Most e-commerce businesses lose money or breakeven on a customer’s first order and make money from the second order. This means that database marketing to existing customers is vital.

Most database marketing is now e-mail and Machine Labs has a great integration with MailChimp to setup your segments ready to mail.

There’s also traditional direct mail which is more powerful than email but also much more expensive. For European companies, direct mail is less regulated than electronic communications with consent generally not required. Machine Labs exports the segments from your database as CSV ready to go to your mailing house.

Finally, database can also be about creating a Custom Audience on Facebook and targeting your customers there. Again, Machine Labs can create that audience for you with a couple of clicks.

Segmentation

The central challenge in database is working out who to mail. This is called segmentation.

The standard method is called RFM where the database is split into segments according to Recency, Frequence and Monetary value. At Machine Labs we use this format for RFM:

Recency Frequency Monetary Value
Current
(0-3 months since last order)
Single Buyer
(1 order)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Multi Buyer
(2 or 3 orders)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Regular Buyer
(4 orders or more)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Recent
(3-6 months since last order)
Single Buyer
(1 order)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Multi Buyer
(2 or 3 orders)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Regular Buyer
(4 orders or more)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Lapsed
(6-12 months since last order)
Single Buyer
(1 order)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Multi Buyer
(2 or 3 orders)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Regular Buyer
(4 orders or more)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Dormant
(12 months or more)
Single Buyer
(1 order)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Multi Buyer
(2 or 3 orders)
Top third of LTV
Middle third of LTV
Bottom third of LTV
Regular Buyer
(4 orders or more)
Top third of LTV
Middle third of LTV
Bottom third of LTV

This segmentation model would allow you to contact less frequently customers who are Dormant, focus more on higher value customers (Middle to Top third of LTV) and focus more on customers who have bought more frequently.

Smart Segmentation

The standard segmentation model works well but can be a bit crude. For example, good customers who would re-order anyway could be targeted with a strong offer. Or customers who are less engaged could receive too much communication (adding cost or reducing inbox delivery rates for email). Good customers could be at risk of leaving for a competitor and require an offer to keep them loyal.

Using a mix of order patterns, open rates (email only), promotional engagement our machine learning segmentation offers a smarter model. We call this Smart Segments which is much simpler for you to work with:

Bronze Customers
(Least valuable)
Silver Customers Gold Customers
(Most valuable)
Just bought
(Green)
Just bought
(Green)
Just bought
(Green)
Order due
(Green)
Order due
(Green)
Order due
(Green)
Order overdue
(Amber)
Order overdue
(Amber)
Order overdue
(Amber)
Lost
Still engaged
Lost
Still engaged
Lost
Still engaged
Lost
Disengaged
Lost
Disengaged
Lost
Disengaged

This model is much easier to work with and allows you to target the right customers with the right message at the right time.