Web analytics helps to answer how data from a web server's log can be harvested to generate useful and actionable business intelligence, particularly when the data is combined with existing customer and sales data in a Data Warehouse. Wiki page defines web analytics as "the measurement, collection, analysis and reporting of web data for purposes of understanding and optimizing web usage".
How web analytics work?
When a request for a web page is initiated by a web browser on a personal computer to a web server, identifying data is sent with the request. This information includes the physical address of the personal computer on the WWW (i.e. IP address), the originating and destination URLs, and the date and time of the event.
This data is required by the web servers in order to return web page content back to the web browser on the the personal computer. This is illustrated graphically:
The web server retains all incoming communications in a log that is stored on the web server. This data can be copied into a Data Warehouse and sliced and diced as needed to generate business intelligence about website visits.
In addition to transmitting basic Internet communication data, many websites store "cookies" on the visitor's personal computers. These cookies contain identifying information about the visitor's previous visit to the website.
What does web server log contain?
A web server log is the file(s) on a web server that contains a history of web page requests that have occurred. The following data is available from server logs:
IP Address: The address of a device attached to an IP network. Every client, server and network device must have a unique IP address in order to communicate. Every communication that occurs contains both a source IP address and a destination IP address. The format of the IP address looks like this: 255.219.12.2
Date & Time: Data and time that the communication request occurred
User Agent: Name and version of the Web browser (e.g. Internet Explorer)
Web Page: URL of the web page being requested. The address can contain parameters with session specific information (e.g. form fields)
Referrer: The URL of the previous webpage from which a link was followed
Analytics on registered and public users of a website
There are two type of Internet Applications: authenticated and unauthenticated. An authenticated application is one that requires a login ID and password. An unauthenticated application allows visitors to use the Internet application without signing in. An authenticated applications can generate much business intelligence about customers, their preferences, interests and purchases.
To obtain a user ID and password, web users are normally required to register. They are normally required to provide a valid email address, their name and address and some personal profile information such as hobbies and interest. This data is stored in a database that can be integrated with web usage data from web server logs.
Sometimes users aren't required to submit a registration over the Internet. Using data from their internal corporate data, the business sends a login ID and password to their customers. Each time the login ID is used, data is collected about the Internet session.
Whether or not a visitor is authenticated, most Internet Applications retain statistics on application usage such as how the visitor navigated through the application. This is required to identify technical issues and usage patterns.
By leveraging data available from authenticated applications, subsequent visits can be customized for users.
How web analytic data is leveraged?
Insight gained from analyzing web server logs and data from Internet application systems can be leveraged to formulate plans for evolving your corporate website. Shorter-than-expected visits, unexpected navigation paths and transaction abandonment prior to completion can suggest design or usability issues that need to be addressed.
It is very useful to analyze the "entry" and "exit" pages to a website. Visitors frequently enter through a 'back door" rather than through the home page. This isn't necessarily a problem. In fact, website visits can often be increased by treating each web page as a "home page". This involves implementing an appropriate navigation design and making full use of meta tags on all pages.
Web analytics should also consider pages in a web as sets of visual "components" that have been assembled into webpages. Some components may be used in multiple pages, while others are not. Some components contain core content that all visitors see whereas other components might only be visible tor certain categories of users.
Because web pages are dynamics, web analytics must account for this. The content on a given web page can be much different from week to week.
Why we need web analytics?
- To identify technical and navigation website issues
- To better understand the customer's unique needs, interests and patterns
- To identify improvements for website design
The business intelligence that is gained can be used to design customized web sessions for groups of users that are based on demographics common interests and similar behaviors. For example, a web session can be customized based on the gender, language or geographic location of the visitor.
Customizations to sessions can entail presenting visitors with web content and advertising that is of direct relevance to them. By improving the user experience, businesses can increase sales and visitors can be motivated to return to the site.