All case studies
CTV & Audience Analytics

Transforming CTV Audience Measurement Through Independent Impression Tracking

A leading US audience analytics provider replaced DSP log dependency with pixel-based ground truth — capturing more signals, higher precision, and full data ownership.

+10.6%
More raw IP signals vs. BAU baseline
0.998
Pearson correlation — near-perfect reliability
91.2%
Jaccard similarity in IP overlap
10.6%
Household match rate vs. 8.3% BAU

The challenge

In the Connected TV ecosystem, independent data validation is a strategic necessity. As advertising budgets migrate toward streaming, the industry's reliance on Demand-Side Platform (DSP) ad logs has created a transparency vacuum — measurement providers operate as observers of walled-garden data rather than independent auditors.

One of the largest audience analytics companies in the USA needed to break this dependency. Their existing workflow relied on FreeWheel's external log delivery, which constrained data control, processing speed, and the ability to independently validate campaign performance. The measured data was derived from the same entity responsible for delivery — a structural conflict.

No independent signals

No secondary, pixel-based tracking mechanism to validate DSP-provided data — unable to audit campaign performance independently.

DSP log dependency

Workflow historically reliant on FreeWheel's external log delivery, constraining data control and processing speed.

Creative-level granularity

Needed more precise data collection directly from the ad creative at the point of execution to improve household matching.

The solution

Ingest Labs deployed a pixel-based impression tracking solution rooted in composability, interoperability, and scalability — designed to bypass the limitations of monolithic, proprietary reporting interfaces.

A dual-tagging strategy employed both a JavaScript tag and a 1x1 image fallback to ensure total coverage across the fragmented CTV device landscape. By leveraging FreeWheel's macro-mapping capabilities, Ingest Labs captured critical metadata in real-time — auction IDs, campaign IDs, creative IDs, and impression IDs — directly from the ad creative at the point of execution.

Privacy and compliance were maintained through a rigorous protocol: rotating daily salt for PII hashing (IP and User Agent) before storage, plus IP truncation to /24 for IPv4 and /48 for IPv6 to ensure household-level analysis remained privacy-safe.

The architecture leveraged Kinesis, PubSub, and SQS for stream buffering with S3 storage — providing infinite horizontal scalability for CTV traffic spikes. Automated S3/SFTP delivery ensured daily CSV files were dropped consistently at 07:00 AM CT.

Results

Validated over a 24-day window (October 2–25, 2025) comparing Ingest Labs pixel-based tracking against the existing DSP log baseline.

Metric Ingest Labs BAU Baseline Insight
Total Impressions 1,237,785 1,247,714 99.2% volume consistency
Raw IP Counts 106,032 95,816 +10.6% Ingest Labs advantage
Verified Household IPs 17,822 16,577 +7.5% advantage for Ingest Labs
Strict Match HH Conv Rate 10.6% 8.3% Higher signal precision
MAID Efficiency 6.5% 4.8% Cleaner, more actionable identifiers

Statistical reliability

0.998
Pearson Correlation

Near-perfect coefficient — Ingest Labs captures the exact same temporal impression patterns as the baseline.

91.2%
Jaccard Similarity

High IP overlap confirms the additional 10.6% IPs captured are valid signals, not noise.

6.5%
MAID Efficiency

Pixel-based method filters out noisy or inactive identifiers for a cleaner dataset.

All primary KPIs met

Validated by the customer as accurate and superior to DSP logs for identity resolution.

Data capture accuracy

All scoped creative and campaign metadata captured accurately.

S3 delivery schedule

Daily CSV delivery maintained consistently at 07:00 AM CT.

Customer validation

Data certified as accurate with superior IP-based signal for identity resolution.

Independent, pixel-based tracking provides a more precise and voluminous signal for CTV measurement than traditional DSP logs. All achieved with pure composable architecture — without any new development work.

Get started

Ready to own your measurement data?

Independent, pixel-based tracking for CTV, web, and app — live in days.

Live in <24 hours No code changes SOC 2 compliant