Connected TV—or CTV—is a cornerstone of modern, omnichannel marketing strategy, yet many teams still struggle to measure its true business impact. Fragmented data and platform walled gardens block leaders from connecting exposure to outcomes, proving incrementality, or measuring true brand lift—leaving organizations with siloed activity reports instead of decision-grade insight.
This guide shows how to track key CTV metrics, evaluate CTV ad performance across the full funnel, and connect media spend to measurable business growth.
Key highlights:
- CTV measurement tracks ad performance across internet-connected devices to quantify reach and incremental impact.
- Effective CTV ad measurement depends on tracking outcome-driven metrics—such as incremental conversions, cross-device attribution, brand lift, and iROAS—to prove true performance, eliminate waste, and optimize spend across the full funnel.
- Keen unifies fragmented streaming data through causal Bayesian modeling, enabling real-time “what-if” forecasting and budget optimization based on true incremental revenue lift.
What is connected TV measurement?
Connected TV measurement is the process of tracking and analyzing ad performance across internet-connected television devices, including smart TVs, streaming sticks, gaming consoles, and set-top boxes. This approach replaces panel-based sampling with device-level and household-level data to quantify reach, engagement, cross-device attribution, and incremental marketing impact, directly linking exposure to business outcomes.
Having a strong CTV measurement is important as adoption accelerates. According to Future Market Insights, U.S. demand for connected TVs will grow from $6.8 billion in 2026 to $25.6 billion by 2036, as more people cut the cord and switch to streaming. As this trend continues, outcome-focused measurement helps advertisers prove ROI, optimize spend, and turn large target audiences into valuable data for business growth.
How does CTV ad measurement work?
CTV ad measurement works by capturing device-level data from smart TVs, streaming platforms, and connected devices to link exposure to downstream consumer actions. Measurement platforms track ad delivery and viewership, assign unique identifiers to households, and connect those impressions to behaviors such as site visits, searches, or purchases through deterministic matching or probabilistic modeling.
To measure more than just ad delivery and understand business results, CTV measurement systems integrate three core data layers:
- Ad delivery data: Platforms provide real-time logs and video completion rates (VCR) to show which households saw an ad and how often.
- Identity resolution: Identity graphs use device signals, such as IP addresses and device IDs, to match devices to household profiles.
- Conversion data: Marketing attribution models connect CTV ad views to actions such as purchases by linking household viewing data to CRM systems, sales records, and website analytics.
11 CTV metrics you need to track
Track these 11 CTV metrics to quantify the incremental business impact of your connected TV campaign:
1. Deduplicated household reach and effective frequency
Deduplicated household reach shows how many unique homes see your campaign across all platforms. Effective frequency measures how many times each household should see your ad to drive brand recall.
- Why measure: These CTV metrics help you use your budget wisely and reach more people. Innovid reports that the average viewer sees the same CTV ad about seven times, while campaigns typically reach only 19.64% of eligible households. This high-frequency, low-reach imbalance drives overexposure to a small audience segment while failing to reach new viewers.
- How to measure: Combine device IDs into one household profile with identity resolution and cross-platform tools. Then set frequency limits so your budget reaches households that haven’t yet seen your ads.
2. Unique households exposed and co-viewing impact
Unique households exposed count how many different homes your campaign reaches. Co-viewing impact measures the number of people watching together on the same screen at the same time.
- Why measure: People often watch together, not alone. TVision reports co-viewing rates from 1.1 times for some streaming programs to 1.9 times for live sports. Ignoring these co-viewing patterns can lead to underestimating your audience and overstating cost per mille (CPM)—the cost you pay for every 1,000 impressions. Counting co-viewers reveals true audience size and shows that CTV delivers reach more efficiently than device-level reporting.
- How to measure: Connect exposure data to household identity graphs and use co-viewing multipliers from automated content recognition (ACR) data or trusted panel benchmarks.
3. Video completion rate (VCR) and quartile completion
VCR measures the percentage of viewers who watch an ad through to the end. Quartile completion tells you at which point—25%, 50%, or 75%—viewers tend to stop watching.
- Why measure: VCR and quartile completion show how well your ad holds viewers’ attention and whether the creative message connects. Since you usually can’t skip CTV ads, a high completion rate means your ad runs smoothly and reaches viewers. Quartile data helps you see where people lose interest, so you can adjust your ad’s opening or length to make sure your message gets across.
- How to measure: Deploy VAST tags to trigger tracking pixels at each quartile and aggregate data within your ad server. Break down the results by creative version and placement to see what keeps viewers engaged.
4. Viewable impressions and CTV viewability rate
Viewable impressions verify that an ad appeared on an active screen. The CTV viewability rate indicates the percentage of impressions that meet visibility standards.
- Why measure: This marketing KPI evaluates actual ad exposure and inventory quality. It identifies technical waste, such as background playback, app crashes, or ads running when the TV is off. Monitoring viewability protects your budget from invalid traffic and ensures you pay only for digital advertising that people can actually see.
- How to measure: Use third-party verification partners to capture device-level signals and confirm playback. Validate exposure against the Media Rating Council (MRC) standards to identify and flag non-viewable inventory.
5. Cost per thousand impressions (CPM)
CPM calculates the cost to deliver 1,000 ad impressions on streaming platforms and devices.
- Why measure: CPM helps you see how efficiently you spend on media and how your costs compare to others in the market. It also lets you compare different inventory sources and shows how factors such as targeting, seasonal demand, and auction competition affect your pricing.
- How to measure: Take your total media spend, divide it by the number of impressions delivered, and then multiply by 1,000.
6. Cost per completed view (CPCV)
CPCV measures the cost per full ad completion.
- Why measure: CPCV connects your spending to actual ad views. It shows where you waste money on ads people don’t finish and points out which platforms, placements, and creatives really get viewers’ attention. By focusing on completed views, you can spot where technical issues or tired creative weaken your message.
- How to measure: Take your total campaign spend and divide it by the number of fully completed views (100% quartile) from your ad server and platform logs. Break down the results by ad length and placement type to see the real cost of attention for each partner.
7. View-through conversions and post-exposure site visits
View-through conversions track digital actions, such as purchases or sign-ups, that viewers take after seeing a CTV ad. Post-exposure site visits measure the web traffic generated by those ad impressions.
- Why measure: CTV ads don’t have clickable links, so viewers often take action later on another device. These CTV metrics show how ads create demand and highlight the halo effect that click-based attribution can miss. They demonstrate that streaming campaigns drive measurable intent and activity beyond just brand awareness.
- How to measure: Connect household-level ad exposure to later actions by using cross-device identity resolution and IP matching. Set clear attribution windows to confirm that site visits or conversions happen after someone sees an ad.
8. Incremental conversions and cost per visit (CPV)
Incremental conversions identify actions—such as sales or sign-ups—that are directly caused by CTV exposure, excluding organic baseline activity. CPV measures the media spend needed to drive an exposed household to a digital or physical destination.
- Why measure: Incremental conversions and CPV distinguish true causal impact from correlation. They show whether CTV generates new demand or only captures conversions that would have happened regardless. Measuring lift demonstrates the real efficiency of your streaming investment and supports budget decisions based on actual growth.
- How to measure: Conduct controlled experiments, such as geo-holdout tests or randomized control trials, to compare conversion rates between exposed and unexposed audiences. Subtract baseline activity from the exposed group’s total, then divide the campaign cost by the resulting lift in actions.
9. Cross-device impact and attribution
Cross-device impact and attribution track the consumer journey from CTV ad exposure to actions on other household devices, such as mobile, desktop, or tablets.
- Why measure: Viewers rarely convert on the TV screen. Cross-device impact and attribution capture the halo effect and assisted impact that single-device tracking misses. It prevents undervaluing CTV by revealing how big-screen exposure triggers downstream search queries, site traffic, and conversions on personal devices.
- How to measure: Deploy household identity graphs to link CTV impressions to individual digital activity. Use deterministic or probabilistic matching to attribute multi-screen actions back to the original exposure within a defined window.
Download our playbook on cross-channel media measurement.
10. Brand lift metrics (awareness, recall, consideration)
Brand lift tracks changes in consumer perception, such as awareness, ad recall, and purchase intent, by comparing people who saw the ad with those who did not.
- Why measure: Standard metrics often miss the psychological effects of CTV. Brand lift shows how ads make your brand more memorable and influence future sales. It proves that media spending builds long-term value, not just immediate sales.
- How to measure: Use controlled surveys from third-party partners to compare answers from people who saw the ad with those who did not. Find the lift by looking at the percentage difference in brand favorability or intent between the two groups.
Read more: Performance marketing vs. brand marketing
11. Incremental ROAS (iROAS) for CTV
iROAS calculates the specific revenue return generated by your CTV investment, isolating it from baseline organic sales.
- Why measure: iROAS shows the real financial impact of CTV, not just a link. It reveals whether CTV generates new revenue or just shifts sales from other channels. Tracking this metric helps you make a stronger case for shifting budgets to streaming that truly grows your business.
- How to measure: Run matched-market experiments or randomized control trials. Compare revenue from households that saw your CTV ads with that of a control group of unexposed audiences. Subtract the control group’s revenue from the test group’s total, then divide that number by your total CTV media spend.
Download our marketing metrics ebook.
Key CTV measurement challenges
Tracking CTV performance only creates value if you can measure it accurately. Fragmented platforms, identity gaps, and attribution complexity distort ad effectiveness signals, undermining decision-making.
| CTV measurement challenges | The importance | The solution |
| Fragmented inventory and walled gardens | Closed ecosystems block data sharing, prevent unified marketing measurement, and distort cross-platform comparisons. | Third-party measurement layers combine data across ecosystems to offer a unified view. Use solutions such as Keen Marketing Measurement, which standardizes KPIs and presents an unbiased, normalized view of ad performance in a common financial framework. |
| Identity resolution (the cookie-less gap) | Cookie loss breaks the link between household exposure and cross-device actions, weakening attribution accuracy. | Household identity graphs link IP addresses and first-party data to build persistent profiles. Keen, for example, employs Bayesian marketing mix modeling (MMM) to unify signals and connect exposure to outcomes without depending on unstable platform IDs. |
| Accurate attribution (the halo effect) | Upper-funnel CTV performance shifts conversions to other channels and misallocates credit. | MMM, incrementality testing, and causal modeling measure CTV’s true contribution. Keen’s modeling engine isolates cross-channel lift and the halo effect beyond last-click attribution. |
| Co-viewing measurement | Shared viewing undercounts audience size and misprices reach value. | Marketing measurement platforms model co-viewing using panel benchmarks and automated content recognition (ACR) calibration to translate device-level impressions into people-level reach and to accurately value audiences. |
Streamline CTV measurement with Keen Decision System
CTV KPIs are valuable only when they connect to financial outcomes across the marketing mix.
With Keen’s MMM platform, you unify CTV data across channels, employing causal modeling and incrementality testing to isolate the true revenue and profit impact of CTV investment. Our marketing measurement solution estimates incremental lift, quantifies the halo effect on search and social, and deduplicates reach across fragmented inventory—turning fragmented performance reports into a unified decision framework. The result: predictive CTV measurement that maximizes marketing ROI.
Book a Keen demo to learn more.
