2 or 3 unique visitor per unique visitor id

[m2masari]

Dear All,

I came across a rather peculiar discovery in visitor traffic data. The presence of two or three unique visitors per unique visitor ID strikes me as highly unusual. How could this be possible? 

Thank you 

[Jennifer_Dungan]

What is your "v94 Visitor ID" representing? Is it a copy of the Adobe ECID? Is this your in-house User ID/GUID for logged in users?

If it's an in-house id, this is pretty common, people would log in on their computer, maybe then on a different browser on that computer, then on their phone, then on their tablet, etc.. this would be one system id mapped to multiple "device" level (i.e. UV) identifiers.

If this is the ECID then that is strange... Are you possibly using the "Append ID" functionality to cross stitch domains? Maybe the way you are sending ECID to your eVar is being impacted by site level caching (cloudfront or another caching service)?

As in, I come to your site, you send my ECID to Adobe... someone else on another computer comes in to the same page... my ECID value is cached, and is then sent to your custom tracking solution instead of their ECID?

The most reliable way to capture the ECID is to use Dynamic Variables (use "D=mid" in your eVar... 

This tells Adobe to duplicate the value from the tracking call param "mid" into the eVar or prop specified. 

No relying on trying to read from cookies, or run a JS script to get the ECID.

If you are already doing the above, and still see this behaviour, then that is very unusual... 

[sonjaz24855261]

As already mentioned, this can have several reasons and probably also depends on the type of visitor ID that you track. 

But we had a similar issue, with 2 or 3 Unique Visitors per Visitor ID. It appeared only for a very small number of visitors IDs, maybe 1000/5000000. In our case the explanation was hash collision. As described here: https://experienceleague.adobe.com/docs/analytics/implementation/validate/hash-collisions.html?lang=...
In data tables with a very high number of values it can happen that a value gets the same hash like another one. I guess this is a sacrifice to have on the other hand a better performance of the tool.