Faridkot, Punjab (India).

Web metrics are established goals and standards for measuringwebsite performance. Web Analytics can be used to analyze andstatistically process user and customer behavior. Web Analyticsespecially refers to the use of data collected from a Website todetermine which aspects of the Website work towards thebusiness objectives. This paper provides a web metrics basedapproach that can be used to analyze and improve web usagepattern. We define a set of 15 web metrics that can play animportant role in understanding web visitors behavior andprovide suggestion how these metrics can help in making awebsite more popular. We describe the approach by consideringa case study of the website gndu.ac.in for the data collected overa period of five months.


Introduction
Lord Kelvin said "If you cannot measure it, you can not improve it". This fact also applies to analysis of web usage pattern. The more efficient a site is the better the design is since it will take up less of the users' time. If the website design is more efficient there will be less steps for the users to go through to accomplish tasks and thus less things to remember, making it easier for them to reestablish proficiency even after a long period of not having visited the web. Requirement for heavy traffic to a website is that the website has to be simple and efficient.
In this paper we provide a metrics based approach that can be used to analyze and improve web usage pattern. We first describe what is a web analytics process followed by metrics concern with web. Then we provide a metrics based approach for web usage analysis. We describe the approach by considering a case study of the website gndu.ac.in for the data collected over a period of five months.

Web Analytics Process
Web analytics is the process of analyzing the behavior of visitors to a Web site. The use of Web analytics is said to enable a business to attract more visitors, retain or attract new customers for goods or services, or to increase the dollar volume each customer spends.
Web Analytics is the study of the behavior of Website visitors. In a commercial context, Web Analytics especially refers to the use of data collected from a Website to determine which aspects of the Website work towards the business objectives; for example, which landing pages encourage people to make a purchase.
Web analytics is a tool for measuring website traffic and can be used as a tool for business research and market research. Web analytics applications can also help companies measure the results of traditional print advertising campaigns. It helps one to estimate how the traffic to the website changed after the launch of a new advertising campaign. Web analytics provides data on the number of visitors, page views etc to gauge the popularity of the sites which will help to do the market research [1].

Web Analytics Process
The objective of Web Analytics is to understand and improve the experience of online customers, while increasing revenues for online businesses. Among other techniques, this can be done by studying the ways customers navigate a website. According to the Web Analytics Association, the official definition of Web Analytics is "the measurement, collection, analysis and reporting of Internet data for the purposes of understanding and optimizing Web usage." Web Analytics is not a technology to produce reports; it is a process that proposes a virtuous cycle for website optimization.
As illustrated in Figure 1 below, the Web analytics process is composed of the following phases [2]: I. Application users use the Web application; Information about this usage is gathered using data gathering methods.
II. Web usage data is analyzed by the Web analytics tool to produce reports; the tool is configured and administered by the Web analytics tool administrator.
III. Developers analyze these reports to define Web application updates.
Updated Web application is tested and deployed.

Analysis process
The analysis and report generation process has following four steps ( Figure 2): I. Gather Web usage data.
II. Parse data, and eventually store the data to a database and retrieve previously parsed data from the database.
III. Analyze data, and eventually store the results.
IV. Generate and distribute the reports.

Different Web analytics Tools
AWStats and Webalizer are the two most common statistics packages available through shared hosting services. Both are fairly basic, offering information about visits over time, mostvisited pages, referrers, search strings, and some data about visitor's browsers and locations. Though Webalizer is a bit more popular, AWStats's reports are generally considered somewhat easier to understand [3].
Google Analytics is a class by itself. It offers substantially more functionality than the basic tools above. Unlike tools such as Webalizer or AWStats, Google Analytics is needed to be installed on site, which involves pasting a chuck of HTM L (provided by Google) into every page. This obviously requires a bit of HTM L know-how, but if it is known what is to done, shouldn't require too much effort. Depending on the size of site and how it's set up, installing the Google Analytics code might take anywhere from a couple of minutes to a few hours.

A Metrics based approach to analyze & improve web usage patte rn
In this section we provide a metrics base approach that can be used to analyze and improve web usage pattern. We describe the approach by considering a case study of the website gndu.ac.in (Guru Nanak Dev University, Amritsar) for the data collected over a period of five months from 17 march 2011 to November 2010 and provide suggestions so that these web metrics can be effectively used to analyze web usage pattern and therefore can help in improving it. The data is generated by Webalizer

Number of hits per month
The number of hits per month received by a website is used in analyzing the usage pattern of the website during various months in a year. From figure 3 it is clearly observed that the month of February 2011 received maximum number of hits. The month received total of 747813 hits. Rise in the number of visitors (possibly students) is experienced as more visitors are interested in session end announcements

Number of pages viewed per month
Pages are those URLs that would be considered the actual page being requested, and not all of the individual items that make it up (such as graphics and audio clips). On the other hand month of November received least number of pages viewed and the number is 30928. Since this month has minimum number of hits, so it is expected that the number of pages viewed would be least as less visitors are interested in viewing the pages of the website.
Reasons should be traced to find out the months in which minimum number of pages were viewed so the appropriate action can be taken to increase it. There can be decrease in the number of web pages to be viewed by the visitor by making use of flash/AJAX. The increase in online video, means fewer page views are counted, even though the same amount of content is being looked at.

Number of visits per month
In

Number of sites per month
Site is the number of unique IP addresses/hostnames that made requests to the server. Figure 3 shows that the maximum number of sites are in the month of February 2011. This number gives an idea that the most common visitors to the website are in the month of February 2011. M onth of march received minimum number of sites and the number is 7301. This metrics may also help to increase traffic after finding the maximum and minimum number of sites from the data provided.

Total Number of KBytes downloaded per month
This metrics is used to show the amount of data that was transfered between the server and the remote machine, based on the data found in the server log. From figure 3 it is observed that maximum number of downloading is done in the month of December 2010. The downloading total is 7891702. Although the month of December does not received maximum number of hits, but still its has maximum downloading. So, the observation confirms that the number of hits and total downloading are not related.
On the other hand month of November 2010 received minimum number of total down loading as in this month less announcements and uploading is made by the university. The count for this month is 2605278. So, files must be up to date in the months that have maximum downloading.

Total Unique Sites
Total Unique Sites is each request made to a server coming from a unique "site," referenced by a name or, ultimately, an IP address. As shown in table 1 the count for total unique sites for the month of M arch 2011 is 7522. It is important to understand that this does NOT equate to the number of unique "real people" visiting site, which is impossible to

Total URLs
This metrics can be used to ascertain the most popular referring URLs to the site (i.e. pages which provide a direct link to the web site. These pages can include search engines). Further, the data indicates the total number of referring or inbound links for the summary period.

Total User Agents
User Agents are a fancy name for browsers. Netscape, Opera, Konqueror, etc. are all User Agents, and each reports itself in a unique way to server. This metrics can be used to ascertain the most popular user agents accessing the site, the operating system/platform on which the software is running (and sometimes the language).
Data collected shows that the highest hits are for the user agent : M ozilla/4.0+(compatible;+M SIE+6.0;+Windows+NT+5.1;+SV 1) With a count of 77303 hits and the least one is

Number of Entry pages
Entry pages are those pages that were the first requested in a visit. This metrics provides information regarding most and least popular entry page of a website. From the data gathered we have to list the most popular entry pages. So that instead of analyzing and improving only the home page, this page may also be improved because the reality is that many people will arrive deep into website through search engines.

Number of Exit pages
Exit pages are those pages that were the last requested in a visit. The analysis of usage pattern is to be done to find the most exiting page in the website. Table 5 list out /Index.asp to be the most exiting page in the website with a count of 333291 hits. On the other hand /about/amritsar.asp is the least exiting page with a count of 471 hits. It reveals that this pages on website is driving people away. Improvement over this page is required so that it does not be marked as the most exiting page in the website.

Number of search strings
Search Strings are obtained from examining the referrer string and looking for known patterns from various search engines. According to usage pattern for the given data we have to find out most popular search string, so that it can be used to inform keyword strategies for the web site. Table 6 shows that the most popular search sting for the website is gndu with a count of 32 hits whereas search string gndu.ac.co is the least one with a count of 3 hits.

Number of Countries
Countries are determined based on the top level domain of the requesting site. This metrics is used in finding out the place on the earth from where the website is  Table 7 is used in finding out accessed. The topmost country is India from where the maximum access to the website is made. The number of hits for India is 77341. The other

Daily access
This metrics is very important from web usage pattern view as it provides information that when in the day access to the website was high. The analysis of the data collected of the usage pattern shows a very high usage pattern at the beginning of the daily service i.e. at 8-9 AM period which is understandable as most of the users start using internet immediately on arriving at their respective offices. However, the usage of the site decreases with the passage of time during the entire day especially in the month of Feb 2011. However there is an appreciable volume of access of the site before the peak hours (8-9 AM ) in the months of Dec 2010 and Jan 2011. There is an unusual high accessibility during 5-8 hours in the months of Dec 2010 and Feb 2011 period which could be attributed to unusual website access which is not explainable.

Conclusion and Future Work
The importance of web metrics has dramatically increased due to extremely fast growth in Internet technology. The web metrics is directly related to the purpose of the good website design. Website with poor traffic can easily destroy the purpose of using web metrics. In this paper with the case study of website gndu.ac.in we proposed a set of 15 web metrics that can be used to increase the popularity of the website by increasing its traffic.
Although we have tried to find out the web metrics that are concern with web usage pattern, but still there is a need to find out the more and new ones. Other metrics that can play role are Key Performance Indicator, Search engine optimization, IP address owner etc.