Closed Bug 1246710 Opened 4 years ago Closed 3 years ago

Add telemetry probe to understand how often users load content in reader mode

Categories

(Firefox for Android :: Metrics, defect)

defect
Not set

Tracking

()

RESOLVED INVALID

People

(Reporter: barbara, Unassigned)

References

(Blocks 1 open bug)

Details

We currently have a probe that tells us if a Reader Mode fetch was successful or not but as far as I know, it's not really telling me how many times users intentionally triggered a Reader Mode view.
We do have data to tell us how often the user hits the reader mode button as opposed to how often it's shown:
http://mzl.la/1Xec0dH

Does this answer your question?
Flags: needinfo?(bbermes)
It might :)

Can you help me confirm the following:

- 15% of the time content was loaded, the content was RV-capable
- 0.62% of the time, the RV button was actually clicked. Would this really mean _from_ the 15% of the time the button showed up, 0.62% clicked it? Should we have this number in the same histogram than the "Button shown"? 
- 0.02% exited RV mode, indicating that people seem to stay in the RV mode?

So maybe this is an indicator that if we were to show the RV button more often, people would use it more often.
Flags: needinfo?(bbermes) → needinfo?(margaret.leibovic)
We also start a session every time a person taps the "reader" icon in the URL bar:
http://mxr.mozilla.org/mozilla-central/source/mobile/android/chrome/content/Reader.js#252
(In reply to Barbara Bermes [:barbara] from comment #2)
> It might :)
> 
> Can you help me confirm the following:
> 
> - 15% of the time content was loaded, the content was RV-capable

Yes. Well, really, theoretically - it just determines if we showed the button or not, depending on our heuristic.

> - 0.62% of the time, the RV button was actually clicked. Would this really
> mean _from_ the 15% of the time the button showed up, 0.62% clicked it?

No, this means 0.62% of all of these histogram events. So you'd need to do some math to determine what % of the time it was shown it was clicked. If I'm doing it correctly, I think that means people click on the button 4% of the times it's shown.

> Should we have this number in the same histogram than the "Button shown"? 

I suppose this could be in a different histogram... histograms are not well suited for this sort of data, we should probably just make this into UI telemetry and then compare the raw numbers of events we're seeing.

> - 0.02% exited RV mode, indicating that people seem to stay in the RV mode?

Yes, there's a big difference in the number of people who click the button to enter RV vs. the number of people who click to exit. This means they must stay in RV until they navigate away or close the tab.

> So maybe this is an indicator that if we were to show the RV button more
> often, people would use it more often.

Sure, but we have to make sure it actually works. The reason we don't show it all the time is because it doesn't actually work in all cases.
Flags: needinfo?(margaret.leibovic)
Does this bug need any action? It seems the reader session should be sufficient to see how often reader view is used (I'm not too familiar how the sessions are analysed though)?
Flags: needinfo?(bbermes)
Flags: needinfo?(bbermes)
Do we still need this probe? Maybe I can help
Flags: needinfo?(bbermes)
I think we are good for now, Nevin. Thanks
Flags: needinfo?(bbermes)
According comment 7, I'll close this bug. If there's future need, please feel free to reopen it.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.