Reprinted Source 1:
https://blog.csdn.net/qq_41455420/article/details/79200553
Reprinted Source 2:
https://blog.csdn.net/weixin_44530236/article/details/89956447
table of Contents
1. Data Analysis Data Analysis Methodology difference
2. The importance of data analysis methodology
3. Common data analysis methodology
Ideas need to be analyzed to determine the marketing and management theory as a guide. These analyzes with data related to the marketing and management theories collectively referred to as data analysis methodology. The methodology can be understood as a compass, under the guidance of the analysis methodology we going to carry out the data analysis, the results of such analysis was instructive, but the situation does not appear poles apart.
1. Data Analysis Data Analysis Methodology difference
Data analysis methodology is mainly used to guide data analysts to conduct a complete analysis of the data, it refers more to data analysis ideas, such as the main areas where to carry out data analysis, all aspects of what is included and indicators? Data analysis methodology mainly from a macro perspective guidance on how to analyze the data, it looks like a pre-planned analysis of data to guide the conduct post-data analysis work. The data analysis rule refers to a specific method of analysis, such as our common comparative analysis, cross analysis, correlation analysis, regression analysis, cluster analysis, data analysis. Data analysis mainly from the microscopic point of guidance
on how data analysis.
For example, the following table we have to make clothing as an example, comparative understanding of data analysis methodology:
2. The importance of data analysis methodology
Many people do data analysis, often encountered several problems: I do not know which side to start to carry out analysis;
content analysis and indicators are often questioned whether reasonable and complete, while he could not say what they were. On
these issues often feel troubled.
Data analysis methodology has the following main functions:
Straighten out the analysis of ideas, to ensure system data structure analysis.
The problem into an associated portion, and shows the relationship between them.
Guidelines for carrying out the direction of the subsequent data analysis.
The results ensure the effectiveness of Health and correctness.
If there is no guide data analysis methodology, although the entire data analysis reports covering all aspects to,
but give people the feeling of something missing. In fact, the main line of the report is not clear, logical analysis of each part is unclear.
3. Common data analysis methodology
3.1 PEST analysis
PEST analysis method for the analysis of the macroeconomic environment. Macro environment, also known as the general environment, it refers to all industries and the impact of various macroeconomic forces enterprises. When the macro environment factors for analysis. Due to the different industries and enterprises have their own characteristics and business needs, analysis of the specific content will vary, but generally deal with politics (Political), economic (Economic), technology (Technological) and social (Social) these four categories affect analyze the main external environmental factors, this method is referred to as PEST analysis.
- Political environment: political environment constitute the key indicators are: the political system, economic system, fiscal policy, tax policy, industrial policy, investment policy, the number of patents, the level of defense spending, the level of government subsidies, public participation in politics and so on.
- Social environment: key indicators constitute a social and cultural environment are: population size, sex ratio and age structure, fertility, mortality, ethnic structure, fertility rate, life-style force, buying habits, educational status, urban characteristics, religion status factor.
- Technical environment: key indicators constitute a technical environment are: the invention of new technology and the progress of wood, depreciation and obsolescence rate, technological upgrading, technical propagation speed, the speed of technology commercialization, support of key national projects, national R & D spending, patent number the number of patent protection and other factors.
- Economic environment: key indicators constitutes the economic environment are: GDP and growth rate, and total import and export growth rates, interest rates, exchange rates, inflation, consumer price index, household disposable income, unemployment rate, labor productivity and so on.
3.2 5W2H analysis
5w2H analysis method is based on five English words beginning with w and two English words beginning with H ask questions, find clues to solve the problems of the answer, that is what for (Why)), what is going on (What), who (Who), when (when), where (where), how to do (how), what price (how much), which constitutes a general framework 5W2H analysis.
The method is simple, convenient, easy to understand and use, enlightening, widely used in corporate marketing, event management, decision-making and implementation of measures for the activities of very helpful, also help make up for the omission to consider the issue. In fact, seven are from this aspect to think about anything, topic sentence analysis for poor people, as long as the practice more to get started, so it is the same also applies to guide the establishment of a data analysis framework.
Now the user purchase behavior analysis, for example, to learn 5W2H analysis.
3.3 Logical Tree Analysis
Logic tree, also known as the problem tree, the tree or the interpretation of decomposition trees. It is all a question of the problem of hierarchical lists ,, start from the top and gradually downward expansion. The problem is known as a tree trunk, and then start thinking about it and what related issues. Every thought that gave the trunk of this problem plus a "branch" and marked the "branches" on behalf of any problems. Logical tree can solve the problem of ensuring the integrity of the process, it will work broken down into tasks easy to operate, prioritize each part, clearly the responsibilities to individuals.
Use logic tree must follow the following three principles.
- Features': the same problems as summarized elements.
- Framework of: the various elements into the framework of the organization. Does not comply with the principle of weight does not leak.
- Association of: within the framework of the elements necessary to maintain the relationship, not simply independent.
- When using logic tree, try to involve issues or feature thoughtful.
3.4 4P marketing theory
4P marketing theory produced in the United States in the 1960s, and it is with the proposed marketing mix theory of the emergence. There are actually dozens of marketing mix elements, these elements can be summarized into four categories: Product (Product), price (Price), channel (Place), promotion (Promotion).
- Product: From a marketing point of view, the product is the ability to provide to the market, used by the consumer and meet them and whatever people need some kind of tangible goods, services, people, organizations, ideas, or a combination thereof.
- Price: is the price at which customers buy products, including the basic price, discounted price, payment terms and so on. The main factors affecting the price of three: demand, costs and competition.
- Channel: refers to all aspects of the product from the manufacturer to the user's hands full flow in the process of going through.
- Promotion: is an enterprise user to stimulate consumption by changing sales practices, short-term behavior (such as profit sharing, buy one get one, the marketing atmosphere, etc.) contributed to the growth of consumption, to attract users of other brands to promote or lead to premature consumption sales growth. Advertising, promotion, personal selling, sales promotion agencies are the four elements of a promotional mix.
Next, we have the overall business operations of the Company for analysis by 4P marketing theory. In practice, we need to be adjusted based on actual business situation, flexible transport month, should not be applied mechanically. Only a deep understanding of the company's business at the same time in order to better data analysis business that would otherwise be out of business reality, no draw instructive conclusions, like on paper.
3.5 User behavior theory
Development of web analytics has been more mature, have a mature analysis indicators. Such as IP, PV, time on page, bounce rate, returning visitors, new visitors, repeat visits, separated by a few days visit, turnover, keyword search, conversion rate, login to rate, and so on. Met so many indicators, all indicators should adopt it? What is the use of indicators? What indicators do not adopt this, What is the connection between index, which index to analyze? Which index after analysis?
So many indicators still ignorant for us to start with, so we need to sort out the logical relationships between them, such as user behavior using the theory to sort out.
User behavior refers to the user to obtain, use of goods or services taken various actions, users first need to have a product awareness, familiar with the process, then the trial before deciding whether to continue to use consumption, and eventually become loyal users.
Now we can take advantage of user behavior theory, sort out the logical relationships between the key indicators of the site analysis, building in line with the company's actual business website analysis indicator system:
3.6 RFM model
RFM model is an important tool and a means of measuring the current customer value and potential profitability capabilities, many customer relationship management (CRM) analysis mode, RFM model is widely used; this model represents a customer purchases by R (Recency) time how far, F (frequency) represents the number of customer purchases within a fixed time, M (Monetary) represents the amount that the customer purchases within a fixed time.
General Analytical CRM focuses on the analysis of the customer's contribution, RFM is emphasis on customer buying behavior to distinguish between customers. In practice there will be adjusted according to the dimensions of different data services by reference.
- R-value: the last consumer (Recency) refers to the most recent spending time with the current time interval. R-value is theoretically the smaller customers are more valuable customers. In the current online shopping convenient environment, customers have more choice and lower purchase cost of the purchase, the removal of geographical constraints, the customer is very easy to lose, if you want to increase the repo rate and the retention rate, you need to always pay attention to R value.
- F value: Consumption Frequency (Frequency) refers to the number of customers buy at a fixed time (for example, one year). In practice, due to the special nature of some commodities, such as 3C products, durable goods, etc. Even a big fan of users is difficult to buy several times in one year. Therefore, in this case, the range of time values F will remove, replace cumulative purchases.
- M value: the amount of consumption (Monetary) refers to the customer the amount of consumption in the corresponding time (such as one year). M and F values are the same, all with time, it refers to the amount of consumption over time.
Practical Application
1, customer segmentation
- High-value customers: Recent spending time near, consumption frequency and amount of consumption are high, belong to high-quality customers.
- Focus on the development of customers: Recent spending time close to the high amount of consumption, but not high frequency, customer loyalty is not high, have a great potential to focus on the development.
- Focus on customer retention: Recent spending time far away, but the amount and frequency of consumption are high, indicating that this is a period of time not to loyal customers, we need to take the initiative and keep in touch with him.
- Focus on customer retention: spending time recently far, frequency of consumption is not high, but the high amount of consumption of the user, may be lost or will have been lost customers, retention measures should be taken.
2, marketing strategy
The following are examples
3, model scoring
In addition to group users directly with RFM model, there is a common method is the use of three properties RFM model to customers to rate the quality of each user is determined by scoring the final screening of their own target audience.
RFM model scoring has three main components:
- Determining indicators RFM three segments and each segment score;
- RFM score is calculated for each of the three indicators of customers;
- The total score is calculated for each customer, and quality customer selected according to the total score.
for example
After determining the segment and the corresponding segment RFM scores, may be scored in accordance with the corresponding user's situation.
How to determine the corresponding segment of the score, there is no more unified, scientific method, it can be scored based on experience, and then use the model to verify the correction algorithm.
3.7 Funnel Analysis Model
It found that in many things in the process of moving forward, will be an inverted pyramid shape. Starting at the same starting line of a marathon runner, with the increase of Dan, fewer and fewer people can continue to stay on the track, and ultimately only a handful of people can end up by the whole point. This phase-out or loss mode can be abstracted into a funnel model.
Funnel Analytics scene
Electricity supplier industry conversions of different customer groups: a scene
A business enterprise customers based on customer spending power, customers are divided into ordinary members, Gold, Diamond members. To enhance the conversion of the user guide, F operating mode to be different for different user groups. 
1 Member and Diamond funnel conversions contrast (Source: Shence Data Products)
By contrast, the apparent conversion of ordinary members "pay order" from the "Submit Order" to significantly lower than Diamond members. To find out "payment order" stage conversion rate becomes low, F company operations staff should depth analysis Regular conversion circumstances, such as comparison of different payment channels (PC side, the phone side, etc.) of the conversions, find short board optimization. In addition, try to pay novice order process guidance, help the novice successful completion of the purchase.
Scene 2: Retail - Commercial Huimin effect within the scientific assessment station promotion bits
Home Promotion bit performance monitoring is an important part of the station operating, monitoring and analysis of data is an important task, it is the station optimization, page guide to enhance the experience. Operations staff can click conversion rate of purchase conversion rate of users can determine the position of a page to promote a different effect. The figure is in the business that benefit the home position to promote the "one dollar promotion", "clean special" two Banner conversion contrast to the situation. (Note: for alleged trade secrets, the following scenarios simulate real-world application scenarios and set, the data are fictitious.)
2 "one dollar promotion", "clean special" contrast to the situation two Banner conversion rate (Source: Shence Data Products)
In addition, funnel analysis model has been widely used in data analysis in various industries to assess the overall conversion rate, conversion rate all aspects of the scientific assessment of the effect of promotions and other special events, and other data through a combination of analytical models were depth user behavior analysis to find the cause of the loss of customers, in order to enhance the amount of users, active, retention, and improve the scientific data analysis and decision-making and so on.
3.8 in freight yard model
3.9 AARRR model
