User behavior analysis of user operations!

From traffic marketing to data-driven, the refined operations of many products are centered around users, and the key lies in user research.

Common methods of user research include: situational investigation, user interviews, questionnaires, A/B testing, usability testing and user behavior analysis, among which user behavior analysis is one of the most effective methods of user research.

1. Understand user behavior analysis

User behavior analysis is the analysis of user behaviors on products and the data behind the behaviors. By building user behavior models and user portraits, we can change product decisions, achieve refined operations, and guide business growth.

During product operation, by collecting, storing, tracking, analyzing and applying user behavior data, we can find viral factors, group characteristics and target users that achieve user self-growth. This can deeply restore the user's usage scenarios, operating rules, access paths, and behavioral characteristics.

2. Purpose of User Behavior Analysis

Data-driven user behavior analysis is particularly important for products in industries such as Internet finance, new retail, supply chain, online education, banking, and securities. The purpose of user behavior analysis is to promote product iteration, achieve precision marketing, provide customized services, and drive product decisions. It is mainly reflected in the following aspects:

For products, it helps verify the feasibility of products, study product decisions, clearly understand user behavior habits, and identify product defects in order to iterate and optimize requirements.

For design, it helps to increase the friendliness of the experience, match user emotions, delicately tailor personalized services to users, and discover deficiencies in interaction in order to perfect and improve the design.

For operations, it helps to increase the effectiveness of exponential growth, achieve precision marketing, comprehensively explore user usage scenarios, and analyze operational problems in order to facilitate decision-making changes and adjustments.

3. Collect user behavior data

User behavior data actually has great commercial value. First of all, we need to clarify the data collection method in order to better support subsequent data analysis. Commonly used data collection methods include: platform setting tracking points and third-party statistical tools.

Setting tracking points on the platform is a very common collection method, which is to describe events and attributes in detail by writing code and distributing log points. Taking user login as an example, when a user logs in to the APP, related operations will be recorded and stored in the form of logs on the designated server.

Third-party statistical tools are generally connected through SDK, and we only need to build analysis models based on the indicators. Common third-party statistical tools include: Baidu Statistics, CNZZ Statistics, GrowingIO, Zhuge IO, Sensors IO, Google Analytics, Thinking Analytics, Umeng, Mixpanel, Heap, etc.

3. User behavior analysis indicators

The key to analyzing user behavior data is to find an indicator to measure the data. Based on user behavior performance, multiple indicators can be subdivided into three categories: stickiness indicators, activity indicators and output indicators.

Stickiness indicators: mainly focus on continuous visits during the user cycle, such as the number and proportion of new users, the number and proportion of active users, user conversion rate, user retention rate, user churn rate, and user access rate.

Active indicators: mainly examine the user access participation, such as active users, new users, returning users, churned users, average stay time, frequency of use, etc.

Output indicators: mainly measure the direct value output created by users, such as page views PV, unique visitors UV, number of clicks, consumption frequency, consumption amount, etc.

The purpose of these indicator segments is to guide operational decisions, that is, to optimize and adjust operational strategies based on different indicators. In short, the fundamental purposes of user behavior analysis indicator segmentation are: first, to increase user stickiness and enhance user awareness; second, to promote user activity and induce user participation; third, to increase user value and cultivate user loyalty.

4. Conduct user behavior analysis

After determining the user behavior analysis indicators, we can use some models to conduct qualitative and quantitative analysis of user behavior data. Commonly used analysis models include: behavioral event analysis, user retention analysis, funnel model analysis, behavioral path analysis and Fogg model analysis.

1. Behavioral event analysis

Behavioral event analysis is the analysis of user-specific events based on key operational indicators. By tracking or recording user behavior events, you can quickly understand the trend of events and the completion status of users.

Taking the user bidding behavior event as an example, the registration, authentication, account opening, recharge, investment and other behaviors performed by the lender in the bidding process can all be defined as events, which is also a complete event for the successful completion of the bid.

After determining the bidding behavior event, we can subdivide the dimensions according to the event attributes: user source, gender, date of birth, registration time, card binding time, first recharge time, first investment time, target ID, target name, term, interest rate, repayment method, etc. Then find out the rules that conform to the indicators and formulate targeted measures.

2. User retention analysis

User retention analysis is a model used to analyze user engagement and activity levels. Through retention volume and retention rate, we can understand the retention and churn status of users. For example, use indicators such as next-day retention, weekly retention, and monthly retention to measure the popularity or stickiness of a product.

Taking user retention of channel visits as an example, we conduct a retention analysis on channel users who have visited the APP. It can be seen from the figure that the next-day retention rate from August 14 to August 20 was above 41%, and the weekly retention rate was above 22%. However, on August 17, the next day, the retention rate suddenly soared to 67%. Such a high retention rate is usually due to event planning or function optimization.

User retention generally conforms to the 40-20-10 rule, that is, the next-day retention of new users should be greater than 40%, the weekly retention should be greater than 20%, and the monthly retention should be greater than 10% to meet business standards. We conduct user retention analysis mainly to verify whether the established operational goals have been achieved, which in turn affects the next product decision.

3. Funnel model analysis

Funnel model analysis describes the user conversion and churn rates at key links in each stage when users are using a product. For example, in daily activity operations, by determining the churn rate of each link, we can analyze how, why, and where users churn. Find the links that need improvement, focus on them, and take effective measures to improve the overall conversion rate.

Taking the funnel model of inviting investment as an example, the inviter shares the event page with friends, and then registers, authenticates, opens an account, and recharges to invest. The funnel model is used to analyze the conversion rates of some key nodes. Among them, the user registration conversion rate was 68%, the real-name authentication conversion rate was 45%, the card binding account opening conversion rate was 29%, the online recharge conversion rate was 17%, and the investment target conversion rate was 8%.

Funnel model analysis can verify whether the design of the entire process is reasonable. After comparison, it was found that the conversion rate from visit to registration was 68%, far lower than the expected 80%. The operation strategy this time is that users must register first before they can receive the new user benefits. Afterwards, we adopted the A/B testing method and optimized it to first receive the new user benefits and then induce users to register. After data comparison and analysis, the registration conversion rate increased by 20%. Therefore, by comparing the relevant conversion rates of each link, we can find out which links in the operation activities have not reached the expected conversion rate, so as to find out the problem and find the optimization direction.

4. Behavioral Path Analysis

Behavioral path analysis is to analyze the user's access path during product use. By analyzing the data of behavioral paths, we can find out the functions and usage paths that users use most frequently. And through multi-dimensional analysis of the page, track the user conversion path and improve the product user experience.

Whether it is a product cold start or daily activity marketing, the first step in doing behavioral path analysis is to sort out the user behavior trajectory. User behavior trajectories include cognition, familiarity, trial, usage, and loyalty. Behind the trajectory are reflected user characteristics, which are of great reference value for product operations.

Taking the user's bidding behavior path as an example, we can record the user's behavior trajectory from registration, authentication, account opening, recharge to investment. By analyzing these user behavior trajectory data, we can verify whether the access path is consistent with the expected indicators.

When analyzing user behavior paths, we will find that there is a certain deviation between the user's actual behavior path and the expected behavior path. This deviation is a possible problem with the product, and it is necessary to optimize the product in a timely manner to find room to shorten the path.

5. Fogg model analysis

The Fogg behavior model is an analytical model used to study the causes of user behavior. The Fogg behavior model can be simplified by the formula B=MAT, that is, B=MAT. B stands for behavior, M stands for motivation, A stands for ability, and T stands for trigger. It holds that for a behavior to occur, three elements must be present simultaneously: motivation, ability, and trigger. Therefore, the Fogg behavioral model can be used to evaluate the rationality of the product and whether it can achieve the expected goals.

Taking activity sharing as an example, investors’ behavior of completing activity sharing must also satisfy the three elements of Fogg’s behavior model. That is, by offering invitations with rewards, users will have enough internal motivation to share the event with their friends on their own initiative, and the event special page will have eye-catching buttons and text prompts to encourage users to complete the task.

The user behavior analysis model is actually also an AISAS model, which represents the user's performance from registration, authentication, account opening, recharge to investment: Attention, Interest, Search, Action, Share, which also affects user behavior decisions.

The user behavior analysis model is a complete behavior model that can verify the functions of the product; it is also a closed-loop analysis system that can analyze the results of the data. In short, the core of users is to understand psychology, the essence of behavior is to explore needs, and the purpose of analysis is to grow business.

Author: Zhu Xuemin

Source: Zhu Xuemin