unique udentifier

[peeyushbansal]

My stakeholder is asking to identify if users are returning to application and i wanted to check if we have any unique variable tagged to each user that i can export to check how many time that user have logged in.

[Jennifer_Dungan]

Well this is more of a question for your team.. as we don't know what you have and what you are tracking...

We can help by giving you samples of a how other sites handle this, but whether or not you have the data collected / available at this time, that is something you will have to determine.

But generally speaking, when sites have a login, there is a dimension used to track the user's ID or Guid (but if this is an eVar, the retention could exist after the user has logged out)... which is also why there is generally an event associated to "successful login action". You can also have another dimension which holds the current "logged in status" in the event that you allow for long running rolling sessions, that allow users to stay logged in for months.

Since the same user can use multiple devices, you can create a calculated metric for "Distinct Count" and your User Guid... this will tell you unique "known" users... 

If you want to see actual "Log In Events by User" - You can also create a segment for "Successful Login" exist (either at Hit or Visit level) and you can use that segment stacked with your calculated "Distinct Unique Count Users" to see Unique Login Events

If you want to see "Active Logged In Users" - You can create a segment for "Logged In Status is true" (or yes, or however you are indicating that a user is logged in) and stack that with your "Distinct Unique Count Users" to see active logged in users. (This is assuming that your dimension continues to track the user's id/guid beyond being logged in... IF you only track the ID when the user is logged in, then your Distinct Count Users should be enough... 

[peeyushbansal]

@Jennifer_Dungan  we are not storing userID as it is PII data but my stakeholder want to know returning users over a period of time. so they can check if people are using application. Can you let me know how i can do that. We dont have evars, props created for user information.

[Jennifer_Dungan]

Well a user id isn't really PII... (it is... but only with a lot of extra work to look that user up... so generally this isn't considered PII - when I say userID I do not mean the user's login name particularly if you are using an email.. but an ID number or GUID, something that when seen in isolation no one can know anything about that user)

Without such a "fingerprinting" technique, you can only use the UV value, which will show inflated values since your users could be using multiple devices, multiple browsers, clearing their cookies, etc.

But if you do choose to use UV, you still need to track a "Login Success" and/or "Logged In Status" to be able to pull the data you want....  you can use these against the generic UV metric.

In the meantime, you can get a generic users on your site simply by looking at UVs against the time frame... UVs will scale itself to the time breakdown...

So if you have a Report Period for last 12 months, then have a freeform table with UVs as the metric and Months as your breakdown; each month will show the UVs that used your site in that month (again, inflated as mentioned above), and the total column will show the number of users in that 12 month period (you will notice that if you were to add the values from the 12 months, they will add up to more than the total, this is because the UV total is de-duplicating the users that came to your site in multiple months)

[peeyushbansal]

@Jennifer_Dungan , Sorry i am novice to adobe, can you please confirm what is UVs and how i can add same to report.

[Jennifer_Dungan]

No worries  

UV is short for "Unique Visitors"

The name is a bit of a representation of historical times... back when identifying users wasn't terribly advanced.... "Unique Devices" would likely be a better name, but that would probably confuse people... this is the equivalent of "Users" in Google Analytics (which has the same potential inflation issue)

[peeyushbansal]

Thank you, but how can UVs determine if users are returning to application. they are interested to know if same user is returning to application over a period of time.

[Jennifer_Dungan]

Unfortunately the answer to this is "ish"

If you aren't tracking a unique user ID, you are relying on cookies.... If the user changes devices, or clears their cookies manually, or uses Incognito Mode, or is on Safari or another browser that is forcing the cookies to be deleted, then you will not be able to tell if the same user is coming back.

If the cookies remain, then the user will be identified by their UV cookie, and will register as the same UV from visit to visit....

But if you are against tracking User IDs... this is your only option. You sacrifice more reliable user fingerprinting for generic and less accurate identification....

[peeyushbansal]

@Jennifer_Dungan , can return frequency helps here. what does this metrics indicates and how should i use it. labels are 1 to 3, 3 to 7 are 3 and 3 inclusive in both rows.

[Jennifer_Dungan · Accepted solution]

Yes, Return Frequency can be used to show bucketed returns. But this won't necessarily represent "Logins".

Even if your site requires login, users would hit a login page and might bounce without login in, or your site may allow for long term logins, meaning they don't have to "log in" each time.. so I guess it really depends what your definition of "login" is - users on your site; users actually logging in (the actual action of logging in); etc...

If you are just looking at basic UV metric (regardless of the login status, or identifying the user based on a unique id/guid), then yes, what you showed can work.

You should also be aware of a few things with Return Frequency:

1. Complete New (first time users) won't show in this report... a user must return to the site to show up
2. The same user can show in multiple buckets... If I come to your site today, later today, tomorrow, 5 Days after that, then 16 Days after that - I will be in "Less than a Day", "1-3 Days", "4-7 Days" and "14-30 Days" buckets

You can also do a histogram using "Days Since Last Use Instances" which could create a nice visual