- Allow users to more easily find the information you need
An article in the blog before - to optimize web site navigation design , how to evaluate the site navigation and optimization based on analysis. But later found missing is a very useful feature in Google Analytics - Navigation Summary, literal translation is "Navigation Summary", but it seems to call on the "downstream" on the page (Baidu statistics, take over first borrow) The analysis is more appropriate. It can be a good analysis of the site navigation (to put it on the direct view is that the navigation function for valid clicks or operating) Here's to introduce this feature.
Better measure of navigation to achieve the degree of
Look at my article topics recommended the navigation index page in the GA Navigation Summary Report (this feature is the Top the Content tab in the Content module below):
From the chart you can see the number of times the page has been viewed (marked in the figure), the ratio of the number of Visits is marked in the access to this page from outside the station began a visit to the site (Figure 2), how much percentage of Visits from Visits from the site within the page to jump label to this page (Figure 3), the proportion of the number of Visits to leave after the page to browse the site marked (Figure 4), what percentage of the page to enter the site other pages (Figure marked 5,4 and 5 of the now seemingly data is a problem); while listing the top 10 upstream of the site internal page (the page before viewing the page the user, marked in the figure 6) and downstream page next page (viewed immediately after viewing the page, marked in the figure 7), and their respective percentage. It should be noted that sometimes appear in the upper reaches of the page and downstream of the page and select the same page as the URI address, such as you select the home page (/) to analyze the upper reaches of the page and downstream of the page Home address (/), this is the main refresh operation, the GA will page refresh statistics to Pageviews which.
The above functions, we no longer need by the departure of the navigation page (Exit Rate) to roughly estimate the number of Visits to stay in the website and may click on the links to the navigation page. And we can not only see the Navigation Summary immediately leave the number of Visits from the navigation page, and by the downstream analysis of the navigation page page more accurately measure the effective CTR, excluding refreshed, returned or transferred to other non-navigation list page operation will add those navigation page link click-through rate (% Clicks) is the navigation page click conversion (CTR), which is the navigation function of the index. The above table, for example, rule out the Home (/), the page refreshes (/ featured-topics /) and jump to the page in the non-navigation page (/ about / / site-map /, etc.) for these clicks, the remaining the navigation function of the effective CTR to the sum of the navigation function, Unfortunately, the upstream and downstream of the GA on the page only shows the top 10.
Above the previous content on the optimization of the navigation design, in fact, the page on the downstream analysis is a very effective site analysis methods, not only can be used to analyze the navigation to achieve the degree described below it is another kind of application - - content recommendation effect.
Related to the content of the site recommended
A prior blog article - optimized website information architecture is described in most of the site may have been based on the tree structure to the acquisition or construction, but the original tree structure there is a problem is the leaf node (or site is no direct link between the contents of the node), users can not from the bottom of the contents page to jump directly to another low-level content page, you need to return to the home or center navigation index page to enter the other content pages from the article tree chart is also reflected in the bottom of the page are not directly connected to the lines. So many sites are provided at the end of the content or sidebar the recommendation of the relevant content, such as Amazon, Taobao e-commerce website product page will have the same class with the price of the product recommended, or the user purchased the product but also purchased a product recommendation, the watercress on the books, music, movies page also provides a recommendation of the relevant content.
Many of these features are based on the content of the algorithm to achieve the previous article - up marketing, cross-marketing with related recommendations based on user behavior associated with recommended method. In fact, many blog has a similar function, that is, the relevant articles of the end of each article Below is a list of articles RFM analysis of e-commerce site for my blog this article:
I am using the Wordpress Plugin - Yet Another Related Posts Plugin to implement this feature, according to the introduction of plug-ins, it is through the calculation of the article title, text, tags, and classification will select the top few shows on the page. This feature is great, it opened up the channel between the article page, perhaps after reading the article, the user would like to browse related articles, related content is recommended to provide a good way, users do not need to come back to retire content to the search page, simply click on the line, to help users to more easily, quickly navigate to the information you want to find.
Content Recommend effects analysis
Site Suggest a feature are automatically generated through machine algorithms, so on some level, the algorithm will exist advantages and disadvantages, we need analysis to assess the function to achieve the effect of algorithm be optimized. Analysis based on user browsing behavior is the most effective way to achieve results of the evaluation function, so the site analysis, there has been useless, the Navigation Summary described above Google Analytics is very suitable for the effect of the tool used to analyze the content recommended . Or e-commerce website RFM analysis of upstream and downstream of the article page analysis as an example to see in the end how I use this plug-in effect:
From the upstream and downstream of the page list the proportion of those from the content page and go the other content pages, of which the highest proportion of the inflows and outflows of which pages recommended ranking list, and then the site to compare, so you can correlation with the user in the eyes of the correlation of the reaction site features are the same, in order to test the functionality to achieve the effect.
Algorithm to adjust the conversion ratio, which measure the content recommended in the algorithm adjusted really been optimized, you can also use this method before and after the test algorithm to adjust the upstream and downstream pages. And we do is through continuous optimization of the algorithm recommended by the content site content ranking and content expected correlation to try to reach an agreement, this will be displayed on the page the user wants to find the content of to meet the needs of users.
Need to pay attention to several issues:
- Perhaps a content page which will be more than a recommendation module, or there will be many links to other content pages provided in the report of the GA is the sum of all inflows and outflows, so if evaluation of a recommendation module effects need to distinguish between the links in the module, maybe add a URL parameter would be a solution.
- Note that the data time interval, and website content changes brought about by the changes recommended by the relevant content. Google Analytics, the default time interval is one month before you can choose a suitable interval to carry out the recommended impact analysis and comparison, note that the site's content update content.
- Recommended some content is not a two-way, such as recommended in the page to buy MP3 earplugs, buy earplugs page might not recommend the MP3, it is sometimes necessary to separate the upper reaches of the page and downstream pages, pay attention to conversion direction.
The above two examples, I think the Navigation Summary function of Google Analytics, you not also think of more applications are welcome to share your point of view.
»In this paper, the BY-NC-SA agreement, reproduced please specify source: website data analysis » optimization recommended "
Related Articles:
Write good · · frequently come to your site to learn knowledge ·
Found ranked good Guizhan! The support!
Very good, good enough. Page is the search engine to consider the issue, but also the problem of seo. Good idea.
If your site is updated affect the content recommended by the judge recommended that according to the way the text content is not accurate. Cause it does not click, instead of the recommended content may recommend the contents of the article just on the line.
Yet Another Related Posts where there is an option recommended only for articles published prior to the current article. The Show only previous the posts
_AT_ yoyo : This is important, content publishing time differences may affect the correlation analysis data, it is necessary to select the appropriate time interval may not necessarily need to select the content released before, such as selecting the one week of data, and all involved in the analysis of the week already exist, then that result is valid.
Joegh : makes sense to you all the articles I have read ... learned a lot, looking forward to see more statistics and mining in the site analysis. Thank you!
Can it be understood, Google Analytics is equivalent to the underlying data of the data warehouse, and two instances in the article is equivalent to data applications?
The the _AT_ tjdyw : Google Analytics, should not be understood as the underlying data of the data warehouse, GA is a data show platform, if the contact of non keep the data warehouse architecture, should belong to the scope of the data mart (Data Mart).
Will the Navigation Summary Entrances previous pages, the Next pages are several indicators is how to achieve? I do experimental data, the GA report displayed these indicators, so very sincerely ask you about, but also look enlighten
Anran : Hello, not quite understand the meaning of "no data experiment Navigation Summary of this function, I saw the old and new version of the GA should have, in the content module page statistics.
The _AT_ joegh : Thank you, have found. Your blog is very well written, very carefully, and hope to continue to share more stuff
The @ joegh Hello, this function is for the analysis of the upstream and downstream of a page, then do the vast amounts of content page of the website is also a sample of some of the contents of pages on the downstream analysis? If so, how to be able to judge the content to do good or bad?
@ Amauri : website content modules recommended algorithm recommended is relatively fixed, do not frequent automatic refresh sample for analysis is, of course, the best nor the artificial one to see, based on the contents of the sample statistics on the downstream page is the probability or proportion of the content. More effective approach is the recommendation module links to add a specific parameter, This makes it easier to distinguish.