E-commerce search personalized recommendation "kill familiar" and annoying
Our reporter Chen Xuening
Different people use the same keywords to search for products on the same e-commerce platform, and the price recommended by the system can be several times worse... Such "personalized recommendations" are receiving more and more complaints and complaints.
Consumers appealed that, on the basis of optimizing machine algorithms, the platform should also provide shutdown options in accordance with the requirements of laws and regulations, and return the right to know and the right to choose to consumers.
The average price difference between mother and daughter searches for the same product is 4 times
"How come there are fewer and fewer cheap things on Taobao." As a senior user with 4,626 registration days and 3,906 orders, Wang Hui has always found it difficult to find cheap small items when using mobile Taobao recently.
"Many of the nail clippers set used to be 9 yuan and 9 free shipping. Now, when I search, the top one is 79 yuan, and the ones below 10 yuan are very rare."
Are prices rising so fast?
When she accidentally used her mother's mobile phone, she discovered that it was obviously the same keyword, and the average price of the goods searched out was significantly lower. The price of 10 to 20 yuan abounds.
The reporter learned that the consumption concepts of the mother and daughter are different. Wang Hui usually buys goods at a higher unit price, while his mother always chooses the cheaper one.
Different consumption habits and user behaviors lead to very different personalized recommendations.
In order to confirm Wang Hui's feedback, the reporter used two different mobile phones to experience it.
The first mobile phone was from a 32-year-old advertising practitioner, and the second mobile phone was from a 61-year-old corporate retire.
Using "vacuum cleaner" as a keyword search, the reporter found that the average unit price of the former was 1336.2 yuan, while the average price of the latter was only 244.5 yuan, which was more than four times the average price.
The reporter used the above two mobile phones to search for computer chairs on JD.com, and the price gap was more than three times.
"If you have high spending power, don't let you eat food stalls, and if you have low spending power, don't you deserve to go to the counter?" Wang Hui questioned, the personalized recommendation of the e-commerce platform has become a kind of "killing familiarity" in disguise.
Repeat the recommendation algorithm intimately to bother
The “E-Commerce Law” stipulates that personalized recommendations have a clear scope, that is, e-commerce operators provide consumers with search results for goods or services based on their interests, hobbies, consumption habits, and other characteristics.
Because the information pushed by merchants has a higher relevance to consumers, it can save the cost of information search. Therefore, when this feature first appeared, it once gave people a sense of shine.
However, as application scenarios become more extensive and penetration deeper and deeper, more consumers complain that personalized recommendations have changed from "intimate" to "interrupting."
I have bought a computer desk, but the homepage is still recommending similar products; the Internet celebrity furniture that has been clicked but not purchased will appear on the recommendation page in a few days... The endless repeated recommendations make the shopping cart of the e-commerce platform become a shopping cart "Information cocoon room", "has been spinning around in similar products."
What more people complain is that nowadays, personalized recommendations do not only appear on a specific page, but fully penetrate the entire chain of online shopping.
Adding to shopping cart, favorites, payment, confirmation of receipt, evaluation...Almost every page has "enthusiastic" recommendations from machine algorithms.
8 clicks to turn off "Personalization"
Article 18, paragraph 1 of the "E-Commerce Law" stipulates that if an e-commerce operator provides consumers with search results of goods or services based on their hobbies, consumption habits and other characteristics, they shall also provide the consumers with non-targeted personal search results. Feature options, respect and equal protection of consumers' legitimate rights and interests.
Why draft such a clause?
The interpretation of the e-commerce law clearly mentions that in more cases, due to the huge amount of information on goods and services on the platform, coupled with the bidding ranking service provided by the platform operator, this kind of accurate information may actually limit Consumers’ right to know and to choose.
However, the reporter's investigation found that the current mainstream shopping software mainly provides the option of turning off personalized advertisements, and only a few platforms have the function of turning off personalized content recommendations.
For example, both Pinduoduo and Jingdong can only close the former, while Taobao can both close.
Take Pinduoduo as an example, you need to click "FAQ" in the settings to find the "Personalized Ads" menu.
Open the page, the whole article is introducing "what is the benefit of turning on personalized ads for me?", only the bottom of the small word prompts "how to set up personalized ads".
In other words, after opening the App, after 8 clicks, personalized ads can be closed.
Jingdong and Taobao need to turn off personalization options in their privacy settings.
In addition to e-commerce platforms, news and information apps also need to put a "tight band" on personalized recommendations.
At the end of 2019, the Office of the Cyberspace Administration of China issued the "Methods for Identifying the Acts of Collecting and Using Personal Information in Violations of App Laws and Regulations". The use of users' personal information and algorithms to directional push information but not the option of non-directional push information is a violation of regulations.