Three magical elements in the customer database

Absrtact: According to Arthur Hughes of the American Database Marketing Institute, there are three magical elements in the customer database that make up the best indicators of data analysis: The most recent consumption (recency) consumption frequency (Frequency) consumption amount (mone

According to Arthur Hughes of the American Database Marketing Institute, there are three magical elements in the customer database that make up the best indicators of data analysis:

Last Consumption (recency)

Consumption frequency (Frequency)

Consumption Amount (monetary)

RFM analysis of the original used in traditional marketing, retail industry and other fields, applicable to a variety of consumer goods or FMCG industry, the United States to the first China copper tube levy high to 60% anti-dumping duties as long as any data recorded consumption can be used for analysis. So for E-commerce sites, the site database records of detailed transaction information, can also use the RFM analysis model for data analysis, especially for those who have established customer relationship management (CRM) systems, the results of the analysis will be more meaningful.

Basic concept Explanation

RFM model is an important tool and means to measure customer's value and profitability. The RFM analysis model is mainly composed of three indicators, the following is a brief explanation of the definitions and functions of these three indicators:

Last Consumption (recency)

The most recent consumption means the last time a user bought, in theory, the closer the last time consumers should be a better customer, to provide real-time goods or services are most likely to respond. Because the most recent consumption metric defines a time period and is related to the current time, it is always changing. The most recent consumption is an important indicator of marketing, involving attracting customers, keeping customers, and winning customer loyalty.

Consumption frequency (Frequency)

Consumption frequency is the number of customers in a certain period of time consumption. The most frequently purchased consumers have the highest degree of loyalty, and increasing the number of customer purchases means stealing market share from competitors and earning turnover in the hands of others.

According to this indicator, we divide the customer into five equal parts, the five-division analysis is equivalent to a "loyalty ladder" (loyalty ladder), the trick is to let consumers continue to climb up the ladder, the short slump in crude oil does not hinder the oil price increase expectations The idea of sales is to push two customers into three purchases and turn the buyer into two times.

Consumption Amount (monetary)

The amount of consumption is the most direct measure of the productivity of e-commerce sites, and it can also be verified as "Paretofa" (Pareto ' s)--80% of the company's revenue comes from 20% of customers.

Data acquisition and analysis

Before extracting the relevant data from the database, it is necessary to determine the time span of the data and determine the appropriate time span according to the difference of the items sold by the website. If you are running fast-moving consumer goods, such as daily necessities, you can make sure that the time span is one quarter or one months, and that if you sell a product for a relatively long time, such as an electronic product, you can determine a time span of one year, six months, or a quarter. After the time span is determined, the data in the corresponding time interval can be extracted, where:

The most recent consumption (recency), the data taken out is a point in time, need to be from the current point of time-the most recent consumption point as the value of the measure, attention to unit selection and unification, whether in hours, days as a unit;

Consumption frequency (Frequency), this index can be directly in the database count users to get the number of times;

Consumption amount (monetary), which can be calculated by adding (SUM) The amount of each customer's consumption.

After obtaining the data of three metrics, the mean value of each metric data needs to be calculated, which is represented by AVG (R), avg (F), avg (M), and the customer can be subdivided into 8 categories by comparing three metrics per customer with the mean value:

Recency Frequency Monetary Customer type

↑↑↑ Important Value Customers

↑↓↑ Important Development Customers

↓↑↑ important to keep customers

↓↓↑ important to retain customers

↑↑↓ General Value Customers

↑↓↓ General Development Customers

↓↑↓ generally keep customers

↓↓↓ generally retain customers

--"↑" expression is greater than the mean, "↓" means less than the mean value

Presentation of the results

The RFM model consists of three indices that cannot be displayed with planar coordinates, so this is shown using a three-dimensional coordinate system, an x-axis representation of the recency,y axis representing the frequency,z axis monetary, and the 8 quadrants of the coordinate system representing 8 categories of users, according to the categories in the table above, You can use the following graphic to describe:

RFM analysis also has a certain flaw, it can only analyze the transaction behavior of the user, the head of the community and the user visited the site but not consumption due to the limitations of the indicators can not be analyzed, so that can not find potential customers. So in the analysis of E-commerce site users, due to the richness of the site data-not only have transaction data, but also can be collected to the user's browsing access data, can be extended to a broader perspective to observe users, this quantitative analysis will be in the Web site after detailed description of users.