Behind the network data--analysis index of the network log

The commonly used quantitative analysis is the questionnaire survey, this can collect the user to the product subjective feedback, its result is affected by the questionnaire question, cannot completely objectively reflect the user how to use the product, they encountered in the actual environment what problem. For the quantitative analysis of the website, the log file of the Web server can reflect the user's current experience, explain the deep characteristic of the behavior, and can improve the product more effectively.

Blog can help us answer a lot of questions, for example, when users browse the Web site, what section of the site is interested in, how to understand the site, how many users will be converted to duplicate users, find the point of interest on the site of the path, how to optimize the use of the process, improve the user experience, and so on. To systematically analyze the log, get valuable user feedback, mainly considering aggregation metrics, session based statistics, user based statistics, and click Flow Analysis four aspects.
1. Aggregation Metrics

It can be understood that a large amount of Web site data is merged and analyzed. The following is combined with the log data of a tourist forum to illustrate common aggregation metrics. In particular, this forum is fictitious, and the data is invented to illustrate the concept.

(1) The amount of browsing on the website.

Compared with the browsing volume of the same time period, the trend of users ' attention can be obtained. Figure 1 shows a tourism forum from June 2008 to December 2010 Browse volume changes, of which around June 09 Browse volume suddenly increased, the December 09 gradually flattened, the changes in the curve may be related to the Forum's marketing tools, design and so on, so can witness the effect of taking measures.

Figure 1 views of a tourism forum

(2) The distribution of browsing volume in each time period of the day.

Figure 2 can be used to determine the user mainly in the rest time browsing the forum, so the forum should highlight the relaxed content of leisure.

Figure 2 The amount of browsing time of a travel forum in one day

(3) The browsing distribution of each section of the website.

Can analyze specific plate, single page, similar page group browsing distribution, judge the user's point of interest. Figure 3 illustrates that forum users are primarily interested in Southeast Asia and Japan.

Figure 3 Browsing ratio of the foreign tourist edition of the Forum

(4) Operating system and browser proportions.

Easy to adapt the website to the operating system and the browser. Figure 4 shows that the operating system used by the user is dominated by Windows, and Figure 5 shows that IE is the majority of the browser, and Firefox and Chrome are also relatively large users. Therefore, the design or revision, the need to focus on meeting the Windows model, the display effect is mainly to ensure the compatibility of IE, Firefox and Chrome.

Figure 4 The percentage of users who use the operating system for this forum

Figure 5 The percentage of users of the forum using browsers

(5) Recommended link site.

The previous page that the user browsed before opening the page. Learn about the recommended link site, easy to know how users enter the site, understand the use of the situation. With this forum for example, nearly 30% of users from the search engine to enter the site, about 60% from the collection or address bar input directly into the forum, which shows that the user loyalty is relatively high, already has a stable user base.

(6) Search engine recommendations and keywords.

If the recommended link from the search engine, which contains keywords can help determine the keyword marketing plan, if it is the site's internal search keywords, it will help to recommend content to users. For example, the most popular keywords of the travel forum internal search are "Raiders", "Self driving", "honeymoon", "Hainan" and "51", what users pay attention to, this need to provide a lot of reference.

2. Session-based statistics

A session is an interaction of information with a Web site through a link.

(1) The average number of pages visited per session.

A typical session can explain the breadth of the user's access to the site. For example, the search class site is small, the user's goal is to find information, to achieve the goal will terminate, so support the main task required less pages, simple framework. However, if the E-commerce site, the use of the purpose is to browse, complete the transaction, the process is complex, more natural access page, so need different guidelines to support the entire process.

(2) Average duration of the session.

From the duration of the length of time to determine the user's point of interest, you can also determine what is the transition page, which is the content page. Figure 6 is the process of querying the surrounding tourist attractions in Beijing, and the contents of the small box are the pages viewed in this step and the length of the view. It is possible to determine from time to time that the list page and the local tourist version belong to the transition page, but the same is the Transition page, which takes a different amount of time, perhaps because there is no obvious entrance to the recommended attraction, so the search is longer.

Figure 6 The process of searching around Beijing tourist attractions

3. User-based statistics

The main use of cookies (web site stored in the user's local terminal data) to get.

(1) Number of visits.

Mainly used to measure the loyalty of users and the trust of the site, the relative number of visits is proportional to the two.

(2) Frequency of access.

Can be used to determine how often the site content is updated. Figure 7 shows the statistics on the frequency of updates and visits of some three sections of the above Tourism Forum, the first version of the update is too slow, you can get the information users want through the interview, increase the frequency of updates; second edition updates and Access synchronization, the release of information to allow users to accept; The third edition is too fast, many content is buried, is a thankless task.

Fig. 7 The frequency of update and visit of three sections of tourism forum

(4) Retention rate.

It is the ratio of the number of users returning to a certain time period and the total number of users, and the length of the time period needs to be decided according to the location of the website itself. It is also explained by the travel forum that users usually check the travel information within one months prior to travel, so you can consider the retention rate at monthly intervals. As shown in Figure 8, the forum retains the user's ability to stabilize.

Figure 8 User retention rates for the Tourism forum in 2010

(5) Conversion rate.

The conversion rate is equal to the ratio of users who enter the next page to the user browsing the current page. From the table below you can see that users from the list of tourism to open the posts of fewer users, the proportion of the reply is also very low, so you can recommend valuable stickers, while considering increasing the response to the incentive measures.

4. Click Flow Analysis

That is, the analysis of how users use the site.

Path analysis. The purpose of the forum is to quickly understand the information of the target sights, and the current process can be reduced to the critical path shown in Figure 9. Therefore, consider whether you want to design the transition page "home list" in Figure 6 as the sidebar of the Forum. At the same time, such as "Badaling", "Qinhuangdao", "Haihe" and other places of concern are high, so you can consider the link to the hot spots directly in the list, the travel layout of this transition page also omitted.

Figure 9 The process of searching for a tourist guide

The indicators mentioned above are presented separately and sometimes need to be considered in combination with multiple indicators. For example, you can use the average duration of the session statistics, the average access page, as well as user statistics based on the number of visits, access frequency and other analysis of user loyalty, this is not elaborated here.

Web log analysis can be used to find out what users do on the site, how to do, but why do need to do through interviews, focus groups, usability testing to further understand.

The above is in the Learning site log analysis of the more useful information collated, relatively simple, more in-depth research hope to be able to communicate at any time ~

References: Kuniavsky, M. (2003). Observing the user Experience:a Practitioner ' s Guide to user. San Francisco,ca:morgan Kaufmann.