Closed friedger closed 4 years ago
Furthermore, a growth of 600% weighs nearly as much as a loss of 100% (-0.12 vs -0.16)
Thanks @friedger PBC team discussed and we think this is a good change. It kind of depends on the larger decisions around how we use Awario scores (if at all). @cuevasm will follow up on this larger topic and revisit this shortly.
We will consider this after finalizing awario scoring method in #135. Moving to backlog until then.
Superseded by #135 (blended awareness)
The blended awareness still suggest this behaviour: For example in November (audit data)
My only hesitation with this is introducing a further delay in getting a full Awario score. @hstove can you check this math and let me know if there are other ways of solving what Friedger is raising please?
Going to continue this discussion on #204 which is a stronger proposal for solving the same issue.
What is the problem you are seeing? Please describe. Timestack lost 100% of reach (58.000 to 0), resulting in a Growth Z of -0.16 Utilo gained 130.000% of reach (29 to 38.000), resulting in a Growth Z of 7.1
How is this problem misaligned with goals of app mining? This score method suggests that it is good to promote the app only every other month. This is against the goal of growing the ecosystem.
What is the explicit recommendation you’re looking to propose? Use LOG(reach(month -1) - reach(month))