How to Measure CTV Ad Performance: The Complete Guide

CTV measurement is different from everything else

Connected TV advertising hit $33 billion in U.S. spend in 2025, growing 20% year-over-year. But the measurement infrastructure hasn't kept pace. Unlike display or search advertising, CTV has no clicks to track, no cookies to store, and no direct path from impression to conversion.

A viewer sees your ad on their Roku while eating dinner. Three days later, they search your brand on their phone and buy. Connecting those events requires a fundamentally different measurement approach than what most agencies are built for.

Here's the complete framework for measuring CTV ad performance in 2026.

The metrics that matter for CTV

CTV metrics fall into three tiers based on how directly they connect to business outcomes.

Tier 1: Delivery metrics (baseline)

These confirm your ads actually ran as planned:

• Impressions. Total ad views delivered. CTV impressions are generally more reliable than display because most CTV environments don't have bots or hidden placements. Invalid traffic rates on CTV are under 2%, compared to 10-15% for desktop display.
• Completion rate. Percentage of viewers who watched the entire ad. CTV completion rates average 95%+ because most CTV ads are non-skippable. This is dramatically higher than YouTube (30-40%) or social video (15-25%).
• Reach and frequency. Unique households reached and average times each household was exposed. Over-frequency (showing the same ad too many times) is the biggest CTV waste driver. Research from Innovid shows diminishing returns above 7-10 exposures per household per month.

Tier 2: Engagement metrics (signal)

These indicate whether the ads are influencing behavior:

• Website visit lift. Measure the increase in direct and organic website visits during and after CTV campaigns. Use a baseline period (4 weeks pre-campaign) and compare. A well-targeted CTV campaign typically produces 15-30% lift in site traffic.
• Search lift. Track increases in branded search volume using Google Trends or Search Console data. CTV consistently drives branded search -- a 2024 TVision study found that CTV ad exposure increased branded search likelihood by 25%.
• QR code scans or vanity URL visits. Adding a QR code or unique URL to your CTV creative provides a direct response signal. Scan rates are low (0.5-2%) but they provide a clean, attributable data point.

Tier 3: Outcome metrics (proof)

These connect CTV exposure to actual business results:

• Conversion lift. The gold standard. Run a controlled experiment comparing exposed vs. unexposed households to measure incremental conversions caused by CTV. Most major CTV platforms (Roku, Hulu, Samsung Ads) offer conversion lift studies.
• Incremental ROAS. Revenue attributable to CTV divided by CTV spend, measured via lift study or MMM. Typical CTV incremental ROAS ranges from 1.5x to 4x depending on category, creative quality, and targeting.
• Cost per incremental conversion. Total CTV spend divided by the number of incremental conversions (from lift study). This is the cleanest efficiency metric for CTV.

Cross-device attribution for CTV

The core challenge: connecting a TV screen impression to a phone or laptop conversion. Three approaches exist:

IP-based matching. Match the IP address of the CTV device to other devices on the same household network. When a viewer's phone later visits your site from the same IP, the visit is attributed to the CTV exposure. Accuracy: moderate. IP addresses change, shared networks create false matches, and VPNs break the connection entirely. But it works well enough at scale for directional measurement.

Device graph matching. Companies like LiveRamp, TransUnion, and The Trade Desk maintain probabilistic device graphs that link CTV devices to household members' phones, tablets, and laptops. Accuracy: higher than IP matching for identified households, but coverage varies. Expect 40-60% match rates.

Deterministic matching. Platforms with logged-in users across devices (Roku, Amazon, Google/YouTube) can deterministically match CTV impressions to conversions on the same account. Accuracy: highest, but limited to within each platform's ecosystem.

Best practice: Use deterministic matching where available (platform-native attribution), supplement with device graph matching for cross-platform campaigns, and validate everything with lift studies.

Setting up CTV measurement

Step-by-step implementation:

Step 1: Establish baselines. Before launching CTV, measure your baseline website traffic, branded search volume, and conversion rates for 4-8 weeks. You need clean baselines to measure lift against.

Step 2: Implement server-side tracking. CTV attribution depends on matching ad exposures to web/app conversions. Server-side tracking captures conversions that browser-based tracking misses, giving CTV credit for conversions that would otherwise go unattributed.

Step 3: Set up exposure logs. Work with your CTV platform or DSP to receive impression-level logs including device ID, timestamp, creative ID, and household IP. These logs are the raw material for attribution.

Step 4: Configure cross-device matching. Connect CTV exposure logs to your attribution system via IP matching or a device graph partner. Map CTV exposures to website visits and conversions within a 7-14 day attribution window.

Step 5: Run a lift study. In your first quarter of CTV, run a conversion lift study to establish a true incrementality baseline. This calibrates your ongoing attribution and gives you a defensible ROI number.

Common measurement mistakes

Using last-touch attribution for CTV. CTV almost never gets last-touch credit because the conversion happens on a different device hours or days later. If your reporting is last-touch, CTV will always look like it's delivering zero conversions. This is a measurement failure, not a performance failure.

Ignoring frequency caps. Without frequency management, CTV campaigns over-saturate high-value audiences. The incremental value of impressions 11-20 is near zero for most brands. Set household frequency caps of 3-5 per week.

Comparing CTV CPMs to display CPMs. CTV CPMs ($20-$40) are 5-10x higher than display CPMs ($3-$6). But CTV completion rates are 3x higher, attention quality is far superior, and fraud is near zero. Cost per completed, viewable, fraud-free impression is often comparable.

Expecting direct response results. CTV is primarily a reach and awareness channel. Measuring it purely on immediate conversion metrics undervalues it. Include brand metrics (search lift, site visit lift) alongside conversion metrics.

FAQ

How long should I wait to measure CTV results?

Allow a minimum of 4 weeks for a CTV campaign to produce measurable lift, and 8-12 weeks for stable performance data. CTV has longer carryover effects than digital display -- the adstock typically extends 2-4 weeks. Measuring too early will undercount the true impact.

What's a good CTV conversion lift benchmark?

Conversion lift varies widely by category and creative quality. E-commerce brands typically see 8-15% lift in conversions among exposed households. B2B and high-consideration categories see 3-8% lift. If your lift study shows less than 3%, the targeting or creative needs improvement before scaling spend.

Can I measure CTV without a lift study?

Yes, but with less confidence. IP-matched attribution and device graph matching provide directional CTV performance data without a formal lift study. Combine these with search lift analysis and baseline comparisons for a reasonable picture. But lift studies remain the gold standard -- run at least one per year to calibrate your ongoing measurement.

---

Go Funnel uses server-side tracking and multi-touch attribution to show you which ads actually drive revenue. Book a call to see your real numbers.