Conversations (2)
Looking at the time-series smoothing, the trailing three data points appear less reliable because the smoothing window shrinks as you approach the present. This creates the false impression of converging estimates. Should we explicitly exclude the last few points from trend analysis, or should we apply a different smoothing method at the boundaries?
Looks good.
Didn't account for the holiday.
What's the methodology for "active"?
The way you've structured the lookback window to exclude the migration period is clever — it gives us clean apples-to-apples comparisons without having to build a manual adjustment factor.
What is the refresh cadence on this?
Thanks for catching that détail.
The units are missing from the colour scale.
Kenya vs Kenya: is retention comparable?
do you know whether 2006 was restated?
Units?
Is the 2018 gap a currency effect, or is the series genuinely missing?
Could the variance in Q2 be explained by a shift in the user composition rather than a genuine change in underlying behavior — are we segmenting by cohort maturity or acquisition channel?
Good catch.
Makes sense, thanks.
Not quite — the 2019 figure is a source revision.
What happened to the US around 2016?
Fair enough.
Do we have this in the Q4 pack yet?
Why does Indonesia dip in 2015?
I do not follow — is population going up or down here?
That is an outlier, not a real move.