Questions about disabled users and Chrome Store stats

[GrungeL]

Hi, I accidentally changed a subdomain in the manifest, and soon the infamous prompt of "the extension was disabled, accept the new permissions" was triggered.

Immediately I republished the previous version, trying to minimize the damage.

After two days, when the stats were updated, there was a huge spike in the uninstalls chart, but I also track the uninstalls by opening a /good-bye page with Google Analytics in it, and GA didn't report more uninstalls than usual for that day.

Does the spike reflect the users who clicked on "Error" and "Remove" (maybe that type of uninstall doesn't trigger the /good-bye URL?), those that didn't interact with and simply left it disabled, or a sum of both?

What happens if before they interact with "Error" I rollback to a version without the new permissions? Is it reenabled? Or the only users I managed to save were the ones that didn't open their computer at all that day and missed the permissions-demanding extension update? 

Also, is there any way to re-engage users that got their extension disabled? 

I really think that Google could help us developers a little more in this regard. It is very difficult to get users and with the slightest mistake we can lose almost all of them irreversibly.

Thanks

[Deco]

To answer the question, the CWS stats do not differentiate between those who click Remove or Error, only the extension state itself.

In a bit more context what this means is what users do is arbitrary, and that the simple statistic is collected from when it is registered on the board (once daily), which is why you seen a spike in the number of people who uninstalled the extension.

The Disabled state means that an extension is physically switched off from any form of operation except for receiving updates, these will still be downloaded and applied as normal, but as far as being able to "interact" with users who have it disabled, this isn't possible.

It isn't really something that the CWS team can assist with, as allowing extensions to interact in a disabled state, would be ripe open to abuse. In the future, take care to triple check everything that you are doing before submitting an update to avoid this arising in the future.

Thanks,

Deco

--
You received this message because you are subscribed to the Google Groups "Chromium Extensions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-extens...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-extensions/988b71b1-bb59-46b1-8ac1-461663c233c3n%40chromium.org.

[GrungeL]

Thanks, but rephrasing my questions:

• The spike in uninstalls is because:
  1. Users clicked in remove extension (and it didn't trigger the /good-bye URL because it was already disabled)
  2. The CS counts the disabling as uninstalls
  3. Both
• What happens if the user sees the Error message but doesn't actively click on "Remove" or "Accept" (I assume the extension is simply disabled) and then I quickly rollback to a version without the new permissions? Is it reenabled? Or the only users I managed to save were the ones that didn't open their computer at all that day and missed the permissions-demanding extension update?
  

Thanks again

[Deco]

1: The answer is yes because it was disabled, and they removed the extension, it will not trigger the URL, as your extension cannot operate at all, and your URL is dependent on the uninstall process triggering the URL, it will never appear.

2: How Chrome works is to automatically disable an extension until user consent is granted for differing permissions, so since you changed it in the first update, the extension is disabled until the user has granted this. This is per-browser, that is, if you quickly "rollback" then people who have not received the update (by not using Chrome) will never have received the original changed permissions, so this state will never trigger for these users to grant user consent.

In the context of a user who sees this prompt but clicks neither, the extension will still remain in a state of being disabled (and as clarified before, will still receive updates, so they will receive any new version which fixes this), however, will remain explicitly disabled until manually re-enabled by them.

Hope this clears up the questions.

Cheers,

Deco

[GrungeL]

OK, thanks, it makes sense.

So, besides the uninstalls number that appears in the Chrome Store, there is a huge number of disablings, I just saw it in the CSV of the enabled / disabled report..

Just another question: if the user doesn't interact with the Error warning and I publish a version without those permissions, does that warning go away? If that's true, rolling back might actually be a bad idea, since it implies withdrawing the warning, which is virtually the only way to get the user to re-enable the extension, right?

Thanks

[Deco]

Yes this assumption is correct, since the browser only detects for changed permissions, if you reverse this then the prompt will no longer appear, note that there are some occasions where outwith this specific scenario if a permission change is more severe, then it will take precedence (assuming you change multiple permissions in one update), causing the prompt to still appear even if one particular one is reversed.

In general, be mindful of this to prevent it happening in the future.

Thanks,

Deco

[GrungeL]

I see, thanks.

The only part that I don't understand is why today the report of disabled users (the downloadable CSV) says I have four time less disabled users than yesterday. Is a bug or did I recover them somehow?

[Deco]

As noted before, the statistics are updated daily, so reported figures can often have anomalies if certain events occur like this. Give it 2 or 3 days to have a stable idea of where your current numbers actually are once things iron out.

Thanks,

Deco