Systems Engineering and RDBMS

Web based Analytics and Mobile Devices

Posted by decipherinfosys on June 5, 2008

A couple of weeks ago, while helping one of our clients with setting up the web analytics and clickstream analytics for them, a question came up about Mobile Devices Analytics. There are a lot of challenges in this area – besides being a fact that that is a new area for us, after doing some research we found that mobile browsers do not always accept cookies which makes identification of unique visitors tough – we do have a workaround though by relying on the OS + IP + the browser.

Another big problem is dealing with the changed IP addresses as the mobile users move between the mobile towers. Many vendors typically use a javascript page tagging mechanism to collect the information on the OS, size and resolution of the monitor. And many handsets do not support javascript🙂 As a result of this, one can miss out on critical data like conversions and keywords. Another approach is to use Log file analysis. Analog is one of the free programs that is available in case you want to analyze your log files and see the different reports.  But for Mobile Analytics, this solution also falls short. Why?  Because any mobile specific information like the handset capabilities cannot be captured and also because all the people using the same carrier and the phone type cannot be distinguished from one another in this case.

A lot of industry vendors propose these solutions to such problems for mobile analytics:

a) Page tagging: This technique is used by Mobilytics which is a mobile analytics provider and instead of tagging the page with a javascript, they use a snippet of code on the web site for a page tag.  This allows them to capture a rich set of data as well.

b) Packet Sniffing: If you install a mobile analytics server in your environment, a switch can then be used to pass a copy of the incoming packet data to the mobile analytics server and at the same time allowing the same packets to pass to the mobile web server without any delay.   Using this approach, you can get information on the handset resolution, basic web data, handset type, mobile operator and the type of the browser.

Mobile Analytics is a very challenging and a very interesting field and as we do more research on the topic and implement solutions for our clients, we will keep writing up more on it.  Stay tuned.

2 Responses to “Web based Analytics and Mobile Devices”

  1. Amethon’s Mobile Analytics product is an example of a packet sniffing solution and there are a number of other advantages to this methodology worth mentioning.

    1. We are able to report on a number of mobile relevant metrics such as data consumption, content downloads and errors.
    2. We are able to leverage operator gateway injected HTTP headers which can be used for visitor identification instead of cookies or IP address
    3. We can use site cookies or custom HTTP headers to filter traffic. For example, one of our customers sets HTTP headers with demographic data once a registered user logs in which can then be used to generate reports for specific demographic types.
    4. There is no delay or latency introduced for the end user which is noticable for page tags on low bandwidth connections.

  2. Thanks Michael. That is very useful information.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
%d bloggers like this: