Everything you need to test content that drives revenue

Test individual content blocks (hero, subject line, CTA) independently within the same message, without creating separate versions.

Test multiple slots simultaneously in a single campaign. Each slot is analysed independently with automatic interaction detection.

Test up to 5 creative approaches per content position. Starter supports 3 variants, Growth and Scale support 5.

Snippets persist across campaigns and channels. Reuse top performers and build cumulative performance data over time.

Define which platform events count as conversions. Choose from any event your platform tracks, not just the defaults.

Reserve up to 20% of the audience as a holdout to measure the total value of having content vs not having it.

Every slot gets a clear outcome: Winner Found, No Meaningful Difference, More Data Needed, or Guardrail Violation.

Know the exact probability each variant is the true winner, not just whether the result is "statistically significant."

Verdicts include incremental revenue with confidence ranges. "Variant B generated an additional £12,400 in revenue."

Automatically detects when variants perform within a negligible range, so you can stop testing and move on.

Watch the leading variant's probability climb toward the threshold in a live chart that updates every 60 seconds.

Compound two-stage model that analyses conversion probability and order value separately, handling revenue skew correctly.

Default for first campaigns. Every variant gets the same traffic, ensuring a clean initial test.

Set your own traffic percentages per variant when you want direct control over the distribution.

Automatically sends more traffic to better-performing variants while still exploring all options. Uses historical performance data across campaigns.

The control variant always gets at least 10% of traffic under Thompson Sampling, ensuring a clean baseline.

New variants added to an existing snippet get at least 20% of traffic on their first campaign for fair evaluation.

Segment-aware allocation that learns which variants work best for each audience segment and allocates accordingly.

Pre-test planning tool that estimates sample size requirements, minimum detectable effect, and test duration based on your audience and channel.

Analysis of variant content attributes (tone, length, structure) correlated with performance to identify what makes content win.

Per-segment variant breakdown with auto-detected audience properties. See which variants win for different customer groups.

Configure analysis parameters including decision thresholds, ROPE width, attribution windows, and prior settings for your workspace.

The statistical engine learns from your workspace history, using past results to inform priors for new campaigns and reach verdicts faster.

Persistent no-contact holdout group for measuring the incremental value of having content vs a baseline of zero exposure.

Monitor variant performance at 7, 14, 30, 60, and 90 days post-assignment. Reveal whether short-term winners hold up over time.

Owner, Admin, and Member roles with appropriate permissions. 3 to unlimited team members depending on tier, managed at the organization level.

Single sign-on support available on all tiers for secure, centralized authentication.

Profile properties, event payloads, and API credentials encrypted with Fernet symmetric encryption per workspace.

All data transmitted over TLS. API credentials never stored in plaintext.

Only stores data required for testing. Profile IDs stored unencrypted for lookups; everything else encrypted.

Automatic validation of assignment completeness, zero-event variants, minimum sample thresholds, and sample ratio mismatch (SRM).

Export campaign results for offline analysis or reporting in your own tools.