Uncover "Ten Websites Analysis myth" truth

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

I've been working on Web marketing and web analytics for nearly a decade, and I've heard almost every web analytics myth that's been circulated in big companies and small companies alike. Here are 10 of my favorite websites to analyze myths and how to expose them.

1. Free Web analytics tools to match the corporate version

There are many reasons to prove that free web analytics tools are not the best solution. I often like to ask "Why do we use the corporate web analytics tool such as omniture or parsing instead of Google Analytics" to refute it, which would, of course, lead to some witty answers, such as "because I had to pay the bills" or "My boss made me do this". Obviously this is not the best answer (not even the answer), the root cause for choosing an enterprise Web Analytics tool is:

· High-level service agreement: What happens when a tool fails? If you pay for the tool, you have the absolute initiative, but if you use a free tool, you have to wait for it to return to normal, and pray that there is no data loss.

· Data ownership: Free does not mean that you have free influence at the same time, someone is paying the bill because the free tool usually provides you with a "0 cost" service. As long as you want, the enterprise version of the solution can guarantee you always have your own data.

· Privacy: An open agreement between the two parties in an enterprise version of the solution will better protect your security and privacy

· Custom: Even if you can make Google Analytics such a free tool to do its best, but also can reach a certain level. The enterprise version of the solution can be customized to your business goals.

2. Bounce rate (or "Some cow x measure") is the best measure

Avinash Kaushik (Translator Note: Web analytics guru, two authors of the classics of WA, do not translate) the "bounce rate" as the sexiest indicator, but it is not the best, because there is no "best indicators." I've learned that many companies employ specialized analyst teams whose only job is to monitor a certain "God-given" indicator, but rarely does it stand the test of time. It is recommended to select a few indicators that can actually drive profits, and then focus on them. (Translator Note: You can refer to Avinash Kaushik's "Web Analytics2.0" in the beginning of the fifth chapter on the importance of output story, very interesting)

3. As long as Avinash Kaushik, Jim Sterne, or Eric Peterson absolutely do it (translator: for these three Daniel interested in, can be circled on Google +)

Don't get me wrong, Avinash's talent is brilliant, but no analyst can provide a plug-and-play (fully-counterpart) metric before it's enough to understand your business background. At a high level, their best advice is important, but nothing is more important than studying your own data and making your own feasibility analysis report.

4. Website Data analysis dashboard or Web site data analysis report should contain 4 quadrants and have very little data

This is a lofty goal that you want to achieve when you create something, such as a profile, a menu, and so on. In order to cater to the needs of various people, it is very difficult to integrate various data, insights and visual effects into one page. A good suggestion is: make something bigger than you need to show your strength; in your organization, just get stakeholder attention, consult with different business teams, and then provide customized site data analysis reports for different audiences on their own needs.

5. Opinion is more important than data

When you make a decision, critical data can often provide you with executable requirements. Will your company provide IE6 official support for the next generation of product design? If there are only 2% visitors to your site through IE6 in the last 6 months, and the development and testing of integrating the application will cost US $ millions of, the answer to the question just now is simple.

6. The number of unique visitors is calculated by the actual number

The number of unique visitors may be the most abused indicator in the past. If you think about it, the metric that is used to measure the "number of visitors" is just the number of persistent cookies in the browser. The number of unique visitors is not the same as the number of browsers, the number of independent real people or computers. (Translator Note: The impact of cookies on site analysis can be referred to flops "Web Analytics Basic Course" Chapter III "Web Analytics Technology Decryption" section I)

7. Analyzing scripting code can degrade Web site performance

All code can degrade Web site performance. Adding any code to a fully-empty Web page increases its load and execution time. In other words, to support Web analytics data collection, some custom code will make your JavaScript files more expansive. When you add any code to your Web page, you can measure the benefits it can bring and the price you will pay if you don't have it. (Translator note: Use Fiddler to test Google Analytics and Baidu Statistics, both for monitoring code request response time within one second)

8. Website analysis is marketing/research/communications/operations/IT or other department responsibilities

Web Analytics is the responsibility of a data-driven organization. If your site affects your business in any way, everyone in your organization has a responsibility to share this responsibility to present actionable business insights that increase revenue, reduce costs, seize opportunities, or avoid risk.

9. Matching metrics, logs, and databases from the analysis providers of different Web sites

Web analytics are fundamentally inaccurate, and few practitioners are proficient in statistical theory, so it is pointless to argue whether data collected from different sources can be right. Here are a few factors that will cause the Web site to analyze data incorrectly:

· Vendor-supplied JavaScript scripting code and browser compatibility

· Audience acceptance of cookies

· Data corruption: During acceptance, execution, and transmission

· Server-side caching, scripting, or configuration issues

· filtering or Processing rules: Reverse DNS resolution inaccurate, data sampling, Data coding problems

Review historical data and analyze trends to make sure your findings are statistically significant and ignore the limitations of web analytics before making a conclusion.

10. The opinion of the website analysis is free

There is no free lunch. Adding monitoring to a Web site using JavaScript scripting code also takes time and effort, and it takes a lot of cost to analyze reports and go deep into execution, and it takes some extra tools to consider, as well as to consume some of the organization's opportunity costs in the practice of web analytics.

(Copyright GUI Lin website Analysis blog All, welcome to reprint, but reprint please specify the source.) )

I am here just to initiate, do you have what I miss the website "myth"?

English Original address: http://searchenginewatch.com/article/2108307/Top-10-Web-Analytics-Myths-Dispelled