Common problems with Google analytics (II.)

This is the second of the FAQ series in Google Analytics, thanks to all my friends who share the questions in the mail, I may not be able to answer your questions one at a time, but I will classify all the problems, publish them in the form of articles and share experiences with you. And thanks to the administrators of the blue whale's web analytics notes group. Thank you for your timely answer to the questions in the group. With less gossip, here are five common questions that begin in this article.

　　

　1, why the site of the ads click and the target page of the PV inconsistent?

This is a problem I have recently encountered, first to describe the background of the problem. When analyzing advertising effects, we usually pay attention to the number of times the ads are clicked, that is, the click Number in the Google Analytics Hot zone map. But many times, the ads click Data and the target page PV difference is very large. For example: In the hot zone diagram, the ad was clicked 1000 times, but when we looked at the PV of the target page, that number might be 600, or 1300. The PV of the target page may be higher than the number of clicks on the ad, or less than the ad click. Even if you have a separate tag for the URL of the ad, such as Http://www.bluewhale.cc/ref=homepage, there is still a difference, either high or low, between the data. What is this for?

　　

Before answering this question, first still must look at the indicator definition, the advertisement click and the page PV two indicators meaning is what, how to calculate? and whether we should use these two indicators to compare.

AD click: Ads click refers to the number of visitors click Ads, in Google Analytics is not click this indicator, so actually we see the click Number is actually advertising behind the target page PV value.

Target page PV: The PV of the target page represents the number of times the page is browsed, and every time the page is loaded, the PV data is added once.

The above two indicators are simple cumulative index, so it is not so accurate. Although logically, click and PV are recorded in the same page PV, then the two values should be the same. But for some special reasons, the actual situation is not the case. Here are some of the reasons I think there may be data differences.

A Ads Click and the target page PV is recorded PV data, but in fact, the two data are still different. The click Data of the advertisement only records the PV value of the target page accessed from the advertisement entrance, and the PV record of the target page is the number of times the page has been loaded. For example, there are 10 visitors to the target page a, but only 2 visitors open page A by clicking on the previous page ad. At this point, the ad click is 2, and target page A's PV is 10. As a result, two data differences were created.

B The second reason is similar to the previous one, when visitors visit the target page by clicking on the ad, but for some reason the page is reloaded (for example, refreshing). At this point, although all the traffic is from the advertising portal, but the number of ad clicks and the target page PV is still inconsistent. The ad click is less than the target page PV.

C The third reason is also more common, you may already have a separate tag for the target URL of the ad, and the visitor still accesses the target page by clicking on the ad, but the visitor may also share the URL with other friends via IM or other means, and when another visitor accesses the page with a tagged URL, Although they do not click on the previous ads, they are still recorded as the amount of traffic from the ad, because the URL comes with a tag from the first page. The reality is that none of the visitors have clicked on the ad. At this point the ad click and the target page PV data can not be consistent.

D is there any other reason? Welcome to the supplementary.

　　2, how to subdivide the site's only independent visitors?

How do I subdivide the unique visitor data in a Web site? For example: I want to know how many independent visitors Baidu has brought, how many independent visitors banner ads bring, and how many independent visitors the direct traffic includes. What is the number of independent visitors to a channel? This problem I have always thought was impossible, because in Google Analytics, the only independent visitor indicators can not be broken down by the high-level group. We must use the profile filter to split the individual visitor metrics. And Google supports up to 50 profiles for each account. Therefore, we can not arbitrarily subdivide the independent visitor indicators. But just a few days ago, I found out I was wrong.

　　

The way to subdivide individual visitor metrics is to customize the reports, and by selecting the dimensions in the custom report, we can easily subdivide individual visitors. The independent visitor metrics are subdivided in terms of traffic dimension, Time dimension and content dimension.

　3, how do you summarize the traffic from the EDM in the report?

In general, I recommend using the Google Analytics tool URL Builder to mark all external traffic, for example, for EDM traffic, we can get all the traffic from the EDM by marking the source as an EDM. However, if you have not marked the flow of the EDM before, or for some reason you cannot add the UTM parameter to the EDM's links, then all traffic from the EDM will be scattered in the referral source report.

　　

How do you count the flow of the EDM at this point? There are two ways to aggregate all the traffic from the EDM.

A, use a filter to summarize EDM traffic.

　　

The first approach is to use search and replace filters to summarize traffic from the EDM, match mail in all referral fields with regular expressions, and replace these traffic sources uniformly with edm_traffic traffic sources. By filtering, the traffic from all the EDM is uniformly aggregated in the edm_traffic source. This method in the report looks more clear, but there are two problems, 1, after the summary can only see the overall EDM traffic performance, can not classify the traffic to the mailbox category. For example: What is the conversion rate for Gmail? What about the jump rate in Hotmail? 2, you cannot use a filter to summarize previous data. If the EDM occurs before the filter takes effect, we cannot summarize the data.

B, use the Advanced group to summarize EDM traffic.

　　

Comparing the first method of using a filter, it is a better way to summarize EDM traffic with advanced groups. We create a edm_traffic high-level group and then use regular expressions to match all the sources that contain mail. When applying this high-level group, profile will produce edm_traffic traffic reports that can be applied to data of any time period. At the same time, in the referral flow report, we can still see the different mail service providers bring the effect of traffic.

　　4, how does the filter fail the goal page?

The background to this problem is this. There are multiple goal pages in the Web site, the URL structure of the page is exactly the same. And often there will be old page invalidation, new page update. However, due to the relationship between SEO, the failure of the old page will not be directly offline, but has been retained to attract traffic. But because these pages have been invalidated, even visitors to these pages cannot be counted as a transformation. Therefore, in order to ensure the accuracy of the site conversion rate, we need to filter out these invalid pages.

In order to maintain the real traffic of the site, we have to create a profile, and in which the failure to filter the page to ensure that the site accurate conversion rate data. But how to filter this part of the invalid page has become a problem. Because the expiration page is the same as the Goal page URL, it cannot be filtered directly from the URL. And because of the SEO relationship, we can not change the page URL or add parameters, or even modify the title. Therefore, in the filter, most of the filter mode we can not use.

At this point, without any adjustments to the page, to ensure that the SEO does not affect the case, filtering the failure of the page method is in the GATC to use the _trackpageview method to rename these pages. Because the search engine spiders will not read the content in JS. Therefore, this method search engine is almost imperceptible. The specific approach is to rename the failed pages uniformly.

_gaq.push ([' _trackpageview ', '/virtual/invalid page ']);

These pages are then filtered using a filter in the new profile.

　　

　5, how to analyze the jump rate of landingpage?

The last question, how to analyze the landingpage jump rate. Here we have to say is not how to optimize the landingpage design to reduce the bounce rate, but to analyze the reasons for the high landingpage jump rate. What do we need to do when we find a page in the site that has a high bounce rate? Do you want to modify the page? No, we need to first analyze the reason for the jump out rate. And that may be due to a number of factors. For example, advertising content, traffic quality, landingpage content, and advertising and content matching degree and so on. May be the reason for the flow itself, or the content of the page.

　　

Typically, a page is used as a landingpage for multiple traffic sources. When the bounce rate on this page gets worse, we need to subdivide him first. If all the traffic hits the page is very poor, then we need to optimize the page. And if only a certain channel of traffic has a poor bounce rate, we should be the flow of this channel to check the quality.

Author: Wang Yanping

Article Source: Blue whale's website Analysis notes reproduced please indicate the source link.

Http://bluewhale.cc/2011-05-14/google-analytics-question-and-answer-2.html