User Behavior Analysis (How to Use Data to Drive Growth) - Reading Notes 2

Chapter 2 Data Planning

1. Three major ideas of indicator system rules

How to use data planning, "OSM model + UJM model + scene"

1.1 OSM (Objective, Strategy, Measure) model

The OSM model will dismantle and measure target strategies so that corporate goals can be structured and clearly presented. Use the OSM model to dismantle specific goals and implement implementation strategies to further evaluate the effectiveness of the strategies and reflect whether the implemented strategies achieve the goals. When selecting goals, you need to pay attention to four principles, namely the DUMB principle of selecting goals: Doable, Understandable, Manageable, and Beneficial.

• O (Objective): Represents the goal, needs to think or answer, what is the goal of our business, product, or even a small function in the product? What problems can users solve? What needs do users need?
• S (Strategy): Represents strategy, which refers to the various strategies we should adopt in order to achieve the goal after clarifying the goal.
• M (Measure): stands for measurement, used to measure whether our strategies are effective and reflect the achievement of goals. Measurement involves two concepts, one is KPI (key performance indicator), which is used to directly measure the effectiveness of the strategy; the other is the target value (Target), which needs to be given a numerical value in advance to judge whether the strategy can achieve the expected value.

1.2 UJM (User Journey Map) model

The UJM model will help you sort out the user journey map and anchor the target objects of data planning. If you want to carry out data planning, the core still needs to be an axis of business processes to connect in series. In the entire business process, based on the user's perspective, the user's participation in the business process is simulated, and the user's entire life cycle is completed, which is called the user journey.

1.3 Scenarioization

Scenarioization helps everyone promote the implementation of the indicator system more quickly. In addition to keeping close to corporate goals, it is also necessary to break down the key scenarios in the user journey map into multi-level goals, corresponding to the corresponding business strategies, as well as data indicators to measure the achievement of the strategic goals.

2. Indicator grading

2.1 First-level indicators

The first-level indicator must be a core indicator that is recognized by the entire company and measures performance. It can be directly used to guide the company's strategic goals and measure the company's business achievement. Essentially, it requires two-way understanding and recognition from management and subordinate employees, and it must be easy to communicate.

2.2 Secondary indicators

Second-level indicators are path analysis and disassembly of first-level indicators, and are indicators in the process. When the primary indicators change, we can quickly locate the root cause of the problem by looking at the secondary indicators and combining it with certain historical experience.

2.3 Third-level indicators

Third-level indicators are path analysis and disassembly of second-level indicators, usually defined in a sub-process or individual manner. Through the third-level indicators, the reasons for the fluctuations of the second-level indicators can be efficiently located. This step will also be dismantled based on historical experience.

3. Data dashboard

3.1 The significance of data dashboard

Data Kanban is a visual tool mainly used for business communication to detect core business status. The meaning of Data Kanban can be summarized into the following three parts:

(1) The data dashboard is a visualization tool. Through data visualization, companies can integrate data information, monitor business processes, and measure and share business results.

(2) The data dashboard is a communication tool. Through data disclosure and presentation, effective information can be shared within the company and communication and collaboration between organizations can be activated.

(3) Data dashboard is a strategic tool for the company to implement data-driven growth. Generally, data dashboards have three major application scenarios: monitoring, analysis and collaboration.

• Monitoring scenario: Monitoring is the mainstream application scenario of data dashboard. Through the large-screen data display, companies can obtain data in real time, understand business processes, gain insight into development trends, and even issue business warning information.
• Analysis scenario: The data dashboard needs to have the ability to drill down to details. When the actual data is inconsistent with project expectations, it can help the business department analyze the details that lead to abnormal situations and get to the core problem. In short, we use data dashboards to dismantle data in multiple dimensions, conduct business comparisons, and analyze the causes of problems.
• Collaboration scenario: After discovering data problems and determining the cause of the problem, the company needs to take action to solve the problem. Solving business problems using data dashboards often requires teamwork.

3.2 Data dashboard classification: strategic dashboard, analysis dashboard, real-time dashboard

(1) Strategic Kanban

The strategic dashboard is the core data dashboard that senior managers or decision-makers of an enterprise pay attention to. It is often an overall overview of indicators. The strategic dashboard should follow the principle of data integration, that is, in the visual information, the descriptions related to the data should be diversified, and other information should be sparse to help managers quickly discover trends and problems.

(2) Analysis signboard

Analytical Kanban is an analysis tool that uses data visualization functions to find the reasons for changes in business processes. The Analytical Kanban serves as an exploratory verification. There is a set of logic in building an analytical dashboard. The company or department first proposes a hypothesis, and then demonstrates the hypothesis through the data dashboard to determine whether the change has the potential to become an opportunity. Then iterates the product or operation strategy, and finally evaluates the effect.

(3) Real-time billboard

The real-time dashboard is mainly used to monitor specific operational scenarios. By determining the operational scenarios and goals, find indicators to measure the goals, and then monitor user behavior through the data dashboard to achieve rapid decision-making. Real-time dashboards must break down very fine-grained data to produce data-driven effects.

3.3 How to build a data dashboard

(1) The data dashboard has the following two main features:

• A data dashboard focuses on a business goal. The indicators and charts in the data dashboard are designed to better describe business issues, help users understand business processes, gain insights into business trends, and quickly identify problems.
• The screen of the data dashboard must show the whole picture. If the screen of the data dashboard can only display one-sided or partial data, the efficiency of users in obtaining information and gaining insight into problems will be greatly reduced. Only when the data dashboard presents a complete picture of the business information, can we successfully find the new curve of corporate growth through the patterns of data insight indicators.

(2) When building a data dashboard, you must follow the following three principles:

• Principle 1: One screen contains all the information you need. Only by integrating the required information on one screen can users quickly obtain full business facts and understand business issues. Once the data is scattered across multiple screens, it will be necessary to scroll or switch pages to obtain fragmented data information, which will seriously affect the user experience and analysis efficiency.
• Principle 2: It needs to be timely. The timeliness of the data dashboard is determined by the company's business goals and business cycle. For example, when an e-commerce company organizes a major promotion event, it is necessary to go to the data dashboard to display minute-level or even real-time level data. As for enterprise service companies, it means that they need to meet the timeliness of hours or days.
• Principle Three: Customization. Enterprises must build data dashboards to meet the needs of the company, departments, and teams. The data dashboard is a carrier of information. Only when the data meets the user's business requirements, will it receive the user's attention.

(3) Data dashboard requires the following four major components:

• Visualization. Visualization is the presentation of data information through charts.
• Able to tell stories and focus on goals. The criterion for judging whether a data dashboard is qualified is whether the data dashboard can tell a good story and focus on a goal.
• Can quickly identify problems. Dashboards should help users track progress toward goals. If the business goals are not achieved, the data dashboard needs to provide information such as outliers and problem points that hinder the completion of the goals.
• Able to analyze and act. When we find a problem, the data dashboard needs to provide data and ideas to analyze the problem and assist the team in taking action. When there is a problem with a core indicator, corporate executives or the person in charge of the indicator can drill down according to the business dimension to find the department responsible for the core indicator, and then drill down through the department manager to find the specific person in charge for implementation.

(4) The process of an enterprise building a data dashboard is also a process of organizing data information and realizing business communication. It can be divided into the following four steps:

• The first step is to clarify the needs. There are three questions to ask when building a data dashboard: first, what are the users’ business needs, second, what are the business goals, and third, how to achieve the business goals. After getting the answer, you can clearly understand the need to build a data dashboard and focus on specific business issues.
• The second step is demand analysis. In this step, it is necessary to disassemble the business demand objectives, choose the appropriate latitude to abstract them into a data indicator system, and determine the basic content of the data dashboard.
• The third step is visualization. Visualization is a core part of the process of creating a data dashboard. Visual charts need to accurately express data information and clearly convey business facts through orderly combinations and arrangements.
• The fourth step is to evaluate the effect. After completing the basic construction work, we need to pay attention to whether the data dashboard has only one screen, whether users can tell a complete business story through the data dashboard, and whether the created data dashboard can help users quickly discover trends, patterns, and anomalies.

3.4 Six common problems in building data dashboards

• Problem one is data fragmentation. When enterprises build data dashboards, a common problem is that one business information is scattered across multiple data dashboards. On the one hand, scattered data dashboards will greatly affect the convenience of data use; on the other hand, only complete data can present a complete picture of the business. Therefore, when building a data dashboard, it is necessary to put the core data domain charts in a unified data dashboard to prevent data fragmentation.
• Problem two: too much data and too little information value. When using data dashboards, companies will encounter situations where there is a lot of data but no pattern can be discerned. A lot of information, but very little value. The correct approach is to put only core information into the data dashboard. Only in this way can users quickly discover business problems and reduce information interference.
• Problem three, the data is too slow. Under normal circumstances, activity operation data should be updated in real time. If the data update speed is too slow, emergency problems cannot be solved, and in severe cases, huge losses may be caused. Therefore, in specific activities, companies should consider building real-time data dashboards.
• Problem four, wrong layout. When many companies lay out data dashboards, they simply fill in charts and ignore the user experience. Unordered and fragmented chart layouts make it more difficult to understand. As a communication tool, charts must be arranged according to a certain logic and convey information concisely and clearly. In this way, when using the data dashboard, you can use it accurately and well.
• Problem 5: Wrong way of visualizing data. The role of charts is to visually present data and help data dashboard users understand data trends. However, when building data dashboards, people often use charts incorrectly or use charts that have no relevance to the business problem. To avoid visualization errors, the goals of data visualization need to be clear. Only by clarifying what the goal is can you find the chart type that suits it. Here are five commonly used combinations.

trend:	line graph. The line graph is suitable for observing the continuous change trend of one or more data indicators, and can also be used for comparative analysis with data of different time periods.
distributed:	People are less sensitive to area and more sensitive to length, so the effect of using a histogram will be better.
Sort by:	Horizontal histogram (bar chart). Horizontal column charts are suitable for distribution and sorting along a certain dimension.
Multi-dimensional and detailed:	Tables are the most intensive way to present information and can analyze multi-indicator and multi-dimensional data at the same time, as well as detailed data.
number:	Large numbers picture. Dashu graphs are mainly used to monitor KPIs.

• Problem 6: The data measurement method is wrong. Different types of data often correspond to different types of data visualization. Whether it is trend, comparison or sorting, they are all corresponding chart display methods. When we use the wrong graphs to depict data, we lose the effectiveness of conveying information. Therefore, when using charts to depict data, you must consider the fit between the chart type and the data information.