Skip to main content
All split test endpoints authenticate via EF-Access-Key with brandAccess middleware. Request bodies are nested under a split_test key.

List Split Tests

string
required
The brand/project ID
number
Results per page (1–100, default: 25)
string
Filter by status: running, paused, finalized, deleted
cURL

Get Split Test

string
required
The brand/project ID
string
required
Split test ID
cURL

Create Split Test

All body fields must be nested under the split_test key (e.g. split_test.name, split_test.config).
string
required
The brand/project ID
string
required
Human-readable name for the test
string
required
split_test (funnel/page variants) or component_split_test (component-level test)
object
required
Test configuration including items array. Each item has name, node_code, and weight.
string
running, paused, completed, deleted (default: running)
string
ISO 8601 date (default: today)
string
ISO 8601 date
string
Primary metric to optimise for (e.g. conversion_rate, revenue)
string
frequentist (default) or bayesian
number
Alpha level (0.001–0.5, default: 0.05)
number
Beta level (0.1–0.99, default: 0.8)
number
Minimum visitors per variant before declaring a winner (100–1,000,000)
boolean
Automatically select winner when confidence threshold is reached
number
Confidence % required for auto-winner selection (e.g. 95)
array
Conversion goal events. Each goal has name (string) and value (alphanumeric slug).
string
The control variant’s name or node code
cURL

Update Split Test

Same body structure as Create — all fields nested under split_test, all optional on update except as noted. To pause a running test: set split_test.status to paused.
cURL — pause a running test
cURL — extend end date

Delete Split Test

cURL

Get Chart Data

Retrieve time-series and comparison data for visualisations.
string
required
Split test ID
string
required
One of: conversion_trends, funnel_analysis, statistical_confidence, performance_comparison
string
Start date (ISO 8601)
string
End date (ISO 8601)
string
Timezone (e.g. America/New_York)
cURL

Declare Winner

Manually select the winning variant and finalize the test. Sets status to finalized.
string
required
The node_code of the winning variant
string
required
Display name of the winning variant
number
Statistical confidence at time of declaration (0–100)
number
Relative lift over baseline (e.g. 25.8 for 25.8% improvement)
string
Free-text note for the decision (max 500 chars)
cURL

Export Results

Queue a CSV/Excel export of all split test results. An email is sent when ready.
string
csv (default) or excel
boolean
Include p-values, confidence intervals, lift (default: true)
boolean
Include CI columns (default: false)
object
Optional {start, end} to restrict export date range
cURL

Get Node Codes

Return the variant names and node_code identifiers for a split test — useful for linking analytics filter calls.
cURL

Quick Access Stats

Counts by status — useful for dashboard badges.
cURL

Notes

Add freeform notes to a split test for team collaboration.

List notes

Add note

string
required
Note text

Delete note

cURL — add note

Using Split Test Data with Analytics

To filter analytics metrics by split test variant, pass the split_test_id and filter[node_code] to the analytics endpoints: