When we are about to apply any apparatus from probability theory, we must perform a very important test – we must check that our data represents a single process. Only when this check is done, we can call values “random variable” and apply statistics to it.
This need is usually overlooked. And, eventually, models became imprecise or BI shows us wrong results, etc.
Let’s have a look at a simple example: one of the pages on your site was under attack by the bots. If you do not separate bot traffic from regular activity, you will count bots as users and come to a conclusion that your recent ad campaign was very effective…
In the very heart of our product, we have math helping you to distinguish information of different nature and then process it separately. Without it, you cannot be sure that you are comparing apples to apples…
