Settings & Node Types

Settings Reference

All settings can be included when creating or updating a split test. Only name, mode, and variants are required for creation; everything else is optional with sensible defaults.

Core

Parameter	Type	Default	Description
`name`	string	required	Human-readable test name
`status`	string	`running`	`running` or `paused`
`start_date`	datetime	now	ISO 8601 datetime
`end_date`	datetime	null	ISO 8601 datetime

Duration & Auto-Winner

Parameter	Type	Default	Description
`duration`	integer	null	Test duration in days
`auto_select_winner`	boolean	false	Automatically declare a winner
`winner_selection_criteria`	array	null	Metrics to evaluate, e.g. `["aov", "add_to_carts_pct"]`
`winner_selection_confidence`	integer	null	Confidence threshold (1–100)
`sample_size_limit`	integer	null	Stop after this many total views
`baseline`	string	null	Node code of the baseline variant

Statistical Configuration

Parameter	Type	Default	Description
`significance_level`	decimal	0.05	Alpha (0.001–0.5)
`statistical_power`	decimal	0.8	1 - beta (0.1–0.99)
`minimum_detectable_effect`	decimal	5	MDE in percent (0.1–100)
`multiple_testing_correction`	string	`bonferroni`	`none`, `bonferroni`, `holm`, `fdr`
`statistical_method`	string	`frequentist`	`frequentist` or `bayesian`
`primary_metric`	string	`conversion_rate`	Main metric to optimize
`practical_significance_threshold`	decimal	2	Minimum meaningful improvement (%)
`require_practical_significance`	boolean	true	Require practical significance
`min_sample_size_per_variant`	integer	2000	Minimum views before declaring winner
`max_sample_size_per_variant`	integer	null	Stop variant at this many views
`min_duration_days`	integer	null	Minimum test duration
`max_duration_days`	integer	null	Auto-stop after this many days
`enable_sequential_testing`	boolean	false	Allow peeking with alpha spending
`alpha_spending_function`	string	null	e.g. `obrien-fleming`
`guardrail_metrics`	array	null	Metrics that must not degrade
`enable_automated_winner_selection`	boolean	false	Auto-declare when thresholds met
`automated_decision_threshold`	decimal	0.95	Probability threshold for auto-declaration

Bayesian Settings

Used when statistical_method is bayesian.

Parameter	Type	Default	Description
`bayesian_prior_mean`	decimal	null	Prior mean
`bayesian_prior_variance`	decimal	null	Prior variance
`bayesian_credible_interval`	decimal	0.95	Credible interval (0.5–0.99)

Advanced

Parameter	Type	Default	Description
`enable_variance_reduction`	boolean	false	CUPED-style variance reduction
`covariate_metrics`	string	null	Comma-separated covariate metric names
`enable_segmentation_analysis`	boolean	false	Segment results by dimensions
`segmentation_dimensions`	array	null	e.g. `["device", "country"]`
`statistical_notes`	string	null	Free-text notes

Custom Goals

Parameter	Type	Description
`goals`	array	Array of goal objects
`goals.*.name`	string	Display name (e.g. “Purchase”)
`goals.*.value`	string	Event identifier matching a page interaction (e.g. `purchase_complete`). Alphanumeric, underscores, and dashes only.

Full Settings Example

{
  "mode": "page_top",
  "page_id": 456,
  "name": "Homepage hero test",
  "variants": [
    { "name": "Control", "weight": 50 },
    { "name": "Variant B", "weight": 50, "page_id": 789 }
  ],
  "duration": 14,
  "auto_select_winner": true,
  "winner_selection_confidence": 95,
  "sample_size_limit": 10000,
  "significance_level": 0.05,
  "statistical_power": 0.8,
  "minimum_detectable_effect": 5,
  "statistical_method": "frequentist",
  "primary_metric": "conversion_rate",
  "min_sample_size_per_variant": 2000,
  "goals": [
    { "name": "Add to Cart", "value": "add_to_cart" },
    { "name": "Purchase", "value": "purchase_complete" }
  ]
}

Allowed Node Types

For Mode 2 variant chains, each node in the nodes array must have a type from this whitelist:

Type	Data Fields	Inputs	Outputs	Constraint
`page`	`value`: page ID (integer)	1	1	Funnel context only
`page_variant`	`value`: page ID (string), `load_page_events`: boolean (optional)	1	1	—
`url_redirect`	`value`: URL (string)	1	0	Terminal node
`product_check`	(none)	1	2	—
`product_check_product`	`value`: product code (string)	1	1	Must follow `product_check`
`product_check_all`	(none)	1	1	Must follow `product_check`
`script_rule`	`script_rule`: JS expression (string)	1	2	—
`set_variable`	`variables`: key-value object	1	1	—
`add_tag`	`value`: tag name (string)	1	1	—

Example: Node chain with product check

{
  "variants": [
    {
      "name": "Product Path",
      "weight": 50,
      "nodes": [
        { "type": "product_check" },
        { "type": "product_check_product", "value": "premium-bundle" },
        { "type": "page", "value": 500 }
      ],
      "next_node_code": "abc123def456gh78"
    }
  ]
}

Getting Started

Endpoints

Settings Reference

Core

Duration & Auto-Winner

Statistical Configuration

Bayesian Settings

Advanced

Custom Goals

Full Settings Example

Allowed Node Types

Example: Node chain with product check

Getting Started

Endpoints

​Settings Reference

​Core

​Duration & Auto-Winner

​Statistical Configuration

​Bayesian Settings

​Advanced

​Custom Goals

​Full Settings Example

​Allowed Node Types

​Example: Node chain with product check

Settings Reference

Core

Duration & Auto-Winner

Statistical Configuration

Bayesian Settings

Advanced

Custom Goals

Full Settings Example

Allowed Node Types

Example: Node chain with product check