3 models + 5 methods of data operations!

Many friends who are just getting started with data analysis, and even those who have been doing data analysis for several years, often don’t know how to start thinking when they are involved in data analysis in different industries or scenarios. They have no ideas at all, and then they go to Baidu to look for this method and that method. Who knows that sometimes they find analysis methodology and sometimes they find data analysis method, but they cannot accurately define the two clearly, nor do they know how to use these methodologies and analysis methods effectively and correctly.

Let's first look at the definitions between the two:

Data analysis methodology: It is an analysis framework proposed from a macro perspective, from the perspective of management and business. This framework can well guide our next specific analysis direction and sectors, such as: 5W2H, 4P marketing, etc.

Data analysis method: It is a microscopic concept, which refers to the method we use in the specific analysis process, such as common comparative analysis, cross analysis, etc.

The role of data analysis methodology:

(1) Help us sort out our analysis ideas and ensure that data analysis forms a structured system

(2) Break the problem down into related parts and show the relationship between them

(3) Provide guidance for subsequent data analysis

(4) The results of the analysis can be guaranteed to be effective and correct to a certain extent, and it is not easy to deviate from the direction of analysis.

Below we will introduce several common analysis methodologies and analysis methods that are often used in daily analysis.

Data Analysis Methodology

1. PEST analysis model

The PEST analysis model mainly looks at the company's macro-environment from the perspective of the company, analyzing the company's current position from four aspects: politics, economy, society, and technology.

After analyzing the PEST of the company, it is necessary to further analyze the opportunities and threats in combination with SWOT. Therefore, the SWOT analysis model should be used

2. SWOT analysis model

The SWOT analysis model mainly analyzes the strengths and weaknesses, opportunities and threats of an enterprise. It looks at the company's opportunities and risks, strengths and weaknesses from different perspectives. It is relatively comprehensive and covers both the good and the bad, so that a strategic development plan suitable for the company can be better formulated.

3. 5W2H analysis model

When analyzing user behavior, you can look at the problem from the seven perspectives of the 5W2H method, which is very instructive for e-commerce user operation analysis, decision-making and management.

Another common analysis model is the 4P theory, which is mainly used in marketing and examines factors affecting the market from four perspectives: product, price, channel, and promotion. The AARRR analysis model is mainly used to analyze growth conversion issues. It improves user growth from the five link relationships of acquisition, activation, retention, revenue, and referral.

Common data analysis methods

1. Trend analysis

Trend analysis is the most commonly used and common data monitoring and analysis method. It focuses on observing the fluctuations of a certain business and paying attention to abnormal values ​​under the overall trend.

However, in actual business, not all indicator trends are valuable. Trend analysis must also select indicators with core value. For example, when analyzing APP users, if we compare the analysis of users who downloaded the app with the analysis of daily active users, we can find that the daily active user trend is the observation and judgment of the behavior of valuable users, which is of substantial help to decision-making.

2. Multidimensional analysis

When analyzing a problem, one indicator sometimes cannot reveal where the problem lies, so it is necessary to split the indicator. For example, when analyzing the bounce rate of an app, you can analyze and observe from multiple dimensions such as user access depth and access duration to find more possible causes.

3. User Segmentation

User grouping can be based on a single dimension or a combination of dimensions.

Single-dimensional grouping, for example, groups users by age, region, gender, etc., but this type of user grouping is often rough and cannot accurately characterize users, so multi-dimensional combined grouping is required.

Multi-dimensional combined grouping, such as users who log in more than 3 times every night, users who log in less than once during the day, etc. These dimensions are analyzed and verified and then confirmed in combination according to the characteristics of the users.

4. Case Study

Before user grouping, it is often not very clear what user behaviors are, nor is it clear how to start grouping. At this time, it is necessary to sample individual users for detailed link analysis, find most of the user behavior events, and establish the most complete user behavior for user grouping in order to establish better user grouping.

At the same time, case studies can help us discover deficiencies or bugs in product design, and then solve the problems in a targeted and efficient manner.

5. Funnel analysis

Funnel analysis is a model that reflects the conversion of user behavior paths and is widely used in user conversion behavior analysis on e-commerce websites. Maybe everyone is very familiar with it, but there are two key points to note:

(1) The overall transformation and the transformation of each link should be paid attention to and analyzed at the same time

(2) The overall conversion can be split according to the dimension and drilled down to further analyze the anomalies

Common analysis methods include retention analysis, ABtest analysis, cross analysis, etc., which can be selected according to different analysis scenarios to solve corresponding problems.

Summarize

The methodology is the overall framework of analysis, within which detailed analysis can be conducted based on different businesses using different analytical methods.

Combining the common analysis methodologies and analysis methods mentioned today, you can make a summary as follows:

(1) Strategic planning uses PEST and SWOT analysis models, and sometimes personal planning can also be used

(2) 4P Marketing Model for Marketing

(3) E-commerce user analysis using 5W2H and AARRR

Different analysis methods under specific model frameworks are actually aimed at finding better solutions and better decision-making plans for phenomena and problems. Combining past analysis experience, we can summarize the simplest analysis angle:

(1) Start from the macro perspective: trend analysis, dimension analysis, funnel analysis

(2) Start from small users: case analysis, sample ABtext, sample grouping, etc.

That is, when you are unable to solve the problem from a macro or conventional perspective, you can start by exploring the problem links in series from individual cases and details.

Data analysis methodology is the experience accumulated by the past market, which can be used as a reference, but it is not immutable. We will introduce more data analysis methods and data analysis and mining cases in the future. Please stay tuned.