Digital Marketing: 4 Commonly Used User Segmentation Models

This article introduces four commonly used user segmentation models for digital marketing: RFM model, AIPL model, 5A model, and user life cycle model. It is suitable for readers who are interested in digital marketing.

After the traffic dividend, refined operations have become a compulsory course for corporate digital transformation.

The essence of refined operations is to conduct differentiated operations for different users, so as to maximize ROI and improve user LTV. For example, in the customer acquisition stage, new users should be stimulated to convert as much as possible, while for long-term active and loyal users, there is no need to invest too many resources in the short term.

When it is identified that a user is about to churn, it is necessary to appease and recall them as soon as possible. After all, the cost of recalling an old user is several times lower than the cost of acquiring a pure new user. So, in the process of refined operations, or when CDP (customer data platform) is building a user portrait label system, what are the commonly used models for user stratification?

1. RFM Model

RFM is the most typical and commonly used model in the field of data mining and user value stratification. This model counts the R value, F value, and M value of each user from the perspective of user consumption time, consumption frequency, and consumption amount, and then divides the corresponding threshold intervals according to business attributes, so as to correspond users to different intervals, thereby dividing users into 8 types of user value, namely: important value customers, important return customers, important deep cultivation customers, important retained customers, potential customers, new customers, general maintenance customers, and lost customers.

This threshold can be set based on algorithmic models and statistical analysis. For example, for high-frequency scenarios such as short videos and news information, users are likely to be lost if they are not visited for a month, while the holiday decision cycle for tourism products is long, which may take 360 ​​days. In addition, RFM is often used when dividing membership levels (general card, silver card, gold card, diamond card, etc.).

• R: Recency , which reflects the freshness of a customer's activity, whether they were active just now or a long time ago.
• F: Consumption frequency (Frequency) , which reflects a customer's loyalty, whether it is only once or always used.
• M: Monetary , which reflects the contribution of a customer, whether he is a freeloader or a whale user. For content products, it may not directly generate orders or monetary benefits. This M can be replaced by usage time.

Image source: Baidu Images

2. AIPL Model

AIPL is one of Alibaba's three major marketing models. It is a means of quantifying and linking brand crowd assets. According to the process from user cognition to loyalty to the brand, users are divided into four categories. When fine-tuning operations, differentiated operation strategies can be implemented for users at different stages.

Image source: Baidu Images

• A (Awareness): represents the group of people who are aware of the brand, generally referring to the group of people who have passive contact with the brand, such as users who are reached and exposed to brand advertisements through various traffic channels.
• I (Interest): stands for people who are interested in the brand, generally refers to people who actively come into contact with the brand, such as those who actively search for search engine keywords, click on advertisements, browse the brand/store homepage, participate in brand interactions, browse product details pages, search for brand words, receive trials, subscribe/follow/join, and add to cart.
• P (Purchase): represents the brand purchasing population, including those who have made purchases. The strict definition can be successful consumption, while a looser definition can be successful payment.
• L (Loyalty): represents brand loyal people, such as those who have made repeat purchases more than twice or have positive comments and sharing about the brand.

3. 5A Model

The 5A model is a marketing model proposed by marketing guru Philip Kotler in "Marketing Revolution 4.0", which is similar to the Pirate Model (2A3R) in user growth.

• A1 Aware: refers to customers who passively receive information, such as those who are exposed to brand advertising and search for category terms.
• A2 Appeal: refers to customers whose brand impression increases, such as those who click on ads, participate in brand interactions, browse product detail pages, and search for brand terms.
• A3 Ask: refers to customers who actively search for information out of curiosity, such as the number of people guided into the store, or click on customer service consultation, telephone consultation, etc.
• A4 Action: refers to customers who take actions, such as collecting, purchasing, and those who consume, use and provide after-sales service after purchase.
• A5 Advocate: refers to customers who are loyal to the brand and promote it, such as people who become fans and share products or content.

Image source: Baidu Images

4. User Lifecycle Model

Based on the process from user acquisition to loss and combined with the product life cycle, users are divided into potential users, new users, activated users, etc.

• Potential users : The target audience of the product, which is related to the product’s own positioning.
• Pure new users : users who have not completed the specified target action, such as users who have never placed an order in an e-commerce business.
• Activated users : users who have completed a specified action for the first time, such as first-order users.
• Mature users : frequent visits or repeat purchases.
• Declining users : visit frequency and contribution decrease.
• Dormant users : those who have not visited or placed an order for more than X days.
• Lost users : those who have not visited or placed an order for more than Y days.

When actually dividing the life cycle, we must first find the business processes and key indicators of the corresponding stages based on different business attributes, and then use algorithm models and RFM models to make comprehensive judgments.

Image source: Analysys Ark

Author: Qian Bingyi

Source: WeChat public account "Data Fan (ID: zhuangxiu1314)"