A/B Tests
Create A/B Test
Create a campaign A/B test
POST
Create A/B Test
Create a draft A/B test for a campaign in
For a sequence test, copy the effective values from the source test’s
A sequence conversion moves that step’s copy off the sequence node and onto the
test’s variants. The response echoes this as
draft or rejected status, or convert a sequence email node into an A/B test step. Provide exactly one of campaignId or automationNodeId. Variant A (the control) is copied from the current email. A campaign or sequence node can have only one A/B test.
Request
string
Campaign to attach the test to. Must be in
draft or rejected status.
Mutually exclusive with automationNodeId.string
Sequence
action_email node to convert into an action_ab_test step.
Mutually exclusive with campaignId. Requires at least one variants entry.boolean
Required as
true when automationNodeId belongs to an active sequence. This
acknowledges that the live sequence graph will change immediately.string
Test name. Defaults to
A/B Test for <campaign name>.number
Campaign-only share of the audience that receives test sends. Must be from 5
to 50. Defaults to 20. Sequence tests use
winnerThreshold instead.number
Campaign-only duration before a winner is selected. Must be from 15 to 1440
minutes. Defaults to 240.
string
Metric used to pick the winner:
open_rate, click_rate,
goal_conversion_rate, goal_value, or event_rate. Campaigns default to
open_rate. For sequence tests, an explicit value overrides the testType
default.Goal and event criteria. goal_conversion_rate picks the variant whose
recipients convert on winnerGoalId most often per delivered email.
goal_value picks the one that brings in the most goal value per delivered
email, so with default-revenue it is the variant that made the most money.
event_rate picks the one whose recipients most often fire winnerEventName
after receiving it. A goal conversion counts for a variant when it is
attributed to that email, meaning the recipient opened or clicked it within the
goal’s attribution window. For campaigns, default-revenue and goals of that
campaign also credit a recipient who received the email in that window without
opening it. If no variant converts, or the leaders tie, campaign
tests fall back to click rate when the test ends, and sequence tests keep
splitting contacts until one variant leads.string
Goal judged by
goal_conversion_rate or goal_value, and required for them.
Use the ID of an active goal that is company-wide or belongs to this campaign
or sequence, or default-revenue for built-in purchase revenue (SaaS
purchases and commerce orders, tracked in cents). Any other goal returns
400.string
Event judged by
event_rate, and required for it, for example
trial.converted.string
Sequence variant strategy:
subject or content. Subject defaults to
open_rate and content defaults to click_rate unless winnerCriteria is
explicit. Defaults to content. Sequence tests only.number
Number of sequence recipients in the test sample, from 10 to 1000. Defaults to
100. Sequence tests only. After the threshold, the variant with the highest
open or click rate wins, excluding bot opens and clicks. Recipients keep being
split across variants while no variant has engagement for
winnerCriteria or
the leaders are tied. The leader is rechecked every 5 minutes. If tracking for
winnerCriteria is off, select the winner manually.array
Extra variants beyond the control. Each item takes
subject (required),
previewText, and blocks. Total variants cannot exceed 5. Variants without
blocks reuse the current email body. Optional for campaign tests; sequence
conversions require at least one entry.settings object:
contentEditing: from then on the
step’s subject, preview text, and blocks are edited per variant with
Update Variant, not through the
sequence node endpoints, and a change meant for the whole step has to be applied
to every variant. The conversion response also returns the rewritten
nodeConfig.