I'm not sure what you're trying to say. The algorithms adjust for bots, spam, cross-platform duplication, etc. 900M - 1B is defined as the Monthly Active Users figure.

Naturally, there are areas where we know the algorithms are not translating the inputs to real users with 100% fidelity, but we know that the discount is relatively minor, not nearly as substantial as youre suggesting.

Multiple counter-parties had their teams diligence our user figures and associated algorithms and found them to be generally accurate representations.

Unless you're using a different definition of "active, monthly user" that deviates from the industry norms?