User-Level Testing

Run randomized control trials (RCTs) on your own user base. Measure the true impact of CRM, direct mail, email, push notifications, and other addressable lifecycle channels.

Centralized Truth

All your tests, one database

Meta
Google
TikTok
Pinterest
Apple
TV
CTV
Netflix
OpenAI
X
Snapchat
LinkedIn
Bing
Spotify
YouTube
Direct Mail
SMS
Email
Notifications
Display
Video
Live Streaming
Influencer
Affiliate
Coupons
Partnerships
In-store
Events
Concerts
Movies
Radio
Discount
Q4 Promo
Holdout
Scale Test
Creative A/B
Retargeting
Prospecting
Holiday
Launch
Powered by Your Test Database

Store & Calibrate

Every user-level experiment becomes a data point in your central truth database, ready to calibrate your MMM.

Central Log

Keep a permanent record of every email subject line test, push notification holdout, and direct mail lift study.

Meta-Analysis

Look back across all your tests to find patterns. Does "Discount" work better than "Free Shipping" across all channels?

MMM Validation

Use your user-level holdout results to validate that your top-down MMM is correctly attributing credit to CRM channels.

Audience Splitting

Automatically split user universes into treatment and control groups with reproducible randomization and clear assignment logic.

Power Analysis

Calculate exactly how many users you need in each group to detect a meaningful lift before you start the test.

Holdout Groups

Maintain long-term holdout groups to measure the cumulative impact of your marketing efforts over time.

Addressable Channels

Built for the Channels You Can Actually Hold Out

User-level testing is strongest when you control the audience, the suppression logic, and the outcome measurement.

Email and CRM

Run holdouts on promotional sends, lifecycle journeys, and subject-line or offer tests while measuring revenue, conversion, and unsubscribe guardrails by assignment.

Direct Mail and List-Based Outreach

Use clean treatment and suppression files to measure incremental response, profit, and reactivation instead of relying on attributed conversions alone.

Push, SMS, and App Messaging

Measure whether messages create net-new behavior, accelerate an action that would have happened anyway, or simply add noise and fatigue.

Trustworthy Experiment Workflow

The strongest user-level tests are not just randomized. They are logged, guarded, and decision-ready.

Power and Duration

Estimate sample size before launch so the team knows whether the test is capable of detecting a decision-relevant effect.

Guardrails and Holdout Integrity

Track leakage, unsubscribes, complaint rates, and downstream business guardrails so local wins do not create hidden harm.

Experiment Log and Reuse

Save the result, the setup, and the context so user-level learnings become calibration inputs and reusable knowledge instead of disappearing into a slide deck.