Dry Goods Sharing

Data operation ability has been recognized as a plus item or even a necessary skill for operators, and its level also determines the salary level and career life cycle of operators to a great extent. On the one hand, it can greatly improve the working efficiency of operators; On the other hand, it can analyze the operation more comprehensively and deeply, so as to better assist the strategy and guide the practice. In order to make operators operate data better, various data analysis models came into being. Through the data analysis model, we can not only reduce the cognitive cost of operators, but also help us simplify complex problems, quickly understand objective things and easily get started with data analysis. 0 1 Introduce the "Operational Essence 1 1 Big Data Analysis Model" put forward by Guan Yi Ark, and summarize the event analysis, attribute analysis, channel analysis, conversation analysis, retention analysis, attribution analysis, heat map analysis, distribution analysis, funnel analysis, interval analysis and path analysis. This 1 1 big data analysis model has different functions in different operating scenarios and has been widely used in data analysis. Event analysis events refer to users' behaviors in applications such as apps and websites, that is, who, when, where and by what means did something. The event analysis model is mainly used to analyze users' behaviors in applications, such as opening an app, registering, logging in, and paying orders. User behavior can be measured by basic indicators such as the number of triggered users, the number of triggered users, and the length of access. At the same time, it can also be used to calculate indicators and build complex business processes for measuring indicators. Specifically, the event analysis model can solve the following sample problems: monitoring whether the number of users, visits, usage duration and trends have changed every day? What are the factors that cause the change? What's the difference in the amount distribution of household appliances purchased by users in Beijing and Shanghai? Today, a topic was initiated in the product. What is the participation of users in each period? What is the number of paying users and ARPU (average income of each paying user) in the past six months? The event analysis model of Guan Yi Ark intelligent analysis products (see figure 1- 1) can be used to analyze and monitor users' behaviors on different platforms in real time. When the event analysis model is used in session analysis, user events are presented in the form of "points". For example, Zhang San registered as a member of a takeaway platform at 10 last night and paid the first order. Li Si reported the vehicle trouble after sweeping a bicycle near Wangfujing at 8 o'clock this morning. According to the real-time recorded feedback of users' behaviors, we can know exactly when users do what. But in fact, some events can't be described by these "points", for example, the average number of visits by users this month, the length of each visit and the average depth of visit. These problems need to connect "points" into "lines" and then analyze and calculate them. Conversation analysis can perfectly solve the "linearity" problem in user analysis. A Session is a session, which refers to a series of user behaviors that occur on the website /H5/applet/app within a specified period of time. For example, a session can contain multiple page views and interactive events. Session has time attribute, and according to different cutting rules, sessions with different durations can be generated. Specifically, the following series of example behaviors can be regarded as a conversation: IOS application: the user's screen is closed, the Home button is switched to the background, and the process is killed; conversation: iOS overAndroid application: the user kills the process, closes the screen, presses the Home button for more than 30 seconds, and crosses the sky. When the session ends; H5/Web application: The user leaves to open a webpage, which is regarded as a session. Leaving includes closing the whole browser, not opening a new page for 30 minutes or triggering an event. If a visit takes several days, it will be divided into two parts. The session analysis model of Guan Yi Ark Intelligent Analysis (see Figure 4- 1) can analyze various indicators to measure the quality of session access according to different time granularity, including visit times, per capita visit times, total visit duration, single visit duration, single visit depth, jump times, bounce rate, exit times, exit rate, per capita visit duration, total page stay duration and average page stay duration. In addition, it can be multi-index, multi-dimensional and multi-filter conditions, and it can also be compared horizontally among multi-user groups. Compared with event analysis, conversation analysis adds some extra dimensions to meet the needs of conversation analysis in specific scenarios, including: channel source grouping: used to distinguish the channel sources of each visit, only applicable to Web/H5/ applets; Pages browsed: count the distribution of pages browsed every time at the interval of step 5; Landing pages: used to distinguish the landing pages visited each time and evaluate the access quality of different landing pages; Exit pages: used to distinguish the exit pages visited each time, and can evaluate the exit situation of different pages and find the pages with high exit rate for optimization; Duration of visit: 0-3 seconds, 3- 10 seconds, 10-30 seconds, 30-60 seconds, 1-3 minutes, 3- 10 minutes, 30 minutes. 04 Retention Analysis Retention means that users have used applets, apps, websites and other applications and are still using them after a period of time. Retention analysis is a method to measure users' health or participation. Based on the initial behavior time of a user group, it describes whether the expected behavior occurs in the same user group after a period of time. Retention analysis can help us deeply understand the retention and loss of users, find the key factors that affect the sustainable growth of products, guide market decision-making and product improvement, and enhance user value. Specifically, retention analysis can solve the following examples: last month, a product iteration was done, how to evaluate its effect? Have you completed the behavior expected by the product manager? As a social APP, is there a difference between users who don't add friends and users who add 10 friends after registration? Short-term retention is low, long-term retention must be poor? The users brought by the two promotion channels are different. Which channel users are more likely to be high-value users? What is the proportion of users who have registered in the past 30 days and have not paid a return visit for half a month? The retention analysis model of Guan Yi Ark's intelligent analysis products (see Figure 5- 1) can define the initial behavior and subsequent behavior, select the number of retained users/retention indicators to check the retention situation, and can filter the conditions of different dimensions for multi-population comparative analysis. In addition, retention analysis can also be used to determine whether new users are willing to come back and continue to use one of your products or functions in a few days, weeks or months. For more detailed model introduction and usage, the official WeChat account replies "09 12 operation analysis" in the background for information.