1353820 - Allow User to stop their LeanPlum Tracking

Status: VERIFIED FIXED

(Firefox for iOS :: General, enhancement, P1)

Description

[:sdaswani]

Ability for users to turn on/off tracking and to reset their device ID. This should tie to the same setting as Adjust.

Comment 1

[:sdaswani]

Robin & Barbara, should we just reset LP tracking the same place that users reset Adjust tracking, or do you want a separate settings item?

Comment 2

[Barbara Bermes [:barbara] - NI please!]

+ Elvin and Marshall

Suhseel is our new engineering manager and will make sure Leanplum gets implemented into our mobile products. We are starting with iOS and we have some questions for you.

Do you suggest we include Leanplum into the one share anonymous data opt-in/out setting in iOS, like we did with Adjust, or should we start thinking of creating a "send data to third party" on/off setting for Adjust and Leanplum?

In addition, we were wondering what you advise do to in the following case:

Users is counted in Leanplum, e.g. device ID 123. User now turns off the data sharing option in iOS, but decides after some time to turn it back on, from a pure privacy-focused angle, do we need to set a new device ID or can we keep the previously sent device ID 123 to be "reactivated" in Leanplum?

Comment 3

[Merwin]

Barbara,

I'd suggest keeping this attached to our "share anonymous data" setting.  The data categorically is still similar to the other data associate with that setting. We would describe the third parties in the privacy notice and Sumo but wouldn't typically identify them in the product.  I'd appreciate Elvin's view on this.

Regarding the last question, I think we generate a new device ID.  Presumably the user didn't want to be tracked or associated with the data when they turn off the data collection, so for them to turn this back on and then be reassociated with the data seems strange.  When they turn the data collection back on, I'd think the user would expect to be associated with the data going forward but not going backwards to when they previously may have had that setting on.

Marshall

Comment 4

[:sdaswani]

Thanks for the info Marshall and Barbara. While we wait for Elvin to reply I will check with LeanPlum that associating a new id device id with LeanPlum does not result in a new MAU.

Comment 5

[Joe Cheng [:jcheng] (please needinfo)]

Similarly for Android, need recommendation on which setting LeanPlum is tied to. Cloned and filed bug 1358117

Comment 6

[:sdaswani]

Attachment: Pull Request

Comment 7

[:sdaswani]

Merged in https://github.com/mozilla-mobile/firefox-ios/pull/2680

Comment 8

[Aaron Train [:aaronmt]]

What is missing here?

Comment 9

[:sdaswani]

(In reply to Aaron Train [:aaronmt] from comment #8)
> What is missing here?

We backed out the change so I re-opened it, not to mention we need to add some extra code to get this working correctly.

Comment 10

[:sdaswani]

Leanplum's recommendation:

"Alok suggested that we use enableTestMode to stop collection immediately as while this will store data to the local device, it will never send this data to Leanplum. If this is not possible, you can write a helper class around Leanplum. "

Comment 11

[:sdaswani]

This is a blocker to the 8.0 release too and will need uplift to 8.0.

Comment 12

[:sdaswani]

https://github.com/Leanplum/Leanplum-iOS-SDK/issues/27

Comment 13

[Bryan Munar (:bkmunar)]

Attachment: PR

Comment 14

[:sdaswani]

Merged in https://github.com/mozilla-mobile/firefox-ios/pull/2798#event-1115958041

Comment 15

[Stefan Arentz | :st3fan | ⏰ EST | he/him]

Has this been tested as follows:

- install the app, generate some events, look in the dashboard that those events arrive
- disable 'send anonymous usage data'
- generate some events
- enable 'send anonymous usage data'
- look in the dashboard and make sure that the events were not collected and sent

I am asking for this because what Leanplum says:

 "Alok suggested that we use enableTestMode to stop collection immediately as while this will store data to the local device, it will never send this data to Leanplum." 

seems to indicate that events will be queued/recorded to disk and then sent later when this mode is disabled. I'd like us to verify if this is the case or not.

Comment 16

[Stefan Arentz | :st3fan | ⏰ EST | he/him]

I reopened this bug because I am unable to properly uplift it to v8.x.

(I tried to merge, but I don't know enough about this code to make the right decisions)

How about we create a new PR on the v8.x branch with the same changes?

Comment 17

[Stefan Arentz | :st3fan | ⏰ EST | he/him]

Attachment: PR https://github.com/mozilla-mobile/firefox-ios/pull/2817

This is the same patch, but done on top of v8.x

Comment 18

[Stefan Arentz | :st3fan | ⏰ EST | he/him]

Comment on attachment 8877628 [details] [review]
PR https://github.com/mozilla-mobile/firefox-ios/pull/2817

I'm confused. This patch is exactly the same as the second PR attached to this bug. And the first PR already landed on v8.x. So I am not sure why I asked to redo this.

Comment 19

[Simon Basca [:SimonB]]

Verifying as fix on beta 8.3(5817). Turning off 'Send anonymous usage data' will not track events in Leanplum.