How to analyze the results of Web site revision to explain the page data analysis method

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

Write an article before, bring to share with you.

There is a need to revisit the results before the analysis.

In general, there are two types of Web site revisions:

1. Because the content expansion of the website itself leads to the current website structure can not carry more content,

2. Web site transformation based on changes in user requirements.

Of course, can not rule out purely for the appearance of the site and make a revised decision. (This relatively thankless)

Read the full text

What is the revision expectation?

Of course, this is also in the website before the need to face an important question: hope that after the completion of the revision of the site to bring what value?

How to get the results

Can be mainly from three aspects

Collect the user's active feedback by arranging the online research form of the new page.

Direct user feedback through the service department.

Analyze the data of the new page and get the result of the revision

———————————— – Gorgeous split line —————————————

The following focuses on the page data analysis method

Because the revised site will certainly cause users of the site a variety of reactions or even rebound? And these reactions are specific to the fluctuations in the data, the analysis of these data can help us understand the user's response, and how to reduce these fluctuations as gently as possible, so that the revision towards better results.

I. Access to Data

First take the revised page before and after a few weeks of PV, session, and the revision of the page (step by step path) or multi-step path, the individual feel to take the line before and after two weeks to one months of data appropriate.

Ii. comparison

The revised page before and after the PV, session can be macroscopic understanding of the revision, page access volume, viscosity changes.

For a jump data (PV, session) to do the data perspective, when you take the data, it is recommended that the large categories of URLs and small classes are taken out, at least the small class to take, so do perspective can be analyzed by small classes, access to the various blocks of the block analysis of the specific URL, this analysis has focused on, with less effort. Analysis of the PV and session before and after the proportion of the session [(each small category pv/total PV) session of the same]

Then get the rate of ascension [(On-line post-occupancy ratio)/pre-occupancy ratio]

This analysis of the relative value of comparison can explain the problem.

Of course, can also be analyzed absolute value (on-line after the pv-on the line before PV), absolute value intuitive but can not do vertical contrast.

Summary: Emphasis on light to see a jump rate is not enough, the successful revision needs more in-depth analysis, the user's active degree, conversion rate is to explain the problem.

Original: http://ued.alimama.com/posts/451