Google Analytics: how to use custom dimension on different website to identify intranet users

4
votes

SITUATION

I have a main public Liferay website, that is therefore accessible both by intranet and not-intranet (i.e. public) users.

I also have a Liferay intranet website, which is accessible only to intranet users because is protected via a login page. The login page to the intranet website is public. After you successfully login, the intranet website is loaded.

EXPECTED:

In my Google Analytics account for the main website, I want to differentiate intranet users from public users (e.g. in order to understand how the 2 categories behave).

Questions

  1. Can I use a custom dimension to solve this problem, or is there a better way?

  2. Custom dimension data has to be sent via hits (UPDATE: by "hits" I meant either pageview or event hits, I am not referring to the dimension scope, cfr. https://developers.google.com/analytics/devguides/collection/analyticsjs/custom-dims-mets), therefore I should:

    • load the Google Analytics tracking code of the main website on the intranet website (the site displayed after successfully logging in)

    • send a pageview hit from this Intranet website to the main website together with a custom dimension, e.g.

      ga('send', 'pageview', { 'dimension1': 'I am a intranet user' });

    Is this correct?

  3. Does the above mentioned solution have any impact on my Analytics data in the main website (e.g. more pageviews due to the tracking code added to the intranet website, or strange behaviours in counting user sessions, etc.)?

Thanks a lot.

UPDATE:

Actually, the solutions proposed below would not work because the 2 websites (intranet and not-intranet) are considered different domains. So, even if I had the following domains

and I sent data to the same UA account (i.e. the company website UA account), they would be counted as different visits. Quoting Google (see https://developers.google.com/analytics/devguides/collection/gajs/gaTrackingSite#profilesKey)

If a user independently visits two sites that are tracking in the same view (profile), such as through a bookmark, these visits will still be counted under separate sessions. In this scenario, the linking methods are not invoked, and thus there is no way to determine the initiating session for a given user.

So, how could I solve my problem? Would it be possible to solve it by implementing cross-domain tracking (https://support.google.com/analytics/answer/1034342?hl=en), and how? Thanks a lot.

2
javascriptgoogle-analyticsanalyticsmetricsdimensions
Do users login via different devices ? If so you might want to set up a userId view, which would allow you to track users across their different devices (aggregated, not individal users, so no privacy risk). - Eike Pierstorff
@EikePierstorff could you please elaborate? I am not sure what you mean. - MDT
I added this as another answer, explanation was a bit long for a comment. - Eike Pierstorff

2 Answers

3
votes
  1. Can I use a custom dimension to solve this problem, or is there a better way?

Yes, custom dimension is perfect for this.

  1. Custom dimension data has to be sent via hits

The User-level scope is more appropriate than the hit-level one for what you want to achieve. The linked document explains in detail why, and gives an example similar to your use case.

  1. Does the above mentioned solution have any impact on my Analytics data in the main website

Yes, impact is mainly that you will have extra data corresponding to the visits to the intranet.

1
votes

A custom dimension works well for your purpose. You will get additional hits for visits on your intranet site, but you can segment them out via the custom dimension to separate between inter/intranet.

Since the intranet requires a login there is one other way you could try, which would have the additional benefit of allowing for cross-device tracking (if that is beneficial to you).

Google calls this "userID", despite the fact that it must not be used to identify individual users. On login you pass in a unique value per user that is set by your backend system (UUID format is suggest but any unique string would work). Since it is not assigned by the tracking code but set by your system it will be the same id on every device. It is used to de-duplicate users, i.e. persons that log in from multiple devices will be recognized as single users (also useful if people delete their cookies - the userID can be used to aggregate sessions into unique visitors).

To make this work you need to set up a special view that contains only data from visits where the userId is set (so you would have a view for your public site and a view only for your logged-in users). You get a few special reports, for example one to tell you how many users log in from different device categories.

What the userID should not do, and in fact must not do according to Googles terms of service, is to identify individuals. The userId is not exposed in the Interface, and you must not store it as a custom dimension. If you store it on the client side in a cookie you must unset it once the users log out. It is merely there to allow continuous tracking of users independently from cookies (plus you need to amend your privacy policy if you want to use this).

Of course you can combine both approaches to get even more insights.