You can create an RFM model from the Order Summary Table and Order Details Table that are imported or stored after a report. You can use the RFM model to analyze and filter people.
What is an RFM model?
The RFM model is a means to measure user value by using three indicators: R consumption interval (Recency), F consumption frequency (Frequency), and M consumption amount (Monetary).
The RFM model quantitatively scores the values of the three metrics of users. The scoring criteria are as follows:
R: Based on the number of days since the last consumption, 0~30:3, 31~90:2, and above 90: 1.
F: According to the number of consumption, 1:1, 2~4:2, 5 above: 3.
M: According to the consumption amount, 0~100:1, 100~1000:2, above 1000: 3.
Then, by comparing the score of a single user with the comparison value (the average score of the entire user group, or the specified score), the relative value level of the user in the group is obtained, and then the user group is divided into 8 types by integrating the three indicators, so as to facilitate targeted operations for different types of users.
RFM User Types and Classification Rules
Comparing the user's RS, FS, and MS scores with the RS comparison value, FS comparison value, and MS comparison value, respectively, the user's relative value level in the group can be obtained:
The user score is greater than the comparison value, and the value is higher.
The user score is less than the comparison value and the value is lower.
RS, FS, and MS are the consumption interval, consumption frequency, and consumption amount scores of users respectively.
The RS comparison value, FS comparison value, and MS comparison value are the average values of the consumption interval, consumption frequency, and consumption amount scores of all users in the RFM model (that is, the weighted average in statistics), or custom values.
For more information about how to set scoring rules and comparison values in an RFM model, see Create An RFM Model.
The value of users in any one of R, F and M can be divided into high and low categories. Based on the performance of R, F and M, users can be divided into 8 types. The detailed types and classification rules are shown in the following figure.
RFM customer type | RS | FS | MS | Description |
High-value users | Greater than or equal to RS comparison value | Greater than or equal to FS comparison value | Greater than or equal to MS contrast value | Define users with recent consumption dates, high consumption frequency, and high consumption amount as high-value users. |
Emphatically keeping users | Low | Greater than or equal to FS comparison value | Greater than or equal to MS contrast value | Define users whose most recent consumption date is far away but whose consumption frequency and consumption amount are high as key retention users. |
Emphatically developing users | Greater than or equal to RS comparison value | Low | Greater than or equal to MS contrast value | A user with a recent consumption date and a high consumption amount but a low consumption frequency is defined as a key development user. |
Emphatically retaining users | Low | Low | Greater than or equal to MS contrast value | Users whose recent consumption date is far away and whose consumption frequency is low but whose consumption amount is high are defined as key retention users. |
General Value Users | Greater than or equal to RS comparison value | Greater than or equal to FS comparison value | Low | A user with a recent consumption date and a high consumption frequency but a low consumption amount is defined as a general value user. |
Generally keeping users | Low | Greater than or equal to FS comparison value | Low | A user whose recent consumption date is far away and whose consumption amount is not high but whose consumption frequency is high is defined as a general user. |
Generally developing users | Greater than or equal to RS comparison value | Low | Low | A user with a recent consumption date but low consumption frequency and low consumption amount is defined as a general development user. |
Potential users | Low | Low | Low | Users whose recent consumption date is far away, whose consumption frequency is not high, and whose consumption amount is not high are defined as potential users. |
Create an RFM model
You can select two analysis types when creating an RFM model:
Order summary data: aggregates the raw data of the last N days into a large wide table with user granularity. Each user in the table has only one piece of data.
For more information about the sample Quick Audience Import Data Table Requirements, For more information about how to import order summary table data, see Order Summary.
Order detail data: uses user transaction data as the basis for analysis. Each row in the table represents a user transaction record. When an RFM model is generated, the system aggregates order details for each user.
For more information about sample order details, see Quick Audience Import Data Table Requirements. For more information about how to import order details, see Order Details. For more information about how to report and store order event data, see Event Hub.
Procedure
Choose Workspace> User Insight > Marketing Model > RFM Model.
Click New in the upper-right corner to go to the RFM model configuration page.
Select a configuration method, as shown in the following figure.
Select Order Summary or Order Details as the analysis type.
Select the imported data table to analyze (use the data table alias set during import).
The unit of the amount. You can select CNY, USD, GBP, EUR, or HKD.
Click Next.
Configure RFM parameters
The following figure shows the RFM analysis of order summary data.
The following figure shows the RFM analysis of order details.
For Order Details only, you can specify the statistical period, order behavior type, and order time fields.
If the Purchase Channel and Commodity Category Name fields are configured when the order details table is imported, you can also specify the statistical channel and commodity category to filter the corresponding data for targeted analysis. You can select up to 100 values for each of the statistical channels and product categories.
For all types: Select the number of intervals (3 or 5) of R, F, and M respectively, and set the range of the interval corresponding to each score. The distribution of the number of people in the corresponding interval is displayed below.
Scoring rules follow:
R consumption interval: The fewer days since the last purchase, the higher the score.
F consumption frequency: the more the number of consumption times in the last n days, the higher the score.
M Consumption Amount: The higher the consumption amount in the last n days, the higher the score.
After you configure the R, F, and M scoring rules, click Next.
Configure the parameter comparison value, as shown in the following figure.
The comparison value is used to compare the score of a single user with the comparison value in subsequent analysis to obtain the relative value level of the user in the group, and to classify different user types. For more information, see RFM User Types and Classification Rules.
RS, FS and MS are the scores of R, F and M respectively. Accordingly, the comparison values of RS, FS and MS need to be set. You can use the average score of the entire user population as the comparison value, or you can customize the comparison value.
If the average score of the entire user group (that is, the weighted average in statistics) is used as the comparison value, the page will display the average score of RS, FS, and MS of the current user group according to the set scoring rules.
If you customize the comparison value, adjust the comparison value based on your business needs and refer to the average score of the entire user group displayed on the page.
Click Finish. In the dialog box that appears, enter an RFM model name and click Save.
The RFM model list appears. You can view the newly created RFM model in the list. For more information about how to manage RFM models, see Manage RFM Model.
Managing RFM Models
RFM models support editing, RFM analysis, updating, renaming, deleting, permission setting, and updating settings.
Edit: Click Edit to edit the model.
Analyze: Click Analyze to analyze the RFM model. For more information, see RFM Analysis.
Update: Click Update to immediately update the model.
Rename: Choose
/> Rename to redefine the model name. Delete: Select
/> Delete and confirm to delete the model. NoteIf the model generates a crowd, you cannot delete it.
Permission setting: Grant permissions to use or manage RFM models to non-administrators. For more information, see Authorize Tags.
NoteIf you are a non-administrator, the RFM model is displayed on the Authorized tab of the RFM Models page.
By default, the administrator has the permissions to manage all RFM models in the workspace. The administrator is displayed on the My tab of the RFM Models page.
Update settings: Select
/> Update Settings. In the Update Settings dialog box, you can turn on the Auto Update switch and set the update time. Within the specified date range, the model will be automatically updated each time the order details table or order summary table that is used completes the scheduling task. 
NoteIf the update parallelism in the space has been set in Central Administration, RFM models that exceed the limit need to be queued for automatic updates. For more information, see Workspace System Configuration.