Website revision results should be how to analyze

Before writing an article, used to share with you.

Before analyzing the results, it is necessary to revisit the reasons for the revision.

In general, there are two scenarios for site revisions:

1 because of the expansion of the content of the site itself led to the current site architecture can not carry more content,

2. Web site based on changes in user needs lead to site transformation.

Of course, can not rule out purely for the appearance of the site to make a revision of the decision. (This is not thankless)

What is the revised expectations?

Of course, this is also an important question that needs to be faced before the website is revised. What is the value that can be brought to the website after the revision is finished?

How to get the result

Mainly from three aspects:

1. Collect user feedback by placing an online survey of the new version of the page.

2. Receive direct user feedback through the service department.

3. The new page of the relevant data analysis to get revised results

The following focus on page data analysis methods:

Because the revised website will certainly cause users of the site a variety of reactions or even rebound? The specific performance of these reactions is the data fluctuations, analysis of these data can help us understand the user's reaction, and how to reduce these fluctuations as smooth as possible , Make revision a better result.

First, take the data

First of all, take the revised version of the PV online before and after a few weeks, Session, and revision of the page hop (step by step to the next step) or multi-step path, personally feel that taking the data from two weeks to one month before and after the appropriate.

Second, contrast

Revamped pages before and after the line PV, Session can understand the macro before and after revision, page views, viscosity changes.

Pivot the data (PV, session) to do the data perspective, when the data is recommended Url categories and subcategories are taken out, at least to take a small class, this perspective can be done by analyzing the small class to get the various blocks Visit volatility, analysis of the specific volatility of the block clearly Url, so that it has focused on analysis, with less. Analyze the proportion of PV and Session before and after on-line [(PV / total PV in each subcategory) session).

Then come to the promotion rate [(on-line accounting - on-line accounting) / on-line accounting].

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

Of course, you can also analyze the absolute value (on-line PV-PV on the line), the absolute value of intuitive but can not do vertical contrast.

To sum up: emphasis on the rate of improvement is not enough light, successful revision requires more in-depth analysis of the user's activity, conversion rate, etc. to illustrate the problem.