Setting sampling to 20% in quick calculation mode does not mean that only 20% of users' data is actually displayed. It means that the values are extrapolated based on 20% of users. Especially for small values, this naturally leads to a significant difference in the evaluation between sampling (e.g. 20%) and raw data (100%).
An extreme example: Assume your shop had only 5 visitors a day. One of the users buys a product, the 4 others leave the website after short browsing. In fact, 20% of your users have bought. However, if the 20% sampling is used, only one of the users is used as calculation basis, so it could be shown that 100% of the users have bought or that 0% of the users have bought, depending on which of the users forms the calculation basis.
For low values (below 4 digits), sampling should therefore not be used in the evaluations.
Comments
0 comments
Article is closed for comments.