Granular Analytics

From ExtractValue Wiki

Granular Analytics refers to the Centricity module where web, marketing and sales activity is segmented per User.

Business Description

Gain insight into the interests of your prospects by leveraging web, e-mail and sales analytics at a user-specific level. The Centricity platform tracks users across its network of websites and web portals--capturing who, where and when web pages and files were accessed. This information can then be used to deliver targeted advertising, trigger-based e-mail campaigns, and even by sales people over the phone.

Web Analytics

Goal

Identify users on different sites and track their activities on each site.

Need to associate the unique user visiting through 3 different situation.

1. Our sites
2. Sites managed by us
3. 3rd party sites that have our code snippets

Actively track all activities the user performs on the sites.

• User IP
• Visit links
• Download documents
• Amount of time on page
• Referrer
• Browser
• OS information

Technical Information Flow

• All sites will contain a JavaScript snippet which will phone home to Centricity.
• The script will pass the cookie (analytics.centricty.com) value to Centricity.
• Centricity will read the cookie information, or if the cookie does not exists, Centricity will pass back information for the script to set a new cookie on the user's browser in response to the user browser's request.
  • The information will uniquely identify the user, it will be a hash of IP, OS and Browser details.
  • There will be opting out cookie which indicate that activities should not be collected
• An application (e.g. JavaScript) or other browser client-side technology will be set to collect user activities
• Defined activities such as downloading document will be tracked and phone home to centricity

Requirements

Tracking of User to be implemented in JavaScript. Tracking must be achievable in 95% of currently used browsers.

1. This currently implies IE6 + IE7 + FF2 + FF3 + Safari ( based on Wikipedia article ) - This means if Netscape 4 does not work, it is not a problem.
2. Create a Test Framework for testing this.
   1. Test multiple users from one server visible IP address.
   2. Test image and url logging.
   3. Have test pages that utilize common widgets from the JavaScript libraries extjs, YUI and jQuery. Test using these widgets to make sure data is not accidentally logged from them
   4. Have a test form and ensure only the submission and the required elements are passed.
   5. Test framework must be accompanied by a sample output and a description of how one should be able to generate the output which matches that sample ( Minus date and cookie data )
   6. Test framework must be tested on all above platforms.
   7. Test framework must have an output created from the final test

Tracking must utilize both remote IP address and a browser cookie to achieve the highest degree of accuracy.

1. Upon loading the script, it must check for a cookie from the host domain, and then set a cookie according to the following rules if not created.
2. If the cookie is new, the script must report the creation of the cookie.
3. The cookie must be in a format which is most likely to be unique, so it must use:
   1. The current ip address
      1. Access the ip address in JavaScript and convert to hexadecimal.
   2. The current time, with as much precision allowed
      1. Access through JavaScript and convert the second and sub-second portions into hexadecimal
   3. A small random value, to discourage accidental aliasing of private addresses
      1. Also converted into hexadecimal
4. This allows laptop users to be tracked between WiFi hotspots )
5. Finally all tracking must return this cookie value to ensure that unique users are tracked.

Tracking must track when a page is loaded

1. If such data can be accessed it should return the user's current time zone and referrer of the page it is tracking.

Tracking code must track, in addition to access

1. Time Spent on Page, using the most accurate method possible.
   1. The less accurate method is : by logging the difference between load and unload time when the page is unload.
   2. The most accurate method is : by checking at a 5 second interval whether the page is being used, and using this value when unloading the page.
      1. A High level example of how this would be done would be to check for an unblur on the document element and use the time between blur and unblur, and adding it to a total.
2. All links clicked on
   1. Pass back link url
3. The Submission of Forms
   1. Pass back the name of the form

Tracking should be verified to work on pages which involve JavaScript libraries currently in use: e.g.: extjs, yui, jQuery.

1. The home pages of these libraries have tutorials if unfamiliar with these libraries.
2. All should be in the test framework from above.

Finished JavaScript will have two forms:

1. First which is the documented code, which is the basis of further work
2. In addition, a compressed version, to be used when sending the code to clients over the web.
   1. Use a 3rd party tool for the compression. This will strip comments and minimize variable names, and makes it unsuitable for editing.

E-mail Analytics

Austinguu 00:58, 12 September 2008 (UTC) says:
SYBRANT: PLEASE IGNORE THIS FOR NOW

• Hard bounce
• Soft bounce
• Delivered
• E-mail open rate
• Click Through Rate (CTR)
• Forward rate

Related Articles

• Centricity
• Lead Management System (LMS)
• Customer Relationship Management (CRM)
• E-mail Service Provider (ESP)

Category Tree