A/B Experiment Example

(Also known as) A/A/B Test

  • 50% of the traffic (variation under test) is split into 2 identifable and individual control groups: A1 and A2 (which use current same configuration)
  • Other 50% of the traffic use an Experiment group: B
  • There must be no difference in targetted same inventory type: position, availability etc.

Click Ad on right (below if viewing on mobile device) to trigger count