Skip to content

Help · Audience and analytics

A/B experiments

Test whether a different destination order converts better, with honest denominators and a card that refuses to call winners early.

5 min · matches the current product

An experiment answers one question: does putting a chosen destination first change what fans click? Visitors who arrive without a campaign tag are split evenly and deterministically — the same fan sees the same variant on every visit — and campaign-tagged visitors are excluded, because they belong to their campaign’s numbers.

  1. Start it from the link page

    On the link’s overview (Artist Pro), the experiment card asks for a name and one thing: which destination variant B should pin to the top. Variant A is the page exactly as you built it. Leaving the pin empty runs an A/A test — useful once, to see that a null result reads as a coin flip.

  2. Let it run

    The card shows visits, clicks, and click-through per variant, read from the same rollups as every other report — the experiment cannot disagree with the campaign page. It will not name a leader until both variants have at least 100 visits, because a 12-visit "winner" is noise wearing a medal.

  3. Stop it and act

    Stop the experiment when the comparison is meaningful. The result stays readable after stopping. If variant B won, reorder the destinations for real; the experiment made the case, the editor makes the change.