Web analytics: What kind of tools do I need? (3)

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

Objective】

Tools are essential elements of web analytics. But how to choose a desirable tool is not so easy. How to choose the right tools according to the needs of different websites? This is a problem that many friends are concerned about. In the first installment of this topic, we discussed the choice of small ecommerce sites for tools, and in the second, we analyzed how midsize and large E-commerce sites choose web Analytics tools. In the third chapter, we continue to analyze the requirements of other tools for large and medium-sized e-commerce sites, as well as the choice of web analytics tools for brand advertisers.

Body】

Requirements and Tools (cont. 2)

Type two: Large and medium-sized E-commerce Web site (continued)

In addition to the requirements mentioned in the second article, large and medium-sized E-commerce sites also have a very important demand-testing. When it comes to testing, you can't just talk about a misunderstanding that people usually have about web analytics.

Many friends or companies with me to consult their website, let me help them analyze the status of the site and marketing results. When I point out the various situations that exist on their site, they always ask me-"Why is this happening?" I often have 55 answers to this question. The 50% possibility is that, based on experience, I can explain the reasons behind the site's problems, but the other 50% is that just using Google Analytics, sitecatalyst or webtrends can't answer for you.

This is what Avinash said, the narrow web analytics tool is to help find what tools, that is, to help find web site problems, but it is difficult to help us pinpoint the reasons behind the problem. The object of the website analysis tool is human behavior, after seeing the behavior of the person, can discover the difference between the visitor behavior and the website design expectation. And it is these differences that we need to solve or need to explain. The discovery of the problem (ie what), the so-called Web Analytics 1.0 phase, explains why (i.e. why) behind the problem is the Web Analytics 2.0 phase. Human nature, not willing to just see the surface of things, but also want to know the things behind things.

So in order to solve my 55 after 50% of the trouble, we need to introduce some new tools. These tools are valuable not only to large and medium-sized e-commerce sites, but are actually valuable to all sites. But large and medium-sized e-commerce sites are worth using them, because the resulting effect directly brings money.

There are a number of tools to help us understand why, as I've mentioned in a number of previous articles, such as survey (research), such as site Pannel, such as viewing the visitor's buzz (message, etc.). But for web analytics, the most important and direct solution to the problem is testing (testing).

The significance of the test is found that the site analysis phase of the recommendations can really affect the performance of the site, although not like research, testing does not actually directly answer why, but directly solve the problem of how to optimize the site. After all, understanding why, after all, we want the ultimate goal is to optimize the site.

The most well-known test method is a/b test, the principle is super simple, I am going to suspend the Web analytics tool narrative in the next article, and go to talk about A/b test. This article first introduces the tools for A/b test.

　　

A/B testing tool, there are several ways. A commonly used method is that the Web site's art and front-end engineers to design a problem page of another version of the page, and then by the engineer technical settings, so that traffic automatically sorted to two pages. This is no use for extra tools, just dim yourself. This approach is not really a big problem, but it increases the workload of the front-end engineers and is not flexible enough. So use ready-made tools to become a popular choice.

A/B testing tool, if you use Google Analytics to do monitoring, then use Google Website Optimizer (GWO). The two are compatible. GWO is also a free tool, so of course you can't implement too many custom functions, such as a comparison of the effects of different versions of a page just to define a common transformation target page--but, in many cases, this conversion is sufficient, and you have GA in your hand to analyze different A/b test pages , so GWO can help you achieve automatic streaming flow (traffic automatically randomly into different versions of A/b test page) is sufficient. So, there is no special need to do a basic A/b test, with GWO is a good choice.

However, if large and medium-sized E-commerce sites require high levels of flexibility in testing, the use of paid testing tools to meet demand. Such requirements are often concentrated in the following points:

It is not only A/b test, but a multivariate test of page local module change;

is not only a single transformation effect, but a variety of transformation and behavior of the effect of comparison;

It's not just about testing a single version of Who's better for the overall performance, it's about directly positioning different versions of different individuals.

I'll give you a little explanation for the 3rd. Because people's preferences are different, GWO, such as the free tool, can only find a higher conversion rate for the overall page version, and a good tool, you can implement dynamic page, that is, display to different preferences of the people they are happy to see the page, in order to optimize the site from a more detailed individual perspective conversion rate.

Adobe's test&target is such a tool. Of course strong, and relatively not complex, have the opportunity to friends are worth trying. The actual effect of this tool, you can refer to Sony's U.S. website. The first time you visit the Sony website, the products displayed on the home page may not have anything to do with your interest, but if you visit several PSP or PS3 pages, and then come back the next day, you will see the site's home page has changed significantly. Yes, this change is only for you and visitors like you. Moreover, every other visitor with different interests has its own "exclusive" version of a sony.com home page.

　　

So, the Adobe Online Marketing Suite (formerly Omniture) not only provides sitecatalyst (corresponding to Google Analytics), but also provides a number of other tools (as shown in the above). These tools are largely designed to address why, as well as the practical solutions to help Finalize site optimization.

Type III: Brand advertising main website

　　

Brand Advertiser's website is a kind of special website. There is a slight overlap between these sites and e-commerce sites, where they may also sell some of their own branded products. But publicity, not merchandise sales, is the more important purpose of this kind of website.

Intel's website is a typical brand advertising site, as is Apple. Of course, Apple's web site is more akin to a branded advertiser site + E-commerce site, but in relative terms, e-commerce sites are much less complex to monitor and optimize than large e-commerce sites that contain many product categories. So, in this section, we're not going to discuss the e-commerce issues of the brand advertisers website (see the first and second articles in this series), but specifically about the role of the site as part of the promotional tool.

If the site is focused on letting customers know, understand, and even like the company's brand or commodity, then analysis needs will be very different from E-commerce. The analysis demand of e-commerce emphasizes the process, but the brand website emphasizes content consumption. E-commerce focus on transformation, brand site also focus on transformation, but the transformation of the two connotations of the two-e-commerce site to emphasize the results (transactions), and brand sites emphasize the beginning (interest in goods and purchase will).

The monitoring requirements for brand sites are usually as follows:

Visitors are who they are, from where, is not the target of the brand to hit the object;

Visitors to the site's first feeling-is to see the light die, or fall in love, or the impression is not deep;

Is the targeted content viewed by targeted visitors?

Which content visitors are more interested?

Are visitors willing to take part in events and events on the site?

Do new media forms (such as gorgeous flash animations) or presentation (such as scrolling graphics or video) have a positive impact on visitors?

is the navigation of the website easy to use? That is, can visitors easily find what they are looking for through the site's navigation system?

is the layout and content of the key pages reasonable?

is internal search reasonable and easy to use?

is the help system reasonable, easy to use and able to solve the visitor's problem?

How many visitors are likely to turn into customers?

And so on and so on and so on. First, the visitor is not a reliable crowd, the second, the content of the site is targeted and combat efforts; Third, whether the site to make visitors easy to browse, four, the site can be fooled visitors to do some interactive activities.

Therefore, frankly speaking, brand advertisers web site monitoring is less difficult than e-commerce sites, while the choice of monitoring tools, depends on the specific use of the site technology and its attached business.

Technical point of view

If the site is based on static pages, few Flash, JavaScript, dynamic page generation and video, use free tools enough. Google Analytics is a good choice. Other estimates Baidu statistics, CNZZ or 51LA, also won't make you feel uncomfortable.

If the site contains a small number of Flash,javascript, Google Analytics event tracking solution is no problem. Be careful when adding code, not a problem.

If the site has a large number of complex flash,javascript and dynamic page generation and call, in fact, Google Analytics also line, is a fee, the implementation of the time is more troublesome. If there is a cross-domain or cross-domain, and so on, the implementation of GA requires a higher, but there is no theoretical barriers to access to the monitoring data itself. Using Adobe Sitecatalyst (formerly Omniture Sitecatalyst) makes this process a little easier because of the guidance of a professional service person.

　　

If the site has more video, and the video playback data need to monitor, then need to customize the video player. Google Analytics and other free tools will be powerless (because of the lack of targeted solutions to make monitoring implementation difficult, although not impossible, but involves a lot of physical activity), and Adobe Sitecatalyst (formerly Omniture Sitecatalyst) would be easier, but it would require a fee.

Business View

If the site is mainly static display, what visitors need to do is to read, then, choose a free tool. It's not necessary to spend money.

If the site contains a small amount of user interaction, such as getting users to download brochures, coupons, and so on, use GA event tracking or similar tools.

If the site contains a small number of user interaction with the process, such as filling out a purchase intent list, then depending on your monitoring needs, you can use GA or sitecatalyst. If you only want to know how many people have submitted an intent list into the next step, GA can be completely satisfied, if you want to know the user fill out the intention to fill in the form and give up the situation, then you can only use Sitecatalyst, because it contains the form of monitoring the plug-in.

　　

If the site allows users to participate in more interaction, for example, the choice of monitoring tools depends on whether you are willing to monitor these activities from a business flow perspective and the need to monitor granularity (delicacy) of these activities, such as submitting intentions, participating in surveys, participating in the lottery, participating in comments, and participating in sharing. You can of course choose GA and other free tools, to a flow of traffic perspective of these activities, the price is from the overall point of view of a site to measure the overall site participation needs to do more manual calculation, relatively troublesome, and more detailed monitoring of activities can not be. If the choice of pay tools such as Omniture sitecatalyst, of course, can solve all the problems, but the cost of comprehensive monitoring is clearly increased the difficulty of analysis. However, using the Adobe Omniture data interface engine Genesis, these business-level data can be directly imported into the enterprise's BI system, which makes the data directly into information, the work is very difficult to reduce. Intel has been taking this approach as I left the OMG.

If the website not only lets the user participate, even provides the SNS function to the customer, then uses what tool to become complex. I'll put it in detail in the article on how to choose the SNS website monitoring tool.

All right, thank you for your patience. Read so much content. If you have any questions to ask, what do you think?

Web analytics What kind of tools do I need? (2)

Original: http://www.chinawebanalytics.cn/how-to-choose-web-analytics-tools-3/

Reprint please specify the Source: Website Analysis in China