How to segment a database manually?
Posted: Thu Dec 26, 2024 10:36 am
Data collection
To conduct RFM analysis, you will need data on all purchases made by all guests and the amounts of all these purchases. You can collect them manually, for example, using pivot tables in Excel or Google Sheets. However, this process is significantly simplified if you automate it using special services.
Guest database in CRM system
Guest Segmentation Criteria
You divide guests into segments based on the time czech republic mobile phone number list since the last purchase (Recency), the frequency of purchases (Frequency), and the amount of money spent (Monetary). You give each guest three ratings corresponding to each of these parameters. For example, on a three-point system (where 1 is good, 2 is normal, and 3 is bad).
The main difficulty is to define the boundaries of segments, because there are no universal recommendations on this matter. Focus on the niche, the guest's life cycle, and other factors.
The range for each indicator will be different for different companies. Let's say you are an expensive restaurant in the center of Moscow. In this case, you give a "good" rating if the customer comes back to you in 2 months. If you are a food delivery service, you can give a "bad" rating for the lack of repeat orders within 2 weeks. The amounts will also vary greatly. For an expensive restaurant, 3,000 rubles is a low bill, for food delivery it is high.
Formation of segments
So, you should have 27 segments (3*3*3) with scores like 111, 112, 113, 121, 131 and so on up to 333.
The segments may be uneven (one segment may include 90% of all guests, another - 1%). Therefore, it is advisable to split too wide groups into several additional ones, and combine narrow ones. You can also allocate more levels. However, this will complicate further work with segments, since there will be even more of them. If there are 4 levels, you will get 64 segments, if 5 - already 125, and so on.
To conduct RFM analysis, you will need data on all purchases made by all guests and the amounts of all these purchases. You can collect them manually, for example, using pivot tables in Excel or Google Sheets. However, this process is significantly simplified if you automate it using special services.
Guest database in CRM system
Guest Segmentation Criteria
You divide guests into segments based on the time czech republic mobile phone number list since the last purchase (Recency), the frequency of purchases (Frequency), and the amount of money spent (Monetary). You give each guest three ratings corresponding to each of these parameters. For example, on a three-point system (where 1 is good, 2 is normal, and 3 is bad).
The main difficulty is to define the boundaries of segments, because there are no universal recommendations on this matter. Focus on the niche, the guest's life cycle, and other factors.
The range for each indicator will be different for different companies. Let's say you are an expensive restaurant in the center of Moscow. In this case, you give a "good" rating if the customer comes back to you in 2 months. If you are a food delivery service, you can give a "bad" rating for the lack of repeat orders within 2 weeks. The amounts will also vary greatly. For an expensive restaurant, 3,000 rubles is a low bill, for food delivery it is high.
Formation of segments
So, you should have 27 segments (3*3*3) with scores like 111, 112, 113, 121, 131 and so on up to 333.
The segments may be uneven (one segment may include 90% of all guests, another - 1%). Therefore, it is advisable to split too wide groups into several additional ones, and combine narrow ones. You can also allocate more levels. However, this will complicate further work with segments, since there will be even more of them. If there are 4 levels, you will get 64 segments, if 5 - already 125, and so on.