When you swipe Taobao, when you click on a product to view the details, Taobao will continue to recommend similar products to you based on your browsing records; the same is true when you swipe Douyin, Douyin will remember what you like Type of video, it will accurately recommend the videos you like to watch for you every time; when you haven’t used Maimai for a long time, Maimai suddenly sends you a text message saying “The XXX you follow has checked your resume, click XXX link to view".
These companies constantly rely on guessing your preferences, recommending your favorite products to you, stimulating your telemarketing list dopamine, so that you can’t stop brushing; at the same time, these companies continue to make profits and expand their marketing scale.
How did they do it? The answer is to create user tags by collecting customer attributes and behavioral characteristics, and then create different user portraits through user tags, and then provide different marketing content for different users through user portraits.
So how to achieve precise marketing and increase corporate revenue by establishing customer tags?
1. What is a customer tag?
A customer label is actually a description of a customer's characteristics. If we have many customers, in order to be able to serve so many customers better and more accurately, we need to let the business know the characteristics of each or each group of these customers.
For example, if some people like strawberry-flavored ice cream, you can recommend strawberry-flavored ice cream to them; some people don't like ice cream, so try to avoid recommending ice cream to this group of people.
In short, I hope that the business can know which people are the key service targets in which scenario, and spend their energy on the cutting edge. Targeted operations can greatly increase the customer's purchase rate and avoid unnecessary human consumption.
2. What customer information is required to build a customer label?
Different business scenarios also have slightly different collections of customer information. Here is an information framework that can be covered by most business scenarios, that is, some common customer information that needs to be collected. You can find the customer information that needs to be collected most according to different business scenarios.
1. Information of individual customers
2. Information of corporate customers
The customer information collected above is only raw data, and if the data is not analyzed and modeled, the information will become worthless. So how to build a customer label system through customer information?
3. Classification of customer tags
Customer tags are based on business scenarios, and customer tags constructed in different business scenarios are also slightly different. But it's all the same in big categories. A common practice in the industry is to classify labels into three types: factual labels, model labels, and prediction labels.
Raw data is the customer information we collect most directly. Based on the original data, gradually build labels at different levels.
1. Fact label
The factual label is the basis of all labels. It refers to the establishment of labels according to the actual situation and actual behavior of customers. Fact labels are mainly divided into demographic attribute labels and customer behavior attribute labels. For example, a person's gender, age, date of birth, place of origin, etc. are demographic attribute labels. A person's consumption times, consumption amount, etc. belong to customer behavior labels.
What data to collect from customers needs to be defined by product managers or product operations based on their own products and customer needs. For example, in educational products, you need to know the age, gender, view courses, favorite courses, etc. of customers.
2. Model Labels
Model tags, also known as rule tags, are some telemarketing list rules defined by product managers or product operations based on customer fact tags. For example, if a customer buys products 3 times, the shopping behavior itself is a factual label, but purchasing products 3 times can be classified as an old customer. The old customer here is a model label.
The definition of the model label requires the product person to have a deep enough understanding of the industry and customers to be able to accurately define it, and is generally defined by product experts.
Here are a few common mainstream model tags that have been defined by the seniors. In many cases, we can use them directly.