Shulou(Shulou.com)11/24 Report--

After nearly two years of waiting, the official version of Wechat keyboard 1.0.0 has finally been launched.

This is what Zhang Xiaolong calls a product developed to protect user privacy. It does not collect data and is not used for profit.

Wechat keyboard interface why does Wechat do input method? If you understand the past life and this life of the input method, you may be able to understand what the giant is thinking.

Although the functions of different input methods of the Chinese character business on the keyboard vary greatly, their principles are exactly the same.

The electronic world is a binary system with only 1 and 0, no matter what Chinese characters are input or output, they all go through binary code conversion.

The GB2312-80 "Chinese character coding character set for Information Exchange" issued in 1980 is equivalent to issuing a unique "ID card" to each Chinese character. The job of the input method is to ensure that the user remembers the ID of each Chinese character in the fastest way.

From then on, the third-party input methods showed their magical powers, and there were all kinds of input methods on the market, such as five strokes, full spelling, double spelling, and so on. The traditional stand-alone input method appeared.

Traditional input method from the input mode, we can divide the input method into two categories: code input method and phonetic code input method.

The most famous one in the code input method is the five-pen font. In 1983, Wang Yongmin, a professor of the University of Science and Technology of China, invented the five-stroke font according to the writing characteristics of Chinese characters, and established Beijing Wangma computer Co., Ltd. in 1989 to sell the "Wangma five-stroke input method".

The five-stroke font solves the problem of Chinese character input on the PC. Xinhua News Agency said that "its significance is no less than movable type printing," and Wang Yongmin has become a "contemporary Bi Li" and a national model worker.

But in fact, as early as 1964, when Wang Yongmin was an undergraduate at the University of Science and Technology of China, Zheng Yili, an compiler, was designing another code input method for the Academy of posts and Telecommunications-- "four-code Chinese characters." unfortunately, it was later interrupted by force majeure.

In 1979, Henan Science and Technology Commission invited Zheng Yili to make a special report on the coding of Chinese characters. Wang Yongmin also rushed to Zhengzhou overnight to ask Zheng Yili about the coding of Chinese characters.

It was not until 1989 that Zheng Yili and his daughter Zheng long invented the "character universal code" Chinese character input method, which is later called Zheng code [2].

Those who are impressed by the five strokes may remember that although the typing efficiency of the code input method is high, it takes a lot of time to memorize the roots, and the cost of learning is very high. How to create an easy to popularize input method became a problem faced by developers at that time, and the answer lies in the phonetic code input method.

Image source: Wangma official website [3] phonetic code is a kind of code that converts Chinese characters into pinyin letters. When users use the phonetic code input method, as long as they input Hanyu Pinyin, they can find the desired Chinese characters, which greatly reduces the threshold of Chinese character input.

In the late 1980s, Dr. Wang Xiaolong of Harbin University of Technology developed a pinyin input method called InSun, which not only allows PC users to get rid of complex roots, but also supports whole sentence input, making a compromise between input efficiency and learning threshold.

However, this input method did not receive widespread attention at that time. In the mid-1990s, InSun was sold to Microsoft by Wang Xiaolong for $100000, which became the built-in Microsoft Pinyin of Windows95 [4].

Also acquired by Microsoft is Smart ABC. The input method, developed by Zhu Shoutao, a professor at Peking University in 1990, combines the functions of phonetic codes and shape codes: users can quickly read and locate the character through the code of the first stroke on the basis of pinyin input.

Although more people just use it as an ordinary pinyin input method, it has not affected the success of intelligent ABC at all. After being acquired, Microsoft built it into Windows2000, making it the most widely used input method in mainland China at that time.

The success of intelligent ABC has led to the participation of more developers, and the competition in the domestic input method industry is becoming more and more fierce. Excellent pinyin inputs such as Ziguang Huayu, Intelligent Crazy Pelling, Dark Horse Shenpin, Pinyin plus and other excellent pinyin inputs have been introduced one after another, and they have also captured a number of followers.

The "melee" of the input method lasted until 2005, until the only four left-Intelligent ABC, Ziguang Pinyin, Pinyin plus and Microsoft Pinyin, accounted for 90% of the market. Later, the intelligent input method represented by Sogou Input appeared, and a new round of changes have taken place in the domestic input method market.

Intelligent input method in 2005, Ma Zhankai, a new graduate from a state-owned enterprise in Taiyuan, found that when the pinyin of "zhoujielun" was typed in the search box, the search engine would prompt "is it Jay Chou you are looking for?" This inspired him-can we connect the input method to the Internet and develop an input method that can monitor current hot words with the help of the function of the search engine?

So Ma Zhankai began to send e-mails and suggestions to Baidu. After several emails sank into the sea, Ma Zhankai turned to Sohu, joined Lightspeed to become a product manager, and joined Wang Xiaochuan's team to develop Sogou Input. Until June 2006, Sogou Input officially launched, and the Chinese input method was declared to have entered the "network era".

Compared with the traditional input method, the intelligent input method with its own search engine not only "guesses words" more accurately, but also creates the account login function of the input method, uploads users' input habits to the cloud, and makes it convenient for users to use the input method and back up their personal thesaurus in different places.

It can be said that the intelligent input method is a user-specific artificial intelligence. Every time users type, they are training AI. As long as users log in to their accounts on the Internet, they can switch to the input method that "knows themselves best."

The popularity of the Internet confirms Ma Zhankai's thinking. Microsoft's native input method and other traditional input methods are becoming more and more popular, and other big companies have also launched similar products. In 2007 and 2010, Google, Tencent, Ali and Baidu successively released similar input methods. However, Sogou, who was the first to enter the game, still has an irreplaceable advantage. By June 2009, Sogou Input's installed capacity had exceeded 80 million, accounting for 70 per cent of the input method market.

More and more companies have launched intelligent input methods, and "coquettish operations" between their companies are also common.

Sogou, as a pioneer, took the lead in pointing the finger at Google. In 2007, Sogou issued a statement saying that the Google input method embezzled Sogou thesaurus, which ended with Google publicly apologizing and deleting the controversial thesaurus. Subsequently, domestic Internet companies began to wrangle with each other.

On June 23, 2009, Sogou sued Tencent for 20 million yuan for unfair competition and intellectual property infringement. However, Tencent did not hesitate, but sued Sogou for false propaganda to the court on November 4 of the same year, and the claim amount was also 20 million yuan.

Tencent even opened a special page on "unfair Competition of Sogou Pinyin input method" in QQ.com for this purpose. The page not only collates the latest reports, expert opinions, and the comparison of the functions of the two products, but also introduces a lot of news about Sohu's lawsuit and 60 patents of QQ Pinyin [6]. This "commemorative web page" has not been taken offline until today, nearly a year after Tencent and Sogou completed the merger.

The Beijing Youth Daily once said that the struggle between Sogou Input and QQ Pinyin shows that the market competition of Pinyin input method has begun to enter a white-hot stage. Since then, there have been several disputes in the input law circle, such as Sogou and Baidu suing each other for input method patent infringement in 2015, Baidu and Baidu jointly suing Sogou Input for traffic hijacking in 2019 and so on.

On the one hand, Internet companies are quarrelling over the patents of intelligent input methods; on the other hand, the advertising problems of third-party input methods disturb users.

The declining Microsoft input method suddenly seized the opportunity to announce in 2010 that it would step up its update efforts and improve the experience of using Chinese input method on Windows10 [11]. On the basis of continuing the pure and advertising-free characteristics of the original input method, the new Microsoft input method has optimized and upgraded a large number of functions--

Microsoft's input method experience is getting better and better, but it is still too late to strive for strength. At this time, the Internet companies have already shifted their business focus to the mobile end.

From PC to mobile phone, input method has become a "necessity" as early as 2008, when universal charging can be seen everywhere, Sogou has already entered the mobile business, targeting users' smartphones. However, at that time, Google had just launched its first Android phone, so Sogou chose Symbian, which was more mature at that time, in the mobile test.

The input method front from the PC end to the mobile phone end, small companies are also emerging to compete with big manufacturers for the market, the emergence of Toubao, A4, Dianxun plum blossom and other products focused on mobile phone input method.

To 2009, LG, Sony Ericsson and other manufacturers quit Symbian Ecology. In the face of the wind, the input method manufacturers began to launch the Android version of App. In 2011, Android's market share in smartphones jumped to 48%, crowding out Nokia to become the world's largest smart system [12].

Symbian is over, and mobile input methods have switched to Android. In 2014, Apple, which only supports native input methods, opened third-party input methods for the first time in iOS 8, and companies announced that they would log on to App Store for two-line development.

The mainstream brand mobile phone input method has been online for a long time, but it has been proved later that even on a new platform such as the mobile terminal, the input method has to "hold the thigh" if it wants to increase the volume. Among the above three companies, A4 and Dianxun Plum Blossom were acquired by Tencent and Baidu in 2009, respectively, and launched the QQ mobile phone input method and Baidu mobile phone input method.

By 2017, only three leading companies, Sogou, iFLYTEK and Baidu, had more than one million MAU among all kinds of mobile phone input methods. In this, Sogou Input also occupies an absolute advantage, his MAU is iFLYTEK and Baidu products MAU combined twice [13]. According to the report of Analysys data, by November 2020, the penetration rate of deduplicated users of Sogou, iFLYTEK and Baidu exceeded 95%, while Sogou alone accounted for 58.4% [14].

In addition, although Toubao, which is still a "free body", has been listed, it has also lost its living space at home and has moved overseas to become a paid product.

Input method has the potential to become a "big business". Compared with social networking, games and other categories of App, tool-based input methods are inherently highly viscous and permeable. According to the data of CNNIC, by December 2020, China has 986 million mobile Internet users [15]. In QuestMobile's 2020 China Mobile Internet Annual report, the number of active users of the input method App has reached 882 million, with a penetration rate as high as 89.5% [16].

Why is there only room for a few giants when the number of users is so large?

Who does the input method for serious people? Even though the input method industry has developed in China for nearly 30 years, there are still many problems to be solved, the biggest problem is that it is difficult to commercialize.

The business model and technical model of the "toy" input method of large companies have determined that this is a "toy" that rich people can afford to play with.

On the one hand, technological progress has brought new changes to the input method industry, and the rules of the game are becoming more and more unfriendly to small companies.

Ten years ago, the input method paid more attention to how to make typing faster and more accurate. Some companies have developed double input, sliding input and Taper 1 keyboards to optimize the text input experience, while others have worked hard on word selection, introducing personalized thesaurus such as regional thesaurus, professional thesaurus and address book thesaurus.

Ten years later, the input method is increasingly testing the company's AI technology.

With the progress of artificial intelligence technology, users are no longer satisfied with the traditional input method. Voice input, translation input, emoticons and other functions are more popular with users.

According to a survey conducted by Analysys data in 2019, when choosing a third-party input method, 47.5% of users will consider whether voice input is easy to use, and 46.6% of users will consider whether Doutu emoji is desirable [17]. In addition, intelligent functions such as intelligent assistant, voice change, AI writing assistant and OCR scan input have also become high-frequency functions for generation Z users [14].

And these technologies are often only available to big companies.

Among the three input method giants Sogou, Baidu and iFLYTEK, Baidu and Sogou are the earliest companies in China to set foot in artificial intelligence search engines, while iFLYTEK started with speech recognition and semantic recognition technology.

On the other hand, the profit model of the input method is still a mystery.

On the PC side, the input method has always been a "basically unprofitable" free product, and its profit model has almost no payment function except for occasional pop-up ads. QQ and Baidu input methods, which rely on the two Internet giants, live smoothly, but small companies either choose trees, are acquired by big companies, or simply go out of business.

A few manufacturers have tested the advertising functions of paid skin and input methods, but they immediately aroused strong disgust among users.

Wang Xiaochuan's "three-stage rocket" model was once highly expected by the input method industry, and he found that although the input method could not make money directly, it could be used as a traffic portal for blood transfusion for other profitable businesses.

As a result, Sogou launched the browser in 2008 and established a product system based on "input method-browser-search engine"-- using Sogou Input (first-stage rocket) to drive the installation of Sogou browser (two-stage rocket). And then through the browser to Sogou search (three-stage rocket) with quantity.

Sogou's "three-stage rocket" model will basically take shape before and after 2012. This year, nearly half of Sogou's search traffic came from Sogou browser [18]; this year, Sogou's revenue was $131 million, an increase of 108% over 2011 [19].

But in the final analysis, the three-stage rocket model is only a diversion means, profit pressure all depends on Sogou search advertising. Later, the domestic search engine business was gradually monopolized by Baidu, Sogou search market share is getting less and less, this model is also out of order.

Sogou has been eating the old capital of the search business in recent years. According to the financial report, Sogou's total revenue in 2019 was 1.17 billion US dollars, of which the contribution rate of search advertising business exceeded 91.5% [20].

Finally, Sogou did not find a better way to make a profit and was finally acquired wholly by Tencent.

In recent years, the input method has also looked for other ways for mobile phone manufacturers.

Sogou has cooperated with Xiaomi, OPPO and vivo to build a deeply customized Sogou Input on some models. Baidu is not to be outdone. In addition to the above three manufacturers, it has also reached an exclusive cooperation with Huawei, which uses a customized Baidu input method for all models.

However, cooperation with manufacturers can not bring real benefits to the input method company.

On the one hand, although cooperation with mobile phone companies can quickly bring active users, it does not solve the profit dilemma of the input method itself. On the other hand, input method companies have to beware of mobile phone manufacturers coming back with self-developed input methods. In November 2021, Huawei launched the self-developed "small Art input method", while the "Jovi input method" developed by vivo is also constantly updated, and the application market is a warning for the input method.

In order to commercialize, some input methods secretly start the business of "reselling data", which brings privacy protection problems to the input method industry.

Where is the user data? on January 19, 2021, Zhang Xiaolong said on Wechat Night that because users frequently complained that their chat records had been stolen, they decided to develop their own input method and launch the input method in order to protect their privacy.

Zhang Xiaolong's words put other input method companies in an unfair place.

User data has always been a "secret corner" in the input method industry. Even if the input method company declares to the user over and over again, "We will properly dispose of the user data and protect the user's privacy", the user still has a bad feeling about the input method.

As the largest information distribution center in users' mobile phones, the input method has established data cooperation relationships with advertisers and analysis service providers. If you open the privacy policy of any input method App, you will see a description similar to "some of the user's information will be shared with partners."

The so-called "partial information" is not the sensitive data of the user, but desensitized forms such as "user profile" and "user tag", which are used to recommend personalized advertising [21].

For example, JD.com and Sogou signed a cooperation framework called "Beijing search Plan" in 2017. JD.com said at that time that cooperation with Sogou will get more traffic entrances, which can improve the accuracy of delivery [22].

Because of this, many users will receive relevant product recommendations on e-commerce App for what they say to their friends.

On December 30, 2019, the State Internet Information Office, the Ministry of Industry and Information Technology and other four departments jointly issued the method for determining the illegal collection and use of personal information by App, which mentioned that App shall not "violate the necessary principles and collect personal information that has nothing to do with the services it provides." [23]

At the beginning of 2021, the Ministry of Industry and Information Technology was "particularly concerned" about the input method industry and suddenly announced 26 App that infringed upon the rights and interests of users. Among them, QQ Input, fast input, palm input method, WeChat business input method four input methods are removed from the shelves for violating the personal information of mobile phone users [24].

It's just an appetizer.

On April 26, 2021, the Ministry of Industry and Information Technology issued the interim provisions on the Administration of personal Information Protection of Mobile Internet applications (draft for soliciting opinions), in which Article 7 stipulates that "those who engage in App personal information processing activities shall have a clear and reasonable purpose, and shall follow the principle of minimum necessity, and shall not engage in personal information processing activities that go beyond the scope of user consent or have nothing to do with the service scenario." [25]

Five days later, the State Internet Information Office issued a circular on the illegal collection and use of personal information in 33 App, including input methods, among which 15 input methods App illegally obtained and exceeded the scope of personal information collection. Sogou, iFLYTEK, Baidu, QQ and other mainstream input methods App are not spared [26].

As a result of stricter supervision, the data management of input methods is becoming more and more standardized. In Sogou Input's "explicit list of collecting personal Information" and "third-party Information sharing list" updated in April 2022, there is no information sharing that can be used for personalized advertising.

Since then, the data business of the input method has ended and has officially become a public welfare product of the Internet giant, guiding the flow of its own products at most. In the latest grayscale test of Wechat keyboard in September 2022, intelligent recommendation and Mini Program, official account sharing and other functions have been provided.

But is that what Wechat is all about? Apparently not.

The "Vanity Fair" of the mobile phone industry, Wechat, the largest social media in China and with more than 1.2 billion monthly active users, holds private data on almost all of China's Internet users. This is an achievement beyond the reach of social products, and it is also a "hot potato".

The input method used for chatting and typing is the only outlet of chat information except Wechat. With the increasingly stringent management of user privacy in the country, it is necessary for Wechat to cut off from these "risks", and self-developed input method is the best way.

It is not difficult for Wechat to develop input methods.

First, Wechat App already includes voice input, voice transcription, text translation, emoji search and other functions, which is only one keyboard short of a complete input method; second, Tencent has acquired Sogou, and both Sogou Input and QQ Input's team have ready-made technology and experience for Wechat team to learn from.

Wechat also does not have to consider the investment-to-return ratio of developing input methods, after all, there is no small matter in front of privacy. As for how to make a profit, it is a question that should be considered in information flow advertising.

So isn't this a loss of jobs for your peers? Yes, but the first job has been ruined by Apple.

In 2021, Apple limited App's ad tracking function in iOS 14.5, destroying jobs for Google, Amazon and then Facebook. Although peers complained and even announced a lawsuit against Apple, users applauded.

Today, however, the protagonist of the event has been replaced by Wechat and the battlefield has become an input method, but the mobile phone industry is still a vanity fair of privacy and profit.

References:

[1] Wang Ma, vice president of the China Invention Association, the inventor of the five-character character, "Contemporary Bi Yi"-Wang Yongmin http://www.wangma.net.cn/ wangyongmin.aspx?sm=10

[2] Wang Ma and Zheng Ma November 8, 2014 https://www.qqxiuzi.cn/ wz / zixun / 1163.htm

[3] Wangma Wubi font version 86, 98, new century edition simplified character root bitmap http://www.wangma.net.cn/ InfoMationDetail.aspx?sm=5&m=41

[4] although typing is fast, typos are fined 50 times to review the development history of input method, September 7, 2011, https://soft.zol.com.cn/ 247/2476547_all.html#p2476650

[5] Ma Zhankai: the Story of the Father of Sogou input method https://blog.csdn.net/ anzi57066381 / article / details / 1819943

[6] Tencent QQ Pinyin Input Method formally sued Sogou for unfair competition in Pinyin input method https://tech.qq.com/ zt / 2009 / qqpinyin09 / index.htm

[7] the Pinyin input method competes to enter the white hot 2009.7.1 https://www.chinanews.com/ it / it-itxw / news / 2009 Ultimate 07-01/1756009.shtml

[8] Sogou sued Baidu input method for 8 patent infringement claims of 80 million yuan https://www.yicai.com/ news / 4702489.html on October 26, 2015.10.

[9] Baidu formally sued Sogou Input for 100 million yuan http://news.imobile.com.cn/ articles / 2016/1028/172563.shtml on October 28, 2016.10

[10] for "traffic hijacking" sued by companies such as 360 and Baidu, the search candidate function was convicted of unfair competition. Sogou Input robbed 2019.7.13 http://www.cb.com.cn/ index / show / bzyc / cv / cv13418711641

[11] the new version of Windows 10 is specially aimed at the Chinese market to improve the simplified Chinese input experience of https://www.51cto.com/ article / 524434.html on December 8, 2016.

[12] Android mobile phones account for 48% of the global smartphone market. Http://www.techweb.com.cn/ news / 2011-08-02/1075173.shtml

[13] Aurora big data: 2018 input method App Industry Research report 2018.31 http://www.199it.com/ archives / 766638.html

[14] China's third-party input method market development analysis 2021.1.11 https://www.analysys.cn/ article / detail / 20020025

[15] the number of netizens in China has reached 989 million 2021.2.4 http://www.xinhuanet.com/ tech / 2021-02 Universe 04 / c_1127061748.htm

[16] QuestMobile2020 China Mobile Internet Annual report: changes give birth to new life, new "consumer touch matrix" set off sales and marketing changes 2021.1.26 https://mp.weixin.qq.com/ s/_y6d43VvzRoL_OLndSREWQ

[17] Special analysis of China's third-party input method market 2019 2020.1.3 https://www.analysys.cn/ article / detail / 20019622

[18] Sogou's surname is Sou, Sohu's search is August 2, 2017.127351 https://www.pingwest.com/ a

[19] interpretation of Sohu's 2012 Financial report: hidden worries under the record 2013.2.6 https://www.huxiu.com/ article / 9979.html

[20] Sogou Financial report https://www.sec.gov/ Archives / edgar / data / 1713947 / 000110465920049011 / A20-15253_120f.htm

[21] will the input method "betray" us? 2021.1.31 https://www.infzm.com/ contents / 200787

[22] JD.com and Sogou push the Beijing search plan to fully open the enterprise's core competence on October 20, 2017.20 https://tech.huanqiu.com/ article / 9CaKrnK5y44

[23] notice on the issuance of "methods for determining the illegal Collection and use of personal Information by App" 2012. 12. 30 http://www.cac.gov.cn/ 2019-12 http://www.cac.gov.cn/ 27 / c_1578986455686625.htm

[24] App report on violations of user's rights and interests by illegally invoking microphones, address books, photo albums, etc. (2nd batch of 2021, total 11th batch) 2021.2.5 https://www.miit.gov.cn/ jgsj / xgj / gzdt / art / 2021 / art_7893db5f4a4d403192f69c72fc0320d9.html

[25] Open soliciting comments on the interim provisions on the Administration of personal Information Protection of Mobile Internet applications (draft for soliciting opinions) 2021.4.26 http://www.gov.cn/ xinwen / 2021-04lap26 / content_5602780.htm

[26] Bulletin on the illegal collection and use of personal information of 33 App, such as input method, 2021.5.1 http://www.cac.gov.cn/ 2021-04 http://www.cac.gov.cn/ 30 / c_1621370239178608.htm

This article comes from the official account of Wechat: fruit Shell hard Technology (ID:guokr233), author: Yang Jingyi, Editor: Li Tuo

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.