10 principles to help you better use web analytics tools (top)

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

Objective】

The article came from a speech at the CWA and WAW conference in Guangzhou on May 28, when the topic was "rational use of web analytics tools." Speech 30 minutes, but covered a lot of aspects, so a friend secretly told me did not understand. The topic may take 2-3 to four hours to make it easy to make clear, so I think it's necessary to open a new article.

Body】

Website Analysis tool is the necessary weapon to do website analysis. From the day the history of web analytics was uncovered, web analytics could not be separated from the tools. The level of tool utilization, to a large extent, determines the level of Web site analysis.

But there is a big misunderstanding, that is that the better tools to help make better web analytics. I have used perhaps the most advanced tools in the world, but I know that the advanced tools and the advanced Web analytics are completely different things. I know that if you don't use your tools with the following 10 principles, any tool can make your analysis a hassle. This is like AK47 is a powerful weapon, but the muzzle is entirely possible against yourself.

Principle one: Do not use tools that are beyond your means

As far as I can see, the strength of the web analytics tool is proportional to its complexity. Behind the powerful, it means more flexible customization capabilities, and customization can only be done by people. Good customization depends on three points:

You have a clear grasp of your business;

You have a clear grasp of the tool;

You have a clear understanding of the relationship between your business and your site's performance.

　　

These three points, cough ..., for most of our friends are very high requirements. For me too, I dare not say I have a clear grasp of the business, the industry and the company's internal situation requires a lot of time to understand and experience. This requires perception. On the other hand, people who have a clear understanding of the business may not be able to get a thorough understanding of tools. Business and tech are like two different languages, using the left and right brain to think separately. My friend Lizzy Fan's boss Yi Shi (CEO of Avazu) is such a person, so he is extremely successful, but such a person is too few, in short I am difficult to meet. After that you have a deep understanding of the relationship between your business and the performance of your site, which is equally difficult.

I've seen a lot of companies in China using Omniture's flagship analytics tool Sitecatalyst, but it's not as much about Google Analytics as it is for GA. If you're just using simple features, then it's probably better for GA in the user experience, even CNZZ or Baidu stats are enough. If you do a custom function, you must understand the logic between the Sitecatalyst flow and the business flow, and the traffic, event, prop, and evar that are triggered by both streams. These, really not so easy.

My advice is to make sure you are more powerful than this tool if you want to use more powerful tools. If you think you're not enough, make sure you have a strong implementation team and a team of consultants to ensure that you can tailor your requirements to the flexibility of your needs, or deal with the tools you know. Of course, you can ask Michael Li and me for help, but we can't keep bundling in your business, so a more powerful tool means you should be more cautious.

Remember, you use tools, not tools.

Principle two: Free tools no good, pay tools will be more difficult

The commercial paid WA tool always gives us a throbbing heart, these "unattainable" tools, how powerful ah! So, we may think that the free tool "bad" is the function of the tool is limited, we use the "pay tool" will certainly be completely different!

I basically disagree with this idea. Just like the following figure-"I can't have children, it's your fault!", but it's a joke, it's not a tool, my mother hates me for the wrong "find objective reason" excuse me. :D

　　

Free tools are not good, and pay tools will not only help you, but you will worry more.

A free tool like Google Analytics or Baidu Statistics brings together some of the most core features of web analytics, and embodies many of the core ideas of web analytics. These tools are great.

You think the tools are "bad" and may really just be because you don't have the potential to maximize the mining site analytics.

In this regard, my colleague, Travy, is an expert who understands the significance of maximizing free tools such as "Squeeze" Ga.

"Because at that time our company was not willing to spend more money, so we tried every means to use GA," Travy said.

The use of this made me an eye-opener. Frankly, before that, I didn't know that GA could have such a flexible way of using it. All of these methods are just a method based on GA, the _trackpageview () method, our default use is not parameter, but if you add the various parameters-including the page title, including the crumbs, including the URL of the page after the interception of the translation of information, GA adds a lot of functionality that we didn't think it could implement at all. Using crumbs, GA can even have a little bit of a path function!

Similarly, for _gaq.push ([' _tracktrans ']) This command can also be dug very deep, you can put more product classification information, prices and according to your request summary information through dynamic variables to pass to the GA. GA can also become more powerful in monitoring commodities.

　　

I started to admit that I don't know enough about the ability to do free tools.

Of course, the ability of a free tool is limited, even if you squeeze it, it always touches the ceiling. But my point is clear, the process of extracting this tool is the process of in-depth understanding and practice of site monitoring. If you're willing to take the time to squeeze on free tools, then when you start using a Pay tool, you'll find it's not that difficult. But if you keep complaining when you're using a free tool, I can imagine that you'll have more complaints when you use the Paid tool.

Because we have said in the first principle that you use tools that are beyond your means.

Start with a free tool, squeeze it, and make it. :)

Principle three: Do not attempt to accurately monitor transactions with WA tools

Another myth is that people always believe that paid tools are more accurate than free tools. Well ... I don't think so. Because of the principle of the monitoring mechanism itself, the accuracy of the tools is different from what you understand.

It is noteworthy that the common pay tools and free tools in the flow monitoring of the implementation of almost the same, so there is no more accurate problem between the two. In the monitoring of transaction data, after all, unlike the E-commerce site internal ERP system, web Analytics tool is to capture the transaction information on the page to achieve transaction data acquisition, so often just order confirmation, or pay the confirmed data, but not the final deal of data (COD transaction success or failure, And the user returns the withdrawal list and so on the situation is very difficult to be monitored by the website analysis tool.

The tools to accurately monitor transactions are your company's ERP system (or invoicing system), but WA tools are difficult.

About the accuracy of WA tools, there is actually more content, read this article-Web analytics-Our data accurate? You can see why we don't need to be accurate. :)

the following references:


extended reading: How to "make" WA tools can also accurately monitor traffic channel transaction data

WA tools to monitor the way to accurately monitor the transaction data is very difficult, although Google Analytics and Omniture can be set to achieve withdrawal and reverse logistics, such as the deduction of income, but too cumbersome, operational feasibility is not high.

However, the WA tool can "monitor" the transaction data of traffic channel almost accurately.

Notice that I've made several quotes because it's not directly implemented through the WA tool. We need to adopt a flexible approach to allow the GA recorded traffic channels and these channels corresponding to the transaction data accurately.

There are two methods, both of which have a premise that you have a media sales management system (or a similar system) that you can use to tag traffic to capture traffic-generated sales. For example, if the tag format of this system is from=, then the URL of the ad that I put on my blog for the catwalk should be Http://www.xiu.com/?from=CWA, and the system will be able to identify the sales of CWA traffic. Such systems abound, and are ubiquitous in electric-business websites.

If you have this system, then congratulations, you will be able to achieve the flow channel and the corresponding sales of accurate matching purposes.

One way is to use GA: let GA's UTM link tag recognize the tag format of your system settings. If you have seen the secret of traffic, then you know, utm_campaign, utm_medium, etc. 5 tag media properties can be customized by name. You can use the from to represent Utm_campaign, so in the above example, FROM=CWA can be identified by the GA as the traffic of a campaign called CWA, and it can be identified by your media sales management system as a traffic from the CWA, And record the sales that these flows can bring.

The practice is very simple, only need to implement GA code when the use of: ....

Another relatively stupid point is to allow your own system to identify the source of the flow of the WA tools, such as your system is identified? from=, now you let your technology move, change to identify? utm_content=, also completely feasible, the modification is very simple.

After doing so, you will find that you can get a very accurate statement as follows:

　　

beautiful! Now you can know exactly how much the conversion rate of each flow source can be! If you have access to the cost data, then accurate ROI is achieved!

Principle four: Again intelligent, also should pay attention to hand

The trend of web analytics tools is more intelligent, such as the introduction of their own very advanced thermal mapping features. Or give some suggested data that has been processed two times.

I have a love-hate relationship with intelligence.

There is no intelligence, a lot of work is unimaginable, but intelligence sometimes hides the truth and makes people lazy to think.

For example, the hot map, the Getoux is particularly unable to intelligently because the page has many of the same URL of the chain access, or page links are often updated (such as our E-commerce site). The way I handle this is to link all the URLs on the page to the URL and the dynamic changes, plus identifying parameters. Then manual statistics. Need strong executive power, not simple, but very accurate.

You may say this will affect SEO, but has been done by my colleague Jay, not a problem.

but this method is not very suitable for GA because GA has no next page report because GA has no path function. This is so painful. V5 version of GA also does not have this function, extremely disappointed. GA without path function, too incomplete. But with Omniture Sitecatalyst or Yahoo!, Web Analytics, Getoux can be made very accurately by the manual method I mentioned above.

Another reason to attach importance to craftsmanship is that no matter how subtle the monitoring is implemented, it is impossible to be perfect. For example, when I was doing this sharing in Guangzhou, I had a friend ask--how can I evaluate the long tail effect of SEO well? This job is hard to read directly from Google Analytics, you can only find a way to export all the long tail organic keyword. Then use their own experience and wisdom to analyze their laws with the data.

If you ask me, my website analysis uses the most tool is what, my answer is not Google Analytics, also not omniture of a tool, but Excel. Many times, I am not particularly concerned about the similarities and differences between Google Analytics and pay tools, many times I am in Excel to achieve the analysis, WA tools is just my analysis of the raw materials needed. This is also why the commercial web Analytics tool must retain and provide raw data (e.g. Omniture Sitecatalyst provides Data warehouse functionality).

the following references:

Extended reading: How to export 20,000 Google Analytics data at once?

With Google Analytics Ordinary data export, you can only export up to 500 data.

　　

But for older versions of Google Analytics, you can export any number of data in a simple way.

1. Click on the export label above.

2. Right-click CSV and copy its link address.

3. Open a new browser window, put the copy of the link address into the Address bar, but do not rush to activate this URL.

4. Add &limit=15000 after the URL in the address bar. You can export 15,000 of data. You can change any number after the equal sign.

This method is not valid for the new version of Google Analytics, and the new method I haven't found yet.

Principle five: WA tools can not satisfy all your inquisitive

In general, the WA tool refers to Web site user click Stream Tools (Clickstream tool) such as Google Analytics, Omniture Sitecatalyst, which can tell you what (phenomenon), But it is generally difficult to tell you the what behind the Why (reason).

For example, when you see the top content report of Google Analytics that the bounce of a certain page is rate high, you know a very important phenomenon--the first impression that a visitor enters your site through this page, Leave your site without entering another page.

But what causes this phenomenon?

If it was me, I would do the following things:

1. Check to see if the page is no link pages, or a less-linked end page.

2. If not, see the source of the flow of this page is where, if the main search engine, then congratulate you, your problem is easier to solve; if it is not a search engine, but the direct flow of the main, then it is very regrettable that your problem through the WA tool to find out.

　　

Why?

The reason is that direct flow is a kind of people like to add but also "extremely egg pain" flow. Direct Flow only tells you a message Google Analytics don't know where the traffic is coming from! Since I don't know where they came from, it's hard for me to know why they didn't make the first impression on my page.

For WA tools, we must treat them correctly. The great thing about it is that it helps us to find out what we're not aware of through the data, and can help us demonstrate whether our improvements will lead to better data performance. But the reasons behind the phenomenon, many times through the WA tools can not be done, we have to introduce other analytical tools and methods.

These analytical tools include AB test tools, research tools, pannel, eye-moving devices, and so on-research, usability testing, Delphi, user tracking, and more.

Speaking here, I want to do an extension of the scope of the WA tool, we commonly used GA, sitecatalyst, CNZZ or Baidu statistics, but is the site analysis of the Click Flow Analysis Tools, web analytics tools should have a greater scope of containment.

All right, here's the story. I haven't opened a new article for a long time, so I'm looking forward to my friends ' messages. Any message is welcome, the question is also welcome!

Thank you for your patience!