Case study: Number of new visits to data black holes and average website

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

Today, the GA is very slow to open, water the flowers poured into the coffee a lot of action down, and finally sit at the table, the report is still loading. While waiting to take a glance at the control Panel on a large number of values, see%new visits this value, startled a bit, and then a closer look, 56%! when the number of new site visits to reach this level?

I was surprised because it was a 99%-based website. For a bidding-oriented site, this ratio iffy, just look at this data is tantamount to tell the boss that you bid money, nearly half of the water floated. Why not? Advertising has not significantly reduced, and UV has not significantly increased. Almost half of them are returning visitors.

According to brain experience, generally do promotional marketing site, the percentage of new visits must not be less than 20%-30%, because the search engine top three is so conspicuous, it is easy to become a return visit of the convenient entrance, whether you pay for it! Paid to pay a return visit, the boss will certainly shake his head, Bidding colleagues will certainly be crazy. The next step is definitely to adjust the bidding strategy.

Can you report that? Wait!

The intuition tells me that maybe the data itself is out of the question. This is an average data! So before you believe this data, it's best to verify the credibility of the data. Open the visitors report and see the ratio of the new interviewee to the caller. is still customary to see the ratio, 80%, which is odd, from 56% to 80%, the percentage of new visits and the difference between the number of Web sites how can this be so large, is GA also contaminated with the statistical office of the problem?

Obviously this is an unusual situation, or this is an unknown situation, carry forward the spirit of Sherlock Holmes, we come to the bottom of the. What is the difference between the new visit ratio and the average?

1. Date Range

First you need to determine the date range. See if the date range selected by the two metrics is consistent. Especially the default time range for site averages. At present there are two kinds of statements, one is the default is January ago, one is the tacit view of the current time period. I think the term "current period" is quite plausible.

2. Whether there is interference of the filter

If the filter has been set before, excluding a certain part of the flow, then compared to the whole station, filtered data and filtered data is not the same. This is often easily overlooked.

3. Choice of Advanced segment groups

The principle of the percent effect of the Advanced Subdivision group is the same as that of the filter, but this is easier to spot. After all, it's just above the date range.

4. Use the footer of the inline filter

This is not commonly used, but it is also possible to search for a specific source to produce data differences. For one reason.

5. Use of custom Reports

The metrics for customizing the report, after filtering, will be slightly different from the regular report. The new GA custom report is added directly to the filter, not excluding this factor.

Based on the above 5 points preliminary view of the results are as follows:

To my surprise, even if I pulled out a completely pure, pure source data that was pure than the mountain springs, the difference between the new number of visits and the average web site was still great. I began to wonder if I had misunderstood the data and thought about the definition of the percentage of new visitors.

The official definition is simply no longer simple: "The percentage of visits to users who have never visited your site before." "To think such a definition, have been staring at absolutely will look silly." Fortunately, to understand the data indicators, the blue whale's commonly used indicators and interpretation of GA provides us with a good reading model, cited here.

Several ways to interpret data:

1, the meaning of the data and the cause

2, and historical data comparison

3, bring the data into the trend

4, is this a summary of data?

5, is this an average data?

From 1 to 5, let's look at the following item:

One. The meaning and causes of the new interview data:

Definition of indicators

New Access percentage: The percentage of users who have never visited your site before.

The real meaning of the index

As the name suggests, this is the new visit as a percentage of all visit. The new visit should be validated from Ip,cookies and other n-data.

The calculation method of the index

%new Visits = New visits/visits

Factors that may have an impact on the indicators

New definition: Is the number of times the GA cookie was first created to user Brower. In other words, it is recorded in the Utma in the cookie, the time is 2 years; the details in the definition are visits, not visitor.

Ii. comparison of new visits and historical data

By definition, if the data is not volatile, the average number of sites in the same time period should be the same as the percentage of new visits.

When comparing this month's and last month's data in a date, this week's trend is not much different from last week's, with a total of around 80%.

Iii. review of the trend of the new visit ratio

In the chart mode at the top left of the chart, select the indicator as the "percentage of new hits" and compare it to the average site. You can see that the basic is a smooth line, and the site averages coincide. No highs and lows. This is clearly inconsistent with the data we have seen before.

Iv. new Visit ratio-is this an average figure? Is it a summary data?

The new visit ratio is a summary data. The prerequisite for analyzing summary data is that all visitors to the site are identical. But this is obviously impossible. Similarly, the new visit ratio is the average data, the different traffic source visitors have different number of new visits percentage.

To solve the new visit ratio of the indicators of confusion, we look at the average site.

What is the definition of an average website?

I did not find the exact number of new visits to the site average definition. Nigel told me that this should be the date selected above and the average of this time period. Like Avgtimeonpage,avgtimeonsite, I feel that his understanding should be right.

What is the meaning of the green (or red) percentage beside the average site?

Web site average green percentage can be considered as the current period of value and average contrast, calculated values.

When the date range changes to year, month, week, day, the percentages in parentheses also change. Mark the good direction green, and mark the bad direction red. For example, the faster the visitor rises, the more green the ratio, but if the jump rate rises, it means that visitors are not interested in the site, is a very good direction, will be shown as red.

Trend percentage of the specific calculation method is not known, if you know please leave a message to tell me. ***

All the way to the interpretation of the data down, in fact, we have from the historical data, trends and several places to see the doubt, that is, the site average value is not credible, it is possible to fall into the "data black hole."

Let me be certain that the existence of this "data black hole" is a casual discovery:

In the case of an error in the 6 replica profiles of the same Web site, there is a single copy that shows and data trends in line with the basic "normal" data, this copy in the settings and other replicas no difference, the only difference is that it enabled the Pacific Time GMT-7 timezone, The other is the Beijing time updated timezone.

It's really black humor if the difference in time zones can lead to such a big difference as the average population. So, I can finally confirm that the data of the new visit has fallen into the "data black hole". Long breath ~ Wrong data kills people.

Original link: http://yuli.in/question/precentnew-visits/