CODEWITHSUNDEEP
google-tag-managerweb-analyticsgoogle-analytics-4
Gus
Gus

Reputation: 83

How would you identify if a visitor to one of your sites is the same person who visited another site of yours before (different domain)?

My question is more of a conceptual one, but in my specific case I am using Google Analytics 4. If the question is unclear, here it is in scenario form: Some guy visits my site x.com after a google search. He closes the tab, does another google search, and arrives at my other site y.com. How do I know it's the same person? I don't think there's anything I can do with User ID's in this situation. How would I solve this?

Upvotes: 0

Views: 451

Answers (2)

Jknight
Jknight

Reputation: 105

This isn't without fault, but if you are implementing it via Google Tag Manager, you have more control over the data being sent and on top of that, if you are transporting the data via Google Tag Manager server side container.

You would use a single server (but possibly different containers) or use BigQuery and either use the templateDataStorage API call or the BigQuery API call.

Essentially, the first time you see a google cid or an IP address or combination of user agent and ip address you would store it in the server or in a BigQuery table as a key and create a random associated value next to it.

At each time, across all your sites, you would check to see if the IP address or CID or combination of user agent and ip exists in the server or in the BigQuery table, then output the random value as a custom dimension and if not, it will create one.

Upvotes: 2

Eike Pierstorff
Eike Pierstorff

Reputation: 32770

Actually you probably wouldn't.

Presumably you could try fingerprinting, but depending on your legislation that might not be quite legal, and it tends to work a lot better in a lab than in real life. Also browsers start to implement anti-fingerprinting measures such as trimming the user agent, and denying access to browser properties such as installed plugins.

I have heard of experimental approaches to recongnize users via usage patterns - e.g. how do they move their mouse etc. I am not aware of any actual product that uses this, and I am not convinced it is a useful (or even legal) approach.

But in general, when it comes to cross-domain detection for unrelated visits (moving from domain to domain works via link decorators, and even that is affected by browser protections) you have the combined power of browser vendor against you, who try to make this harder (either for genuine concerns about privacy, or to establish themselves as the single gatekeeper for user identity. E.g. Google has a huge user base that is almost constantly logged in to Google accounts or Android smartphones, which helps with identifying users all over the web).

Upvotes: 1

Related Questions