8 Big Data Analysis Models - User Model (I)

A model refers to a formal expression of a practical problem, objective thing or law after abstraction. Any model consists of three parts: target, variables, and relationships. By clarifying the variables and changing them, you can directly present the results and achieve the goals. In daily data analysis , we often use 8 major models (user model, event model, funnel analysis model, heat map analysis model, custom retention analysis model, stickiness analysis model, full behavior path analysis model, user grouping model). Starting from today, we will interpret a model every Tuesday. This article starts with the user model.

1. What is a User Model?

Let me first use three sentences to explain why the user model is the basic analysis model. It is so important that it should be analyzed first: because if you don’t know who your users are, you don’t know what services to provide; if you don’t know what stage your users are in “interacting” with you, you can’t know what services to provide first; the marketing strategy cannot be focused, and the service is not systematic and continuous. Therefore, let’s start with the definition, popularize what the user model is and how to build a user model in the traditional way.

Persona is a systematic method of studying users mentioned by Alan Cooper in the book About Face: The Essence of Interaction Design. It is an important tool for product managers and interaction designers to understand user goals and needs, communicate with development teams and stakeholders, and avoid design pitfalls.

Traditional user model construction method:

Alan Cooper proposed two methods for building user models:

-User model : It is established based on research results such as interviews and observations of users. It is rigorous and reliable but time-consuming.

- Temporary user model : built based on the understanding of users by industry experts or market research data, which is fast but easily biased.

1. Build a user model based on interviews and observations (orthodox approach)

In Alan Cooper's method, interviews and observations of users are an important basis for building user models. The complete steps are as follows:

2. Build an ad hoc persona

When there is a lack of time and resources to interview and observe users, a "temporary user model" can be established based on industry experts' understanding of users or demographic data obtained from market research.

The process of building a "temporary user model" is very similar to that of building a "user model", except that the data basis of the "user model" comes from interviews and observations of real users, while the data basis of the "temporary user model" comes from the understanding of users. There are differences in accuracy and precision between the two.

2. Build a user model based on behavioral data

It has been nearly 20 years since Alan Cooper first proposed the concept of user model (Persona). During this period, the process methods of software product development and the way companies operate have changed a lot: the agile development method characterized by rapid iteration has replaced the traditional waterfall model, and the lean startup method with the "development → measurement → cognition" feedback loop as the core is gradually influencing and changing the way companies operate.

The traditional user model building method has not changed much since its inception. For product managers and interaction designers who are used to agile and fast work, spending a long time studying users and building user models requires considerable determination and effort to obtain the required time and resources. In addition, the time it takes to cold-start an Internet product is getting shorter and shorter. In order to reduce costs and risks, product teams often choose to push products to users as quickly as possible during the startup period and obtain feedback as quickly as possible for "quick trial and error." Reality and pressure force most PMs of new products to not invest a lot of time and energy in in-depth user research.

It is easy to understand why everyone thinks that user models are good, but few people actually use them in their work. In order to resolve the contradiction between time pressure and lack of energy , we proposed a lightweight method for quickly and iteratively building user models based on user behavior data.

First, organize and collect any recognizable user experience and data that has been obtained, including: you and your team's understanding of users; user-related information recorded in the product's business database (such as the user's gender, age, level, and other attributes); and any forms filled out or information left by users (inside or outside the product) (such as questionnaires filled out by users, WeChat accounts left behind, etc.).

We map this information into user description information (attributes) or user behavior information and store it to form a user profile (as shown in the figure below).

Zhuge.io New Retail DEMO User Profile (Virtual Data)

As shown in the figure above, we can clearly understand the user's attribute information, behavioral data, and environmental data from the user profile.

3. Advantages of Building User Models Based on Behavioral Data

1. Efficient and real-time insight into opportunities

In the world of data, accuracy is everything, and speed is even more crucial. The faster the analytical system processes and interprets this information, the faster and clearer it can grasp the business status, helping companies make decisions earlier. For example, one of our clients, a shared bicycle , discovered abnormal fluctuations in real-time data indicators: the number of retained users on the next day fell "cliff-like". An emergency investigation revealed that competitors were attracting new users at low prices. Therefore, the operations team took active measures at the first time, thus maintaining the market share in the city. Similarly, the market is constantly changing, and operators and decision makers need to pay attention to fluctuations in their own data in real time, because failure often comes from a minor negligence.

2. Record history, not just results

Behavior is a label. In the past, we often gained user insights by labeling users. In fact, behavioral data itself has become increasingly valuable. The user model based on user behavioral data records every behavior of each user and objectively and truly restores the interaction process between the user and the product. Compared with simply marking "user tags", the recorded user behavior data has more value in multi-dimensional cross-analysis, and the constructed single user portrait is more complete and scientific.

3. 360° user profile covering the entire user life cycle

The user model based on user behavior data changes dynamically in real time. As the user grows in the product, from a visitor to a stranger and finally becomes a high-value user, every step of the user's growth is recorded through behavior. Based on the different stages of the user's life cycle, targeted operational strategies such as attracting new users, conversion , and retention are adopted for new users, lost users, active users, and silent users.

In order to extend the user's lifetime value (LTV), it is necessary to collect data on the user's entire life cycle, connect CRM data, historical data, business data, and third-party data, and associate the user's attribute information (gender, age, country, etc.) with the user's behavior data; connect the data of external promotion platforms to solve the problem of where the users come from; connect the data of different product platforms, and synchronize the user's behavior on the app mini program micro site official website in real time. Only in this way can truly user-centric statistics and analysis be achieved.

This article provides a method for quickly and iteratively building a user model (Persona) with the help of behavioral data and tools. It is more suitable for the working methods and rhythm of today's Internet teams. The user model based on user behavior data, on the one hand, simplifies the traditional method and lowers the threshold for data analysis; on the other hand, it makes data analysis more scientific, efficient, and comprehensive, and can be more directly applied to business growth and guide operational decisions.

The author of this article @朱葛君 is compiled and published by (Qinggua Media). Please indicate the author information and source when reprinting!

Product promotion services: APP promotion services, advertising platform, Longyou Games