Conversions FAQ
- How is the winning version of the split test selected if the test was based on conversion while no conversion was generated?
- How is the winning version of the split test selected if the test was based on the number of conversions while the same number of conversions was generated by all versions?
- How is the winning version of the split test selected if the test was based on the value of conversions while the same value was generated by all the versions?
1) How is the winning version of the split test selected if the test was based on conversion while no conversion was generated?
The winning version will be chosen based on the fallback mechanism: clicks, then opens, and finally – if all the statistics are the same for all versions – randomly.
2) How is the winning version of the split test selected if the test was based on the number of conversions while the same number of conversions was generated by all versions?
The winning version will be chosen based on the fallback mechanism: conversion value, then clicks, then opens, and finally – if all the statistics are the same for all versions – randomly.
3) How is the winning version of the split test selected if the test was based on the value of conversions while the same value was generated by all the versions?
The winning version will be chosen based on the fallback mechanism: number of conversions, then clicks, then opens, and finally – if all the statistics are the same for all versions – randomly.