Analytics shows not plausible data

[Juraj M.]

For several of my extensions, when I check installs and uninstalls, the data doesn't make sense.

For example:

This graph shows that in the last 30 days I have 421 installs from US and 730 uninstalls from US.

That cannot be right.

Not to mention China which seems to be loosing a lot of users but not gaining?

And the Hong Kong is gaining and not loosing?

And I can see similar inconsistencies in some of my other extensions too.

For example:

Can YOU see a similar pattern in your stats?

[Mythical 5th]

Here's one of mine that's been puzzling me for a while.

Over the last 180 days, installs and uninstalls have been low and fairly regular. Just setting the scene, the fireworks come later...

Page views, also low and uneventful.

Here's where it gets weird.

Somehow, a ~40x increase in impressions across the store for several weeks had no impact on page views.

Prior to the algorithm granting its favour with that bunch of inconsequential impressions, and despite there being no change in the pattern of installs, the userbase nearly tripled in about 6 weeks.

Net installs for the 180-day period were 443, user growth was 1373 at peak.

Here's the 5-year chart to highlight the extension's dismal growth story and the weirdness of recent events.

It's highly unlikely that the jump in weekly users is due to lots of old users returning, but the stats show they didn't come from anywhere else.

I'd celebrate the victory, but I don't know if it's real. There were no new reviews from all those extra users.

Back to 180-day charts.

Installs by OS is also odd, with a large mismatch of installs and uninstalls like that observed by Juraj.

The last 30 days:

With those install stats and ~60% of users being new during the time period represented by that data, what are the odds that 60% of current users are on Linux, and so few on Chrome OS?