The new secrets of Google analytics--How to define visit

　　

"Preface" Visit this metric is the cornerstone of web analytics. But even with such a basic metric, Google Analytics's definition of it is not entirely immutable. In order to adapt to new browser changes and people's access to Web site habits, Google analytics is even evolving in basic metrics. Oh, this may also be the analytics of Google, she has reached this height, but also faster than others to climb.

What are you going to learn in this article:

GA defines significant changes in visit;

How GA deals with the definition of visit under special circumstances;

Visit the change of associated attributes (past and present);

Why most pages are visit much smaller than the number of visitors to these pages (visitor).

　Body

If the title of this article I just write "Google analytics How to define visit", you will not read it?

But maybe things are more complicated than you think, so hopefully you can read on. Of course, if you are not so aware of a measure of such an important foundation, then it is recommended that you read this article first: the most basic measure of web Analytics (1)--visit.

　Google Analytics Significant changes in statistical visit

We all know visit's two formulas-"No movement for 30 minutes" and "Don't close the browser". It means that if you browse a Web site for 30 minutes without opening a new page or some action (these action needs to be monitored by GA code, such as event tracking), Then your subsequent visit to this site will be recorded as a new visit. Similarly, if the browser is closed and then the browser is opened to browse the site, GA will also think that a new visit has occurred.

Now, you don't have to remember these two formulas, because the Google Analytics definition of visit has changed a lot.

　　

In fact, this formula must now be replaced by the phrase "30 minutes of quiet" and "a change of source to go in". Why change these two formulas, let's take a look at Google's new definition of visit (if you know English, and you have a wall, then you can read this article: http://analytics.blogspot.com/2011/08/ update-to-sessions-in-google-analytics.html):

The interval between two adjacent pageviews of the same visitor, if more than 30 minutes, is recorded as a new visit.

At the end of the day, the continuous browsing behavior is automatically cut into two visit.

A visitor who has replaced the site with a different source from the original source will also be credited with a new visit, even if it is replaced within 30 minutes. But replace with direct flow except.

What effect does this change have on visit statistics?

For the first to second case, this is industry-accepted, and the GA update visit definition does not change this. The third is the most significant change.

What changes did this change bring to the count of visit? Please read the following words carefully, perhaps some lingering question marks can be solved.

In the old definition, the increase in visit is not in fact related to the source of traffic. In the following scenario, there is a significant difference between the old and new definitions of visit.

Scene One:

Steven visits my website www.chinawebanalytics.cn often through search engines. On this day, he searched "website analysis" on Baidu, then clicked the search result to enter my website, browsed a bit on the website, did not seem to find the article that he once remembers not to read. So, he immediately returned to Baidu, search "E-commerce key digital optimization", found this article he wants to read: http://www.chinawebanalytics.cn/kpi_optimization_part2/. This process occurs within 30 minutes and the browser is not closed. Excuse me, Steven's This series of behavior for my site to bring how many visit?

　　

For this scenario, the new and old visit are defined differently, and the visit calculations are very different.

In the old definition, only the browser is closed or the interval between two PageView is more than 30 minutes to calculate the new visit, so in the above scenario, visit will not increase, still only 1.

This time, produced a trouble, for the flow of statistics, a bit tricky. Our report on traffic sources, two search terms (the previous one is "Web Analytics", the latter is "E-commerce key digital optimization") will be how to statistics it? This time Google Analytics has several options:

Give up the second word of the record, the added value of this visit into the previous keyword "website analysis";

Give up the first keyword of the record, that is, the second keyword washed out the first, the increase of the visit to "E-commerce Key digital optimization" on;

Two keywords are recorded, and they bring to the site of the number of visit increased 1--This means that the traffic source report of the source of the total number of visit numbers, the total number will be greater than the total number of visit site;

Record the first keyword "website analysis" visit increased by 1, while also recording the second keyword "E-commerce key digital optimization", but visit does not increase, if the second keyword did not bring any traffic, then it will appear in the report, but its visit data is "0".

Students, what do you think of GA? The answer is fourth. Interestingly, that's why we might find some traffic source visit=0 in the GA report. If you've ever seen such a report, you'll think that GA must have made a mistake. Not the GA error, but the visit defined by the mechanism of the Black Dragon.

But in the new definition, the change of flow source will cause the increase of visit. The above scene, because traffic source changes (from one keyword to another), so the visit will increase by 1. In the traffic source report caused visit=0 Trouble also does not exist, two keywords each for the site to take visit all add one.

Now the Google Analytics report, under no circumstances will appear visit=0 traffic source. But there was a report that was there. On the Han, this definition changes will change the historical data in Google Analytics, so you're looking for visit=0 in the past and there's nowhere to go.

Scene Two:

Now let's look at another scene where the visit change. Still Steven, he opened my website, and then the phone rang, he subconsciously turned off the browser, 5 minutes later, the phone ended, he opened the browser and re-enter the Web site of my website-- Www.chinawebanalytics.cn in the browser's address bar, and then continue browsing the page. In this process, will create a few visit for my website?

In this scenario, in the old visit definition, visit will increase by 1, that is, the entire process will have 2 visit. Because closing the browser will cause a temporary cookie for the statistic visit to expire (the so-called temporary cookie, which is the cookie that expires immediately as the browser closes).

However, in the new visit definition, visit does not increase by 1 because there is no case of browser shutdown in the definition. In fact, the previous temporary cookie still exists and will still expire, but its content does not change.

Another important new rule for visit--it can also be used to describe the page

Another notable change for visit is that Google Analytics has given up its past adherence to certain metrics in relation to the page category and site category. A bit of a mouthful, the so-called uncorrelated, meaning that some metrics are only used to describe the site level, not the page. Visit is typically such a metric that in the past, visit will not be used to describe a page, so you can't find the visit metric in the middle of the all Pages report, even if the custom report doesn't get them together. There is only one page-level metric similar to visit--unique PageView.

This definition is actually very rigorous because visit is used to describe a process, and for individual pages, it is a step in the process, so it is problematic to describe a page with visit. But the unique pageview can--a visit visited a Page 3 times, then the page PageView increased by 3, but the unique PageView is only 1. In the following illustration, a visit visited two pages, A and B, where a visited 5 times and B visited 3 times (because there was a refresh, so there was a->a and b->b access), the unique pageview=2 in the process.

　　

But now, part of this rule is history. In the new rule, visit can be used to describe the page. We've found clues from GA's custom report.

　　

As we can see in our custom reports, visits this metric can be completely associated with page this dimension. However, in order to avoid a sudden unacceptable, visits still does not appear in the default page report, such as the All Pages report.

Associating the visits with the page, and the increase in the previous section of the change--visits also affected by the change in the flow source, the two significant changes to the visit such a fundamental measure, the most important reason being that the GA multi-channel Traffic Channel Attribution report) makes it easier to process attribution (especially thanks to the good friend Robin Li's corrections in this place). Another reason is that Google Analytics wants to keep the visit metric better aligned with the session, to avoid people not understanding it. For example, in the case of the visit=0 mentioned above, people will mistakenly think that GA is bad, and now the new definition will no longer have such a misunderstanding happening.

why is visit less than the unique Visitor?

The attentive reader will try what I said above, if you do experiment (I must support you to do this), then when you generate the custom report above, you will find a strange phenomenon-most pages have a smaller number of visits than their unique visitors. And some small, some small, there seems to be no law. The following figure:

　　

is there a problem with Google Analytics?

Frankly speaking, Google Analytics has no bugs on the metric, and any number has a reason, even if the numbers "don't look reasonable".

Why is this happening?

Originally, because visit is to describe a process, so many visit will go through different pages. Under the new definition, visit can be associated with a specific page, then, when a visit through multiple pages, the visit of these pages are increased by 1?

Not so.

Originally, in the process of visit access to multiple pages, GA rules only add visit to visit pass the first page, and after the page no longer assigned to them visit, that is, after the page visit does not increase.

　　

However, the Unique visitor is not the same value. The GA assigns the addition of UV to the pages visited by each visitor. Therefore, the number of UV is coincident (if you add the total number of Uvs of each page, is much larger than the total number of Uvs), but visit did not coincide, you put the visits number of each page plus total, the total number just equal to the total site visits. No wonder there are cases where visits is less than the unique visitors.

Now we look back at the example on the blue a page and the Red B page above, and the situation does change, as follows:

　　

If you think about it again, you might think that, since visit does not assign value equally to every page that the visit passes, then theoretically there may be a page visit the visit value is 0! If you think like that, congratulations, Your earnest spirit ensures that you can become an excellent analyst. In fact, I think we are not difficult to find those visit=0 pages, just do a visit from small to large sort of good.

　　

Conclusion

GA is not immutable, its functional changes make us refreshing, and in metric changes, this is not the first time. In any case, these changes are moving in a way that allows users to better understand. However, since the definition of a metric needs to be very rigorous, far-reaching, so when you find something unusual, don't panic, it's probably just a matter of GA handling data rules, not your code adding wrong, or GA not monitoring your data well.

Finally, are there any problems with the friends? There are other strange phenomena in GA, please leave a message, we and all the reader friends can brainstorm and participate in the discussion. Thank you!

Original address: http://www.chinawebanalytics.cn/how-ga-define-visits/