Statistical data fights: browser wars are on the rise

Browsers cannot be opened, but the real market share has become a mess

Lead:Foreign media wrote today that with the rapid development of the Internet market, internet devices and Internet protocols have become increasingly diversified, making it increasingly difficult to count the market share of browsers. Data from different organizations is even different. However, this also reflects from the other side that the browser war is growing.

The full text is as follows:

Data disputes

The new round of browser competition once again, this time, statistics have become the focus: Microsoft and Google, who are the most popular browsers in the world?

StatCounter, an Irish browser market research company, said last month that Google Chrome has surpassed Microsoft IE to become the world's largest browser with a share of 33% and 32% respectively. The company even said this advantage has been around for a while. However, Microsoft executives refuted StatCounter's data and quoted data from their competitors as saying that IE is still the choice of more than half of the Internet users around the world, with a share of 54%, far higher than Chrome's 20%. Last week, StatCounter again issued a message to fight back.

However, despite the dispute, the two sides have never clarified a key question: What are the users of various browsers?

The answer is of great significance to Microsoft and Google. They all regard browsers as the core elements that affect people's Internet access and search methods. It is equally important for website developers because they need to decide which browser to optimize the website accordingly. "Whoever owns a browser has the opportunity to attract more users for their products and services," said Al Hilwa, a software analyst at IDC at a market research firm ."

Inherent difficulties

This confusing status quo highlights the inherent difficulty of the network traffic monitoring industry: online behavior will automatically generate a large amount of data that can be tracked, but it is very difficult to interpret the data, it will become more and more difficult in the future. For example, both IE and Safari are bound to Microsoft and Apple operating systems and are usually sold together with new computers. Chrome and Firefox must be downloaded separately.

Before the first round of browser wars in the 1990s s, it was much easier to monitor Internet traffic. Nathan Newman recalled that at that time, IE began to challenge the market leader Netscape and eventually succeeded. Newman has a doctorate in sociology and is currently working in the Internet consumer group NetAction to study the browser market share. "Before the 1990s S, the network was much simpler than it was now. At that time, a computer was installed with a browser ." "So it's easy to track what browsers people use," he said ."

However, with the rise of smartphones and tablets, Internet access devices are no longer limited to PCs. In addition to browsers, other protocols and RSS readers can also crawl network content. Even the subject of Internet access is no longer just human: various automated programs, including search crawlers, are also growing at a high speed. The browser itself has also changed. For example, Chrome sometimes guesses the user's intention based on past data to preload the webpage. Internet traffic has also increased significantly in the world, and China has become the country with the largest number of Internet users in the world.

These problems make the share of Web browsers more difficult to calculate. These enterprises install tracing tools on their websites to analyze traffic and determine the browser users are using. However, websites installed with these tools cannot represent the entire internet. For example, some websites are very popular in the United States, but not in China. In addition, if the website is accessed by a search engine crawler or browser preload system, the statistics process is more complex.

Pre-Load

These causes make it difficult to determine the browser share. Roger Capriotti, product marketing director of Microsoft IE, published a blog article in March this year, criticizing StatCounter for taking Google browser pre-loaded webpage computing in the background. StatCounter subsequently corrected this issue, but found that it only compresses Chrome's market share by 0.3 age points.

Removing pre-loaded data from the browser makes sense because the website records these activities as access data, but this is not due to the user's choice, which leads to a high browser share.

As a competitor of StatCounter, the data of Net Applications is more trustworthy by Microsoft. According to the company, their approach has better removed the impact of crawler programs on market share. According to statistics, Internet Explorer's global market share is still twice that of Chrome. Vincent Vizzaccaro, executive vice president of marketing and strategic relations for Net Applications, said StatCounter's approach was "vulnerable to Crawler attacks ". Aodhan Cullen, founder and CEO of StatCounter, said the company had already screened out known crawlers and found other crawlers by surging traffic.

Different Methods

The data of these two companies is quite different, mainly because of the different statistical methods. They all provide traffic monitoring tools for websites, and then extract browser data with the collected information, but different methods are used in the analysis process.

Net Applications acknowledges that visitors to its customer websites may not be able to represent all netizens, but the company has referred to the US Central Intelligence Agency (CIA) Factbook ), therefore, global Internet users are weighted. For example, if the proportion of Chinese Internet users in the world is only half of the CIA data based on the company's data, when calculating browser data, net Applications multiply the browsing data in China by 2.

StatCount uses its own data, and even if other organizations list China as the world's largest Internet user, the company's data shows that China's Internet usage ranks only 22nd in the world. The company pointed out that Cisco data shows that the average traffic produced by Chinese netizens is much lower than that of American netizens.

Kulun explained the StatCounter method in two aspects. First, he believes that the key measure of browser usage is actual browsing activities, or page views. Net Applications statistics are independent user visits, which are usually the data that advertisers value, because they do not want to pay constantly because the same person repeatedly watches the same advertisement. However, when measuring the market share of the browser, the actual usage is more important. A Microsoft spokesman declined to comment. The company only reiterated last week that it is more inclined to use Net Applications data due to weighted factors. Google declined to comment.

In addition, Kulun also said there is no reliable statistics on Internet users. Different organizations make different estimates, and most of them do not distinguish mobile Internet users from desktop Internet users. However, these two groups face completely different browser choices. "There is not a weighted system that is accurate enough to choose from ." "That's why we didn't consider any weight," he said ."

Huge impact

Wei zaluo said that the statistical method of Net Applications can "provide more accurate market share, so as to show the actual number of users, rather than clicks ."

StatCounter's processing of Chinese data has a great impact on the statistical results. The company said in a blog article that, if weighted like other companies, IE's market share in May this year will increase from 35% to 43%, leading competitors significantly. However, even so, StatCounter data shows that IE usage has declined over the past year, contrary to the conclusion of Net Applications.

Asa Dotzler, Mozilla Firefox product manager, cannot confirm any piece of data, but he believes that competition in the browser market shows that "we are working towards the goal ."

Kulun said StatCounter is not biased towards any browser. But he also agreed with dozale. "It is a good thing to have three or four important browsers at the same time," he said. "The online world is no longer the only one ."