Sign in

v12

latestOpenAPI 3.1.0AGPL-3.0raw.githubusercontent.com2026-08-0464108237.1 KB
Analytics

A/B test experiment results

Statistical significance results (control vs. variant) for a routing A/B test experiment, with guardrail evaluation.

get/analytics/experiment/{experiment_id}/results

Path parameters

experiment_idstring required

Query parameters

start_msinteger
end_msinteger
min_sample_sizeinteger
guardrail_threshold_ppnumber
evaluation_marginnumber

Headers

x-tenant-idstring required

Tenant to resolve. The shipped config files define only the "public" tenant. No fallback — omitting this fails with TE_03.

Response

A/B test experiment results

experiment_idstring
merchant_idstring
delta_ppnumber double

Variant minus control auth rate, in percentage points.

p_valuenumber double nullable
confidence_intervalnumber[] nullable

Lower and upper bound.

verdict'collecting_data' | 'not_significant' | 'variant_wins' | 'variant_loses' | 'guardrail_breached'
min_sample_sizeinteger
net_delta_bpsnumber double nullable
evaluation_marginnumber double

Example response

{
  "control": {
    "arm": "control"
  },
  "variant": {
    "arm": "control"
  }
}