On almost every other service capable of A/B testing, you can determine what percentage of users see version A vs. version B.
E.g. "Direct 20% of all visitors to page B and keep 80% on page A."
To do that, two options come to my mind which include:
- Recording past visits.
- "Rolling the dice" and splitting through probability calculation.
Both seem suboptimal to me and that's why I would love some more perspectives on routing traffic while keeping a ratio (include example code if you like - any language will do).
Anything helps, really.
Cheers!
Aucun commentaire:
Enregistrer un commentaire