How can I track a user using two devices on my site?

Durga G August 26, 2013
Pinterest Stumbleupon Whatsapp

A user enters into my online purchase website and looks into purchasing an item using a mobile device and when he returns to his office, he started to purchase the item he reviewed using his mobile using his desktop. I want to figure out that he is the one who reviewed the item on his mobile and purchased on his desktop.  How can i do this?  This is called as multi tracking. Are there common attributes present in both devices to identify user activities?

  1. Oron Joffe
    August 27, 2013 at 12:46 pm

    Unless the user logged on to your site on both devices, there's no way of doing it. Two devices, probably on different networks (e.g. office network and 3G) could belong to *anybody* and there's simply no way of linking them automatically in a reliable way.

  2. Jan F
    August 26, 2013 at 2:51 pm

    The only common attribute I could think of is the public IP address. This would obviously require both devices connected to the same network.

    If the desktop is using a leased line but the other device mobile data then I don't see how one could identify it being the same user except maybe using a probability calculation with all the items being reviewed and the final purchase order.

    Other then that the best you can do is encourage users to log in.

    • Durga G
      August 29, 2013 at 4:15 am

      Hi Jan,

      Thanks for your answer.. If we can do Probability calculation, What are the attributes we can consider to do this calculation?

    • Jan F
      September 2, 2013 at 7:57 pm

      If the system was setup to log user sessions with all the items (pages) they reviewed you could compare them and see if sessions A and B were visiting the same pages. But this will never really give you the answer you are looking for! It will be less than 50-50 accurate

      Let's take the situation you mentioned as an example:
      The person reviews items using a mobile device and then buys them using a desktop. In almost all cases a person will buy far less items than they reviewed. So the person will review e.g. 15 items using the mobile device and then buy 5 of them using the desktop. It's unlikely the person will revisit all 15 items on the desktop but instead only visit the 5 items selected and just buy them.

      So comparing the mobile session and the desktop session for this doesn't really give you a conclusive answer. They only match 33.34%

      Anything further is circumstantial:
      If you have 10000 items in your shop and the above occurred there is a reasonable chance that it's the same person.
      If you have 200 items in your shop and the above occurred then you can roll a dice whether it was the same person or not.

      If all the items reviewed and bought are from the same or complimenting categories (e.g. mainboard, cpu, ram) then it's more likely coincidence.
      If all the items reviewed and bought are from completely different categories (e.g. computer parts, kitchen appliances and health care) the chance goes up again.

      At the end of the day we are back to the general idea:
      Encourage users to log in!