Patron Segmentation in the Library using RFM

Understanding our library patrons has been crucial to improving library services to the public, and one way the library has done so is through patron segmentation studies. Segmentation studies are typically done by dividing customers up into groups based on shared characteristics, so that businesses can tailor their services or offerings to different customer segments.

While there are many different ways to segment the customers, we will share how the National Library Board uses RFM analysis to better understand patron behaviour.

What is RFM?

RFM stands for Recency, Frequency, and Monetary. Patron behaviour is based on these 3 factors.

In a retail setting,

• Recency refers to how recently a customer has made a purchase
• Frequency refers to how often a customer makes a purchase
• Monetary refers to how much a customer spends

But of course, since there are no sales in a library setting, we have tweaked the RFM measurement to be as such:

The process of creating patron segments is then to create buckets for each RFM measure, and then to create the segments based on specific RFM combinations.

Creating the RFM buckets

There are many considerations as to how to bucket the RFM measures. Grouping by percentile is a common method adopted by many companies. However, groupings in NLB are also based on domain expertise.

For example, because NLB allows patrons to borrow books for 3 weeks at a time, one of our “Frequency” brackets was based on customers who borrowed 18–51 times in a year. On one side, this separated this group of customers from our most frequent customers who tended to visit NLB weekly as part of their routine, or 52 times in a year. On the other side, this also differentiated this group of customers from those who would visit NLB every 3 weeks (or 17 times a year), whenever they needed to return and borrow a new batch of books.

Creating the RFM segments

Lastly, the buckets for each RFM measure are combined (for example, a patron could be “555”) and mapped to a patron segment. For example

• Patrons with the highest RFM values could be considered to be “Champions”;
• Patrons with high Frequency and Magnitude but low Recency, may have been Champions previously but might need a nudge to remind them of the library.
• Compared with them, a patron with high Recency but low Frequency and Magnitude may be a Fledgling borrower whom we want to nurture.

Let’s look at an example. This patron with the following borrowing behaviour (“345”) seems to be an avid reader, who could be a potential “champion” if they were to come to the library more often. We therefore map this patron to the “Potential” segment, which may mean that NLB shares more library services with them, as a way to encourage them to return to the library.

Using RFM in other projects

A separate study on patron churn revealed that patrons who have “Never Tried” our digital segments, and are at low RFM values for physical items, are most likely to become inactive. For this group, NLB looks at how it can encourage them to also try out our digital collection. Once they have tried out our digital collection, they may find that even if they are too busy to visit the library in person, they can “visit” NLB and borrow a book at any time and from any place in the world, to find a book, browse a magazine, or just to read the newspaper.

We hope this article has been useful in explaining how RFM works, and for creating strategies to reach your network. If you are interested to get started or to see what more you can get out of the NLB Mobile App, you may refer to the following links:

• Quick Guide: https://mobileapp.nlb.gov.sg/get-started-with/nlb-mobile-troubleshoot/
• Getting Started: https://mobileapp.nlb.gov.sg/get-started-with/nlb-mobile/

Text by

National Library Board Data Office