Digitalization without user portraits is just a decoration.
User portrait is user labeling. Labeling is to facilitate the identification of users, and it is automatically identified by computers.
Only by accurately identifying users can we provide accurate services to them. Give me a pillow when you feel sleepy.
Automatically identify users, automatically complete instant information and policy push, and automatically form a new user profile the moment the user changes their shopping behavior.
The power of digitalization can only be demonstrated through user portraits.
02
In the past, offline sales were “one-to-one” and “face-to-face”, whether on the B-end or the C-end.
One-on-one communication provides ample time for interaction. If it is someone you know, you can recommend based on their preferences. KA stores only allow users to choose by themselves.
Nowadays, when users are online, they need to face massive numbers of users at the same time. For example, on Double 11, hundreds of millions of users are online at the same time, tens of millions of merchants are online at the same time, and hundreds of millions of SKUs are online at the same time. And instant push and instant update are required. So, how can we quickly generate products or content that meet user needs every time the user clicks, without making the user feel any time lag during the whole process?
The answer is user portraits automatically completed by computers.
The platform automatically matches product portraits and user portraits, completes the task instantly, and pushes messages instantly.
It’s just that the user portraits of various Internet platforms are completed by the platforms, and users and merchants are unaware of it.
When it comes to digital marketing, fast-moving consumer goods manufacturers cannot rely on platforms, but must rely on the brands themselves to complete user portraits.
03
User portrait, a professional term. Ordinary people don’t understand, nor do they need to understand. Even if users have been profiled, they are unaware of it. In fact, as long as you open various e-commerce platforms, Douyin, and Toutiao, the user has been profiled by the platform. Otherwise, how could the platform push products or content so accurately?
Whether it is a trading platform or a content platform, the platform is a matching system, that is, it matches products and users and recommends products to the most suitable users at the right time.
Matching requires precise matching, and the push notifications are exactly what the user needs. Only when there is a match can the transaction rate (conversion rate) be high.
There are hundreds of millions of users and hundreds of millions of SKUs on the Alibaba platform. Without matching products with user portraits to achieve instant push, wouldn’t the matching process be a mess?
Toutiao and Douyin have a huge amount of content and a huge number of users. If users rely on searching and finding by themselves, it would be a waste of time.
Fortunately, we now have AI technology for user profiling!
User portraits are created based on the traces (data) left by users’ browsing, viewing, and trading behaviors on the platform, in accordance with certain portrait rules and purposes, and then content is pushed accurately based on the user portraits. For example, media such as Toutiao and Douyin will push content in a targeted manner based on user preferences and tendencies.
As the United States cracks down on TikTok, China promptly issues policies to ban the export of algorithm-related technologies. A very important function of AI algorithm is user profiling.
AI algorithms are used for both user profiling and product (content) profiling, and then matching the two. For example, the platform determines a user's demand preferences through user portraits, such as price preferences, brand preferences, etc. At the same time, the product also has a portrait, and then the product is matched with user preferences. Only when there is a match can the conversion rate be improved.
The platform has already created a portrait of the product and users, but why don’t users feel anything? This is because user profiling is done by the platform, merchants and users are just profiled, and the profiles are used to push products or content. Users just feel that the product is more palatable, but they don’t know that they are being profiled.
04
User portraits of 2C platforms have become a daily routine for the platforms. So, does the digitalization of fast-moving consumer goods manufacturers require user profiling?
Of course! Only user portraits can be more accurate.
For example, when a new product is launched on the market, in the past deep distribution would be adopted for comprehensive distribution. However, with product upgrades, the distribution of new products needs to be precise. If a high-end new product is launched on the market, it is necessary to find precise distribution terminals, so it is necessary to create user profiles for the terminals.
It is assumed that users who distribute high-end new products must meet three criteria: 1. The terminal has the ability to recommend new products; 2. The terminal has a high-end user base; 3. The terminal has an advantage in this category.
Then, we can create a terminal portrait based on the above standards, screen out terminals that meet the above conditions for cognitive education (such as experience), and then distribute the products. Because the portrait is accurate and sales are good after distribution, more comprehensive distribution can be carried out.
So, where does the data for new product distribution and terminal portraits come from?
There are two major sources:one is that if you have terminal data yourself, you can create a terminal portrait based on historical data; the other is that if you have just gone online and have no data, you can find a system platform or a third-party professional company, and they will create a terminal portrait based on data from other companies.
Whether it is a manufacturer, agent, or new retail, as long as it is online and digital, there must be a user portrait.
In today’s new retail scenarios, user profiling can be achieved even offline. For example, when a user is shopping in a store, when he walks to a certain place, the special screen will recommend suitable products based on the user portrait. For example, as long as a chain store recognizes the user entering the store (such as through facial recognition), it can actively guide the user to shop based on the user's portrait, which is more accurate than the passive shopping guidance in the past.
05
Even traditional sales have simple user portraits. For example, the cash register keyboards at 7-11 convenience stores are counted based on gender and age. However, this kind of portrait can only be a collective portrait, such as gender ratio and age ratio, and not an accurate portrait of a single user.
User portraits can be roughly divided into two categories: one is user attribute portrait; the other is user behavior portrait.
User attribute portrait, such as gender, age, income, interests and hobbies, active time, and place of residence. The high-end new product distribution terminal mentioned earlier also belongs to the attribute portrait of B-end users.
User attribute portraits can be used in product development, such as finding target users; they can also be used for product recommendations. For example, if the user portrait is "mom", then suitable products can be recommended to the "mom" user based on the mom's demand characteristics.
In the online recommendation system, users with the same profile are called “neighbors” and recommendations are made to users based on their “neighbors’” preferences.
Traditional marketing also does user profiling, but it is more about user attribute profiling. Big data is also used to profile user attributes, such as new product development and B-side user profiles. However, the big data portraits for C-end users are more behavioral portraits.
If user attribute profiling is to "guess" user behavior based on the profile, then user behavior profiling is to predict the next behavior based on previous and current behavior.
Once the behavior has occurred, it is relatively easy to predict the next behavior.
The most important portrait in digital marketing is the behavioral portrait. Behavioral portraits are different from attribute portraits. Attribute portraits have a certain degree of stability because gender, age, and interests and hobbies are also relatively stable. However, even for the same person, behavior changes greatly. For example, a TikTok user may like content today, but want to change his taste tomorrow. Then, as long as the user makes a change, the user behavior profile will change immediately.
Many users of Toutiao have complained that Toutiao has solidified their preferences and they actually want to see more new things. However, even if you "want to see" it, the portrait will not change unless you take action. As soon as the user attempts to make a change, the portrait changes immediately.
So, what is user behavior?
On Toutiao, user behaviors include clicked content, reading time, likes, comments, etc. Profiles are created based on these behaviors, and then the content to be pushed in the future is determined.
Amazon is the originator of using user portraits for recommendations. Amazon uses user behaviors on the site, including browsing items, purchasing items, adding to favorites and wish lists, as well as user feedback such as ratings, to form user portraits, which are used for the following purposes:
1) Recommendation of the day. Based on the user's recent browsing and purchasing history, a comprehensive recommendation is given in combination with currently popular items.
2) New product recommendations. Adopt a content-based push mechanism to recommend some new products to users. Since new products have less user preference data, content-based push solves this problem.
3) Related recommendations. Use data mining techniques to analyze users’ purchasing behavior and find sets of items that are often purchased together or by the same person. When purchasing books, there are many such recommendations.
4) Others purchase/browse items. This is a collaborative filtering recommendation of items. Through social mechanisms, users can find products of interest more easily.
06
User portrait is the soul of digitalization. It is hard to imagine that digitalization can be done well without user portrait.
Traditional marketing also has numbers, but they are all statistics, such as annual sales, daily sales, etc. These numbers are useful, but not very valuable for online development.
The online nature of users has put forward a requirement: instant recommendations. Every online behavior of the user, including browsing, shopping, rating feedback, group buying, etc., is changing the user portrait. In other words, user behavior portrait is a dynamic portrait.
Every online action of the user will be followed by the next action. Between the two actions, they must decide what product or content to push next. Therefore, in an online environment, user portraits must be able to make instant recommendations.
Online instant recommendations require that user portraits be modeled (modeled) based on original data, calculations be completed instantly, and then products, content or policies are pushed. The page update after each click on content platforms such as Douyin is the result of user portrait calculation; every page click on Alibaba and Pinduoduo is the result of instant calculation and push of user portrait.
The digitalization of fast-moving consumer goods manufacturers has basically only been completed online, without modeling, profiling, or instant push capabilities. For example: one item, one code, and now most people just send red envelopes indiscriminately. Without user portraits, users will be treated as having the same portrait (the same appearance).
User portrait is a big topic and will be discussed repeatedly. The next issue will discuss the differences between user portraits from the perspectives of brands and retailers.
Author: Liu Chunxiong
Source: Liu's New Marketing (ID: liuchunxiong1964)
