Now let’s see how the A/B for this scenario will be done: Determining a null & alternate hypothesisĪ Null hypothesis is when the two versions being compared have no differences and are precisely the same. For this, the BlueStacks app player has made some changes in the code that should increase the conversion between stage 0 and stage 1 in version d.e.f. Taking the number of users that did not try to install the Bluestacks app player after downloading as an exception, we focus on solving this problem for the users that tried installing but faced some issues. ![]() the install of the version was low, which meant that although there were a lot of people that downloaded BlueStacks app player, the majority of them were not able to install it due to some issues. ![]() For this version, it was observed that the number of users that reached from stage 0, i.e. Let’s say the Bluestacks app player has a version a.b.c that is being provided to the users from the website. To understand this concept better, we’ll establish a base example which we will reference in the blog at different points: Stage 0 is the number with which every other stage is compared. These stages are stage 0, stage 1 and so on. In part 2 of the blog, we will look at how we get to the point in the A/B tests where the statistical significance is measured and calculated.Īs we discussed in part 1 of this blog, BlueStacks compares the performance of two different versions based on the number of users that reached certain stages defined right from the download of the said version until the user installs an app on the version. The higher the possibility that the improvement is due to the modifications done in the code between the control and variation, the greater will be the statistical significance value and the fewer chances that the improvement is observed due to random changes. To brush up, it is the likelihood that the improvement of the different stages is not a random result. ![]() In part 1 of this blog, we shared some statistical significance insights.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |