Personalized advertising and information protection, is it difficult to find a balance?
arts/
This is an era in which personalized recommendations are popular, with explosive growth of network data, and big data provides the possibility for personalized advertising. As a result, major Internet platforms can collect, analyze and define their historical behavior on the terminal according to the characteristics and preferences of users, and finally obtain user characteristics and preferences that conform to the rules of the platform, so as to recommend interesting information and products to users. ——This is the charm of personalized recommendation.
Personalized advertising has benefited many platforms, and the market seems to have acquiesced to this business model. However, with the conflict between personalized advertising and personal information protection in the Internet, under the high-level protection model of data protection legislation , Personalized digital advertising based on user behavior is facing changes and adjustments .
Two months ago, the "personalized recommendation" switch launched by major Internet platforms one after another is the result of the adjustment. So, does the "personalized recommendation" switch really solve the problem of personalized advertising and personal information protection? Under the commercial pursuit of profit, how should the boundary between digital advertising and privacy protection be established?
The Privacy Conundrum of Personalized Recommendations
From a technical point of view, personalized recommendation means that social media, search engines, and short video platforms use specific artificial intelligence rules to infer the user's possible preferences and selectively present content based on the user's usage behavior and data. Since the personalized information stream that the user is interested in is pushed, it can increase the user stickiness, drive the user's usage time, and obtain more advertising revenue.
A study by MAGNA and IPG Media Lab, "The Pursuit of Relevance," quantified the effectiveness of personalized advertising on brand KPIs and found that brands can build better relationships with people through relevancy in the form of personalized messages. Strong Connections – Increase purchase intent by more than 5% by using more relevant personalized messages and pre-play video ads.
In fact, since the advent of the first commercial online advertising in 1994, advertising is no longer just advertising itself, but a set of complete and reasonable market logic choices behind it - a free service based on advertising. It can be said that advertising has nurtured the digital industry and deeply transformed the marketing model of the business world.
From the most traditional web banner advertisements to the core business model of search engines - bidding ranking advertisements, from programmatic buying advertisements to information flow advertisements, especially the development of personalized advertising technology, the efficiency of advertising reach has been greatly improved. Digital advertising It has formed a unique competitive advantage in the advertising market, and has benefited a number of Internet platforms.
Among them, App developers represented by Facebook obtain traffic through free services and realize monetization through advertising business, that is, advertising + free service model. The "2020-2021 Global Advertising Industry Status and Trends Report" shows that global online advertising spending in 2020 will be US$303.3 billion, accounting for 54.4% of total global advertising spending, which is half of the total.
At the same time, personalized advertising also provides market development opportunities for the majority of small and medium-sized enterprises. For small and medium-sized enterprises, targeted digital advertising that can be billed according to the advertising effect greatly reduces the marketing cost and is a more reasonable choice of marketing plan. The personalized product information presentation method breaks the "28 laws" in the traditional market, fully activates the long-tail market, and allows a large number of small and medium-sized businesses to gain development space. Therefore, small businesses can promote their products and products to millions of people. Purpose of service.
However, personalized advertising has developed and promoted the prosperity of digital advertising, but it has also brought challenges to people's information protection - the pursuit of personalization and efficiency in the digital age is based on the collection and processing of personal information. . Obviously, under the algorithmic mechanism, users' personal data has become the most important nutrient in Internet business value. The platform can understand user preferences through user behavior data, and place more effective and suitable advertisements for them. Similarly, brands can make them information is seen by target customers.
The more detailed the platform understands about the user's core data, it means that the higher the match between the user and the product, the closer the user's real needs are. In order to understand the data in depth and dig deep into the needs of users, users need to accept the use of personal privacy to some extent, especially the privacy security crisis caused by the large-scale invasion of information.
Therefore, while the advertising algorithm mechanism brings a favorable situation for the three parties, user privacy has also become a problem that they have to face. After all, digital services that directly reach users, such as online shopping, social networking, travel, education, medical care, etc., all need to go through privacy policies, service agreements, and obtain users’ informed consent for the processing of users’ personal information.
Personalized recommendation not only brings considerable privacy issues, but also generates more discrimination, information cocoons, and the accompanying binary opposition and hostility on the Internet. Personalized advertisements configured based on personal preference algorithms seem to be user-tailored information to reduce unnecessary information recommendations, but in fact, specific labels have been affixed to each user.
The "freedom" in the "freedom of individual choice that still retains human beings" that Schoenberg believes in "The Age of Data Capital" is the freedom that has been screened. One of the key conditions for freedom of information is to have an information environment without human intervention and without screening, according to Keith R. However, under the personalized recommendation, there is no longer a purely free information environment, and users under the label can only enjoy the most suitable information and advertisements in a limited free environment.
Switches for Personalized Recommendations
It is difficult not to cause people to resist the difficulties caused by personalized recommendation.
After the EU General Data Protection Regulation came into force in 2018, a Dutch company prompted all visitors to its website in a clear and direct way, "whether to choose whether to share data with advertisers", and as a result, 90% chose quit. For this company, it is equivalent to effectively abandoning custom advertising revenue completely, and it needs to open up new revenue streams.
In addition, under the high-level protection model of increasingly stringent personal data protection legislation, personalized digital advertising based on user behavior is also in urgent need of reform and adjustment.
In January this year, China's "Internet Information Service Algorithmic Recommendation Management Regulations" was issued, which clarified that algorithm recommendation service providers should inform users in a prominent way of their provision of algorithm recommendation services; Users are provided with a convenient option to turn off the algorithmic recommendation service.
After the "Regulations" were promulgated, many mainstream apps took action one after another. As of March 15, apps such as Douyin, Toutiao, WeChat, Taobao, Baidu, Dianping, Weibo, and Xiaohongshu have launched an algorithm shutdown button, allowing users to turn off personalized recommendations with one click in the background.
However, although there are clear regulations, the platform must launch the algorithm shutdown button in a prominent position, but in actual implementation, each platform needs to go through tedious procedures to close the algorithm, and the product manager is also using some methods to allow users to obtain the right to choose. For example, the close button in WeChat needs to go through Settings - Personal Information and Permissions - Personalized Advertising Management to turn off personalized recommendations.
Moreover, as far as the current situation is concerned, a simple switch in the APP is obviously not enough: the personalized recommendation algorithm is still a black box for ordinary users, and the user does not know what kind of personal data is involved in the decision-making of the algorithm during the whole process. . In addition, after the personalization algorithm is turned off, whether the recommendation algorithm will still have basic personal information involved, etc., the implementation in different apps is also different.
The privacy page in the Google account or Apple ID settings provides users with more detailed settings than those apps that just provide a simple and crude switch . In the Google account or Apple ID settings, people can see how much personal data Google or Apple has collected about your purchase orders, browsing history, etc., and can also be corrected, deleted or exported in it, and even directly prohibit them from collecting the user's personal data data.
The more realistic situation is that even if the algorithm recommendation is turned off, it is easy for APPs to make consumers re-open obediently. When consumers find that the pushed things are not interesting to them, they can't even find what they want, and the user experience becomes very poor, people may give up this APP. At this time, if the APP pops up a sentence, "Did you not find what you want to watch, open the recommendation and try it?" "Open the recommendation and quickly get the content you want", perhaps most consumers will reopen it.
According to media reports, a senior user of Douyin said that he would not choose to turn off personalized recommendations. Said that if you really want to turn off personalized recommendations, "it's better not to play, and you won't want to watch it at all." From this point of view, the "personalized recommendation" switch cannot be said to be completely useless, but it seems to be of limited use.
An Approach to Balancing Personalized Recommendations with Privacy Protection
It has to be admitted that advertising, as a bridge connecting the supply side and the consumer side, is a reasonable existence in itself. In the digital age of information explosion, all kinds of personalized customization, including advertising, can overcome information overload and match more effectively. The supply and demand side obviously has unparalleled advantages.
From this point of view, personalized recommendation does not need to be changed for the sake of change, and the core purpose of the change is only to solve the problem of user privacy protection while providing a more efficient way of accessing information. Personalized advertising and user privacy should coexist, but only if we find ways to coexist.
In fact, traditional media also have "personalized recommendations", that is, how to make it more convenient for consumers. Youth Daily, Young Pioneer, Reference News, Readers, Half Month Talk, Storytelling, etc., each media, each magazine attracts specific readers by presenting unique content. These target readers pick up newspapers and magazines to read, and it is easy to find what they like to read.
An old man does not have to look for an international news among a lot of fairy tales, and a young pioneer does not need to look for a sample essay in a pair of articles about family relations and young people's marriage. More specifically, a newspaper will be divided into international, domestic, commentary, news, entertainment and other content. After the reader gets it, it is easy to directly turn to the content they want to see. Therefore, personalization is a deep-rooted and long-standing thing in business, and it is not realistic to rely on turning off personalized recommendation to combat the drawbacks of personalized recommendation.
More fundamentally, in the advertising scenario, the audience is not an individual but a demand. More precisely, it should be a collection of demands. Insight into demand through data mining is the essence of marketing and the direction of continuous improvement. The same person has different needs in different scenarios, and different people may have the same needs in the same scenario. These needs are predicted through data, and the advertising information is more accurately connected to the needs to influence consumption decisions. This is marketing data mining, which is essentially demand mining. rather than personal information mining . Advertising targeting to find suitable audiences is the direction and value that the advertising industry has consistently pursued.
In such a situation, the solution is to balance the problem of how to balance personalized advertising and user experience. On the one hand, while the Internet platform is making personalized recommendations, users should be given full and personalized right to withdraw, instead of just providing a simple and rude switch like the current one.
In fact, this was initially proposed in the "Technical Specifications for Mobile Internet Advertising Logos" issued by the China Advertising Association on December 10, 2020. "Technical Specifications for Mobile Internet Advertising Identification" aims to form a set of China Advertising Association Internet Advertising Identification Scheme (CAID), which requires limited uniqueness, stage consistency, connectivity, switchability, resettability, and anonymity. and stability .
Among them, the switchability requires that the CAID should allow users to open or close it in an effective way; the resettability requires that the logo should be reset in an effective and compliant way, and the reset logo cannot be associated with the reset. Anonymity requires that the identification should not be associated with any personal information, including generation parameters, service processes, etc.
On the other hand, for some personalized recommendation methods with poor user experience, users are given more detailed right to choose and reject. Closing personalized advertisements does not mean closing advertisements, but turning them into general advertisements. Many general advertisements are not what consumers need, and more redundant advertisements will appear in consumers' field of vision, which will also arouse consumers' disgust. Finding a more balanced solution between advertisers and users is worth further research in the future.
But all the prerequisites need to be built on the fact that manufacturers truly respect users and make users responsible for their own content intake. Obviously, there is still a long way to go before this day comes.