Update activity any bank wire and fund accounts 

I saw data at https://uma.googleplex.com/histograms?endDate=latest&dayCount=28&histograms=Autofill.SaveCreditCardPrompt.Upload.FirstShow%2CAutofill.SaveCreditCardPrompt.Upload.Reshows&fixupData=true&showObsolete=true&showMax=true&filters=isofficial%2Ceq%2CTrue&implicitFilters=isofficial

Although the volume is very small. Does reshow mean user click on the icon in url bar beside "bookmark"? Probably this means very few people will click on it?

I have not implemented this code. I was hoping you would investigate and determine if it's broken. Perhaps it's not useful so it can be taken out?

Double checked the code: https://cs.chromium.org/chromium/src/chrome/browser/ui/browser_command_controller.cc?l=451
Reshow does mean user click on this button: https://screenshot.googleplex.com/Uu04wpz68So.

Agreed that we can take out that feature. In the last 31 days, save bubble was shown 388,595 times, while it was reshown only 90 times.
+jsaul, dlkumar, mikeying, rogerm for their opinion. 

Yes, Reshows are when the user closes the bubble and then opens it back up by clicking on the toolbar icon as Shanfeng posted.

I'm okay with merging these metrics; it's good to try to log whatever we can, but this metric in particular hasn't been of any use to us.

Reshow is only capturing times when the bubble was hidden (by scrolling?) and subsequently reshown using the UI elmement shown in the linked screenshot? I.e., what are the scenario's under which this event happens?

Is it important to capture these events separately from showing the prompt?

If you're capturing the frequency of user clicks on that UI element, then tracking the outcome of the reshow events would at least confirm/deny that users intentionally select that UI in order to save.

That said, the current volume of events would suggest that the CC icon in the omnibox is either undiscoverable or not-compelling/useful most of the time.

Mathieu, I may misunderstand you. When you say "it can be taken out", do you mean remove the reshow icon or the metrics?

I'm thinking about removing the whole feature. If we decide to keep the feature, I would prefer to keep the metrics as well. So we can know the usage of that button.

Scrolling doesn't hide it; the scenario here is clicking the bubble to give it focus, then clicking away from it (or clicking a TOS or Learn More link), which makes it go away.  After this point one can click the credit card icon to make it show up again.

Shanfeng: I assume the metrics.  Keeping the icon would be the intentional in case the user changes their mind.

I meant the metrics, although with such low UI it's debatable whether the feature is useful at all and warrants the additional complexity (albeit low)

*such low usage

From my point of view,

If we want to keep the feature, I don't think we should remove the metrics. At least this can help us understand that few people is using the feature.

About whether we want to remove reshow, I think we should:
1) Very few people use it. Reshow upstreamed 23 card over the last 30 days. (Not sure how many of them are just for testing).
2) We can always upstream the card when user use it next time.
3) It somehow adds complexity to the upstream code. Especially after CVC fix flow code was added.

Please leave a comment if you are not OK with this.

My two cents: I'm personally not a big fan of removing the feature.  Reasons:
1) Clicking a TOS link or the Learn More link makes the bubble go away, and this is the only way to get it back.  We'd essentially be penalizing users who actually look at the Terms of Service by saying "Nope you viewed the TOS, can't save your card now!"
2) I think it's consistent with other features in Chrome.  If you get the "Do you want to save your password for this site?" and click [No thanks], the key icon still sticks around if you change your mind (and I've changed my mind before).

Agreed with Jared. Actually talked with Kumar yesterday and I understand that we should not remove a feature because no usage. As long as the feature improves usability or user experience, we should keep it. So let's keep it for now.