
The question isn't whether CTV works anymore. Your clients already know it does. The real question keeping CMOs up at night: which half of their TV budget is actually driving results?
Traditional TV metrics are dead weight in streaming-first media plans. GRPs and frequency caps made sense when viewers had three channels and appointment viewing. Today, with streaming capturing 44.8% of total TV usage and audiences scattered across dozens of platforms, those legacy measurements tell you nothing about business impact.
Smart advertisers have moved beyond surface-level stats. They're tracking cost per acquisition from CTV campaigns, measuring foot traffic lift to physical stores, and connecting ad exposure to actual revenue. The best are using AI-powered tools to predict creative performance before campaigns launch and optimize in real-time based on attention scores, not just impressions.
This shift demands new frameworks. Performance marketers need CTV measurement that mirrors their digital standards. Brand builders require proof that premium video drives recall and consideration. Retailers want to know if their streaming ads create store visits. Everyone needs attribution models that actually work across walled gardens.
We'll break down the metrics that matter in 2025: from completion rates and attention scoring to incrementality testing and offline attribution. You'll learn why some brands see 65% sales increases from performance TV while others waste millions on poorly measured campaigns.
Agencies are stuck between two worlds. Their clients demand performance accountability while TV sellers still pitch reach and frequency. The disconnect grows wider every quarter as streaming fragments audiences and traditional measurement systems fail to keep pace.
GRPs worked when mass reach equaled business results. That correlation broke years ago. A campaign delivering 500 GRPs tells you nothing about whether it drove downloads, store visits, or sales. Panel-based measurement can't track viewers across Netflix, YouTube TV, Hulu, and dozens of FAST channels. Even Nielsen's pivot to big data integration can't solve the fundamental problem: exposure metrics don't predict outcomes.
The streaming migration exposed these flaws. Linear TV measurement assumed captive audiences watching ads in predictable patterns. CTV viewers jump between platforms, devices, and content types throughout the day. They watch on phones during commutes, tablets in bed, and smart TVs in living rooms. Traditional measurement captures fragments of this behavior at best.
Yet many CTV campaigns still optimize for impressions and completion rates. Brands celebrate 95% view-through rates on non-skippable inventory without asking if those views changed consumer behavior. They buy frequency caps of 15+ when research shows diminishing returns after 14 exposures. They measure reach without considering whether they're hitting the same households across multiple platforms.
Performance marketers learned this lesson in digital advertising a decade ago. They abandoned CPMs for CPAs, impressions for conversions, reach for return. TV advertisers cling to legacy metrics because the alternative requires admitting they've been measuring the wrong things.
The data exists to do better. CTV platforms generate granular viewing data. Attribution tools connect ad exposure to online and offline actions. AI models predict which creative will drive results before campaigns launch. The gap is now philosophical. Brands must decide whether they're buying audiences or outcomes.
Modern TV ad measurement starts with a simple principle: align metrics to business outcomes. Every brand cares about different results, but they fall into four measurable categories that actually connect advertising spend to company growth.
Direct-response advertisers live and die by two numbers. Cost Per Acquisition (CPA) tells you what you paid to gain each customer. Return on Ad Spend (ROAS) shows revenue generated per dollar invested. Both metrics require attribution models that connect CTV exposure to conversions across devices and time delays.
Calculate CPA by dividing total campaign cost by conversions attributed to that campaign. A $50,000 CTV campaign that drives 1,000 new customers yields a $50 CPA. ROAS follows similar logic: $200,000 in attributed revenue from that same campaign delivers a 4:1 ROAS. These aren't estimates or models. They're measurable business outcomes that CFOs understand.
The challenge lies in attribution accuracy. Modern platforms like Elevate's Custom KPI Optimization algorithm abandons one-size-fits-all metrics. Processing data from 20+ performance indicators, it lets you optimize toward the business outcomes that actually matter. The AI continuously adjusts campaigns to hit your specific targets, whether that's aggressive CPAs, maximum ROAS, or custom KPIs that mirror your P&L. The technology adapts to your business reality, not the other way around.
For app-led goals, a mobile measurement partner like AppsFlyer or Branch can attribute CTV exposures to installs and in-app revenue events, so CPA and ROAS reflect real post-install value rather than just first opens.
{{how-to-1="/tables"}}
Brand campaigns require different proof points. TV advertising lift studies measure changes in consumer perception by comparing exposed and control groups. Key metrics include ad recall (can viewers remember your ad?), brand awareness (do they know your brand exists?), and consideration (will they consider purchasing?).
Modern brand lift studies leverage AI to increase precision and reduce costs. Instead of waiting weeks for survey results, platforms now deliver real-time lift measurements by analyzing behavioral signals and engagement patterns. For instance, Elevate's predictive planning engine runs the numbers before you spend a single dollar. The system crunches multiple data points to predict outcomes and optimize budget allocation, setting targets based on what similar campaigns actually delivered.
The most sophisticated approaches go beyond asking "did awareness increase?" to understanding emotional resonance and message comprehension. AI-powered analytics can now detect whether viewers understood your key message and whether it shifted their perception in the intended direction.
For ongoing tracking beyond the flight, YouGov BrandIndex or DISQO can monitor shifts in perception and purchase intent over time.
Incrementality answers the hardest question in advertising: what sales wouldn't have happened without this campaign? It's the difference between correlation and causation.
Three core incrementality metrics matter most:
Smart incrementality testing uses holdout groups and geo-experiments to isolate TV’s impact, with lift platforms such as InnovidXP (TVSquared) and Upwave making those designs practical at scale. In parallel, deterministic TV attribution from iSpot or VideoAmp links exposed households to downstream outcomes across devices and time to validate causation. Taken together, this multi-dimensional analysis separates lucky timing from effective advertising.
For retailers and location-based businesses, online metrics tell half the story. Real ROI measurement tracks whether CTV ads drive foot traffic and in-store purchases. Modern attribution connects digital ad exposure to physical world actions through:
The measurement tech exists. Platforms can track a viewer from CTV exposure through website visit to store purchase. The question becomes integration and privacy compliance. Smart Supply's AI-powered approach ensures clean, direct paths to inventory, making attribution more accurate by eliminating the bid stream recycling that muddles tracking.
But measuring ROI effectively takes more than isolated tools—it demands integration across planning, buying, and analysis. Why? Because true measurement starts before campaigns even launch, using inventory and targeting data to predict outcomes. As campaigns run, real-time signals drive instant optimizations. Afterward, automated insights reveal exactly what worked.
The payoff: TV ad ROI becomes as clear and actionable as search or social. With the right frameworks, the guesswork is gone.
So what’s fueling this shift to integration and full funnel, and what will truly matter for TV advertising analytics in 2025?
Analytics without action is just expensive reporting. The brands winning in CTV have abandoned vanity dashboards for systems that connect media metrics to business results in real time.
Leading advertisers now cascade business objectives directly into campaign analytics. A retailer targeting 15% same-store sales growth translates that into specific cost-per-store-visit targets for each DMA. A D2C brand seeking 10,000 new subscribers sets CPA thresholds that ladder up to customer lifetime value goals. Every media metric ties to a business metric that executives actually care about.
Research shows this approach works: brands using outcome-based optimization see 65% higher sales increases compared to those optimizing for completion rates. The key is building analytics frameworks that track both immediate performance signals and longer-term business impact.
👉 For a broader look at the forces shaping today’s media environment, read The 2025 Media Reality.
Walled gardens are crumbling. Advertisers running CTV across multiple DSPs demand unified reporting that shows true performance regardless of platform. The days of accepting different attribution windows and metrics from each platform are over.
Cross-DSP transparency requires three elements:
Smart advertisers bypass platform bias with independent supply optimization. Tools like Smart Supply sit above The Trade Desk, Amazon DSP, Google DV360, and other DSPs, normalizing data to reveal actual performance. They also expose hidden costs like bid stream recycling that can inflate CPMs by 36% or more.
Third-party cookies never mattered much in CTV, but their decline across digital advertising forced a broader reckoning with data strategy. Smart TV analytics in 2025 leverages two sustainable data sources: first-party customer data and contextual signals from content consumption.
First-party data integration transforms CTV from broad awareness plays to precision targeting. Brands upload hashed customer lists to find high-value audiences across streaming platforms. More importantly, they use first-party data to measure actual business outcomes, tracking whether exposed customers increase purchase frequency or basket size.
Contextual data provides the scalable complement. Modern contextual analytics goes beyond simple genre targeting to understand mood, themes, and viewing patterns. According to research, contextually relevant CTV ads see 2.2x higher purchase intent compared to demographic targeting alone. AI-powered crawlers now analyze video content frame-by-frame to ensure brand safety while maximizing relevance.
Static campaign analytics died with linear TV. Modern CTV analytics predicts performance degradation before it happens and adjusts automatically. This requires three technical capabilities working in concert:
{{how-to-2="/tables"}}
The shift from reactive reporting to predictive analytics changes everything. Instead of analyzing what went wrong last month, brands prevent problems before they impact performance. Analytics becomes a competitive advantage rather than a compliance requirement.
{{how-to-3="/tables"}}
👉 If you want to know more about AI's impact on TV advertising, check out the dedicated piece on the topic: The Rise of AI in TV Advertising: How Agencies Can Win the Next Frontier.
"Does TV advertising work?" The question persists because we've been measuring the wrong things. Counting impressions is like counting billboards passed on a highway; it tells you nothing about whether drivers noticed, cared, or changed direction.
TV advertising effectiveness in 2025 rests on three pillars that actually predict business impact:
Recent studies show that overall attention to CTV ads rose to 51.5%, with premium CTV content garnering an even higher 56.1%, surpassing linear TV. AI-powered contextual targeting can further boost attention by a factor of four by aligning ads with the content being watched.
Hard data validates TV's effectiveness when measured correctly. CTV consistently delivers 90%+ completion rates—not because viewers can't skip, but because targeted, relevant ads in premium content environments capture genuine attention. Compare that to 15-30% completion rates for skippable YouTube ads or 0.1% engagement rates for display banners.
Brand lift studies show even stronger results. According to research, CTV campaigns targeting local audiences can see a +106% lift in aided brand awareness among exposed viewers. That's what I call memorable reach that changes consumer behavior.
The most compelling evidence comes from cross-channel comparisons. When brands run identical creative across CTV, online video, and social platforms, CTV consistently drives higher consideration and purchase intent. The lean-back viewing environment, larger screens, and premium content context create optimal conditions for brand messaging.
The numbers don’t lie: Interactive CTV drives audience action at scale, with engagement rates 4.6x higher than mobile and a staggering 10.3x higher than desktop.
Impressions became the default TV metric because they were easy to measure, not because they predicted success. In the linear TV era, massive reach correlated with results often enough to justify the simplification. That correlation broke when audiences fragmented across hundreds of streaming services.
Today's viewers actively choose what to watch and when. They pay for ad-free tiers or tolerate ads in exchange for free content. This fundamental shift in the value exchange means every impression carries different weight. An engaged viewer watching your ad during their favorite show impacts your business differently than someone who left Pluto TV running in the background.
The old metrics also ignored the compounding effects of frequency and timing. Showing someone your ad 15 times in three days wastes money and annoys viewers. Spacing those same impressions across three weeks with sequential messaging builds awareness, consideration, and intent. Effectiveness requires strategy, not just reach.
Proving TV advertising effectiveness requires integrated measurement across the entire funnel. Start with attention metrics during exposure—are viewers actually watching? Layer in brand lift studies to measure perception changes. Connect to digital behaviors like site visits and searches. Finally, close the loop with sales data, whether online conversions or offline foot traffic.
This full-funnel approach reveals TV's true impact. A streaming campaign might show modest direct response rates but massive lift in branded search volume. Traditional last-click attribution would miss TV's assist effect. Modern multi-touch attribution properly credits TV for starting customer journeys that convert through other channels.
The technology exists to measure every aspect of TV advertising effectiveness. Platforms like TVScientific and Vibe provide unified dashboards tracking attention through conversion. The challenge isn't measurement capability but organizational willingness to move beyond impressions.
{{how-to-4="/tables"}}
The most effective TV campaigns in 2025 share three characteristics:
When brands embrace this modern framework, TV advertising often outperforms every other channel.
Drowning in data while thirsting for insights—that's the state of CTV measurement without AI. The platforms generating 50+ metrics per campaign create noise. Smart advertisers in 2025 focus on four AI-enhanced metric categories that actually predict and improve outcomes.
Completion rates tell a partial story. A 90%+ completion rate means nothing if viewers left the room, muted the TV, or scrolled their phones throughout. Modern viewership measurement combines multiple signals to understand true engagement.
Completion rate plus context tracks not just whether ads played through, but how they played. AI models analyze patterns like:
A high completion rate with high interaction signals indicates engaged viewing. The same rate with no device activity suggests background play. The difference changes everything about campaign value.
Attention score represents the breakthrough metric for 2025. Instead of binary "viewed/not viewed" classifications, attention scoring uses AI to rate engagement quality on a 0-100 scale. The models ingest dozens of signals:
Research from Adelaide shows that ads scoring above 65 on attention metrics drive 2.5x higher brand recall than those below 40, even with identical completion rates. The granularity enables optimization strategies impossible with basic viewability: shift budgets to high-attention inventory, adjust frequency caps based on cumulative attention rather than impressions, and sequence messages based on previous attention levels.
Traditional brand lift studies take weeks and cost thousands. AI-powered brand lift measurement delivers continuous insights by combining survey data with behavioral proxies.
Real-time brand recall estimation uses machine learning models trained on thousands of historical campaigns. The models identify behavioral patterns that correlate with survey-measured recall:
Instead of waiting for post-campaign surveys, brands see predicted recall scores update daily. When recall projections dip, campaigns adjust creative rotation or frequency strategies immediately.
Purchase intent signals move beyond asking viewers about future behavior to observing actual behavioral changes. AI models track:
The combination of stated intent (from targeted survey panels) and revealed intent (from behavioral data) creates purchase intent scores more predictive of actual sales than surveys alone.
Reach means nothing if you're reaching the same people repeatedly while missing your target audience entirely. AI-enhanced audience metrics reveal who you're actually reaching and whether it matters.
Incremental reach analysis uses machine learning to identify truly new audiences versus those already reached through other channels. The models match:
This analysis answers critical questions: What percentage of CTV impressions reach people unexposed to your social or display campaigns? Where does incremental reach justify premium CTV CPMs versus extending existing channels?
Audience quality scoring goes beyond demographics to predict audience value. AI models analyze hundreds of signals to score each impression's potential impact:
A million impressions reaching high-quality audiences outperforms 10 million reaching unlikely buyers. Quality scoring enables strategies like bidding more for high-value audiences while maintaining efficiency targets.
Cross-platform duplication rates reveal wasted spend from hitting the same households across multiple streaming services. AI-powered identity resolution tracks:
Understanding duplication enables smarter budget allocation—why pay for the 10th impression on Paramount+ when those dollars could find new audiences on Tubi?
Creative testing no longer requires expensive focus groups or waiting for campaign completion. AI analyzes creative elements in real-time, comparing performance to category benchmarks.
Dynamic creative scoring uses computer vision and natural language processing to deconstruct ads into elements: key messages, visual pacing, color schemes, talent demographics, music tempo, and emotional tone. The system then correlates element combinations with performance outcomes across thousands of historical campaigns.
Before launch, AI predicts which creative versions will resonate. During flights, it identifies which elements drive performance. A 15-second ad with fast pacing and upbeat music might score 85 for awareness but 45 for consideration. The same brand message with slower pacing and emotional music reverses those scores.
Category-specific benchmarks ensure relevant comparisons. AI models maintain separate benchmarks for:
These benchmarks answer crucial questions: Is 2.5% CTR good for a luxury auto campaign in Q4? Should a meal kit service expect higher or lower brand recall than the CPG category average?
Creative fatigue prediction identifies when ads lose effectiveness before performance crashes. AI models track:
Instead of running creative until metrics crater, brands proactively rotate messages based on fatigue predictions. This maintains performance while extending creative lifespan.
The real power emerges when these metrics work together. AI doesn't just track individual KPIs but identifies metric relationships that predict success. High attention scores plus strong audience quality typically predict brand lift. Creative effectiveness combined with optimal frequency drives purchase intent.
Smart platforms in 2025 automatically act on those insights: shifting budgets to high-attention inventory, rotating creative before fatigue sets in, and capping frequency when incremental reach disappears. The metrics become self-optimizing systems rather than static reports.
💡 So, how to track TV advertising? Full-funnel tracking means real-time analytics dashboards, third-party verification for delivery and viewability, and brand lift studies to gauge real impact. The latest solutions bring together linear TV, CTV, and digital video metrics for a unified, crystal-clear view of campaign performance.
The agencies thriving in 2025 share one trait: they stopped asking comfortable questions and started asking uncomfortable ones. While competitors debate CPMs and completion rates, winning agencies interrogate the metrics that make clients renew contracts and increase budgets.
{{how-to-5="/tables"}}
Most agencies still optimize campaigns toward media metrics that sound impressive in quarterly reviews but mean nothing to CMOs facing board scrutiny. Your client doesn't care about a 2.5% CTR. They care about same-store sales growth, customer acquisition costs that make unit economics work, and market share gains that justify marketing spend.
Start every campaign by asking: What business outcome keeps my client up at night? Then work backward to identify which media metrics actually predict that outcome. A retailer driving foot traffic needs different KPIs than a D2C brand maximizing LTV/CAC ratios. Yet most agencies apply the same templated metrics regardless of business model.
The hard question: Can you draw a straight line from your media optimizations to your client's P&L? If not, you're optimizing the wrong things.
Clients fund TV campaigns from budgets that could go to search, social, or retail media. They need proof that TV delivers superior returns, not just different metrics in another dashboard. Yet most agencies still treat TV as an isolated channel with its own success criteria.
True cross-channel clarity requires uncomfortable standardization. How does a completed CTV view compare to a Facebook video view? What's the relative value of TV-driven brand searches versus SEM clicks? Which channel deserves credit when TV exposure drives social engagement that leads to purchase?
Agencies avoiding these questions hide behind platform-specific metrics. Agencies winning new business answer them with unified measurement frameworks that show relative performance across every dollar spent.
After every campaign, try explaining what worked and why to someone outside advertising. Not "we achieved 89% completion rate with a 76 attention score." Real explanations like "Your funny ads worked because they ran during comedy shows to people who actually buy your product, driving 23% more store visits per dollar than last quarter."
If you can't explain performance in business terms, you don't understand it. Machine learning models that surface correlations without causation create false confidence. True insight requires knowing why certain inventory, audiences, and creative combinations outperformed.
This matters because clients increasingly demand strategic rationale. They want to understand the mechanics of success to inform broader business decisions. Agencies that explain the "why" become strategic partners. Those that only report the "what" become vendors.
Here's the question that separates modern agencies from those stuck in 2015: Are you still evaluating media quality based on CPMs? Price matters, but optimizing for cheap impressions is like buying discount sushi. The savings disappear when you account for waste, fraud, and zero business impact.
Performance-based evaluation changes every conversation. Instead of celebrating negotiated CPM reductions, you discuss cost per quality attention minute. Rather than bragging about reach efficiency, you analyze cost per incremental customer acquired. The shift from cost-based to value-based thinking unlocks budget increases because clients see returns.
The best agencies in 2025 grade themselves on four dimensions that clients actually value:
Notice what's missing: impressions delivered, CPMs achieved, completion rates hit. Those operational metrics matter for optimization but not for strategic evaluation.
In 2025, TV advertising rewards agencies willing to confront complexity and chase real results. The technology is here, the data is here, and the AI is ready to reveal what others miss. What’s rare is the agency brave enough to ask tougher questions, push past vanity metrics, and demand transparency from every platform. The top agencies know TV works when it’s measured ruthlessly and optimized for business outcomes.
If you’re ready to prove what really drives growth, reach out at daniel.friscia@aidigital.com or LinkedIn and let’s talk.
.svg)
.svg)
Inefficiency
Description
Use case
Description of use case
Examples of companies using AI
Ease of implementation
Impact
Audience segmentation and insights
Identify and categorize audience groups based on behaviors, preferences, and characteristics
Automated ad campaigns
Automate ad creation, placement, and optimization across various platforms
Brand sentiment tracking
Monitor and analyze public opinion about a brand across multiple channels in real time
Campaign strategy optimization
Analyze data to predict optimal campaign approaches, channels, and timing
Content strategy
Generate content ideas, predict performance, and optimize distribution strategies
Personalization strategy development
Create tailored messaging and experiences for consumers at scale
