Jim answers questions from fellow Drillers
Hi again folks, Jim Novo here.
This Driller knows his stuff, using advanced database marketing techniques on the customer acquisition side. But in a business where the bill gets paid every month and it’s about the same (electric utility), how do you get to the R and F (Recency and Frequency) to model potential customer defections? Great question, and a pithy answer.
Remember, you can use RF profiling on any behavior not just payment related activity! If you don’t know what we’re talking about with this RF stuff, see explanation here.
Q. Jim, Ordered and read your book AFTER reading EVERY page in your website. Your newsletter is outstanding and you seem to be one of the few with real life experience in database marketing with the skills to simply explain with pragmatic examples of how RFM and LTV should be used.
We are a technology based Call Center company with over 70 clients – we do a lot of the “operational” CRM stuff you refer to – Siebel, Onyx, Kana, Webline, …….., as well as a lot of custom developed SFA solutions and data warehousing solutions we developed – mostly the premise of investing to collect enough information to do the 360 view of the customer across communication medium (email, chat, phone, fax) and reason for calling (campaign, sales, orders, info, customer service….)
We have a good mix of B-B as well as B-C. We already do a lot of the demographic modeling for list acquisition (SIC codes, size, number of computers, Geo ….). One thing I noticed is that we do a lot of lead generation based upon list acquisitions along with inbound marketing campaigns that seem to address one shot Sales, not recurring sales.
For example, we sell and service de-regulated energy for one client – this is sell once, then service. Since they pay every month for the service, how do you suggest the RFM model be used for service based sales since there is not really an R or an F??? We still have acquisition and retention problems, but we mainly focus on operational efficiency through technology, not strategic use of CRM data. I would really like to be able to add real value based upon the data collected.
I know this is not your forte, but I was just curious if you had any opinions using CRM data in an RFM model when the product is basically recurring service.
A: Thanks for the compliments on the site, book, and newsletter. I hope they will be helpful to you as we try to get a firmer grasp on these subjects this year!
It’s a little tough to provide you a direct answer to such a broad question without more details, but in general, R and F are highly predictive of any action-oriented behavior. In a “billing / service” business like a utility, you sometimes have to hunt a bit harder for the action you want to model as predictive.
For example, at Home Shopping Network, use of the automated ordering process (touch-tone interface to the ordering system circa 1990) was very highly correlated with Future Intent to Purchase. Not exactly a traditional RF action, to be sure, but a falling RF score on use of the interface was very highly predictive of a defecting customer.
In interviewing customers with falling RF scores on the interface, we found what this really meant to them was “the thrill is gone,” meaning they felt no urgency to order anymore so used the interface less – the “hard” data point representing the “soft” feelings of the customer expressed through their behavior. In other words, the beginning of the end of the LifeCycle.
How’d we figure that out? It took a long time, and we used some advanced modeling techniques to locate the correlation. Once found, it became part of the RF modeling
process and put into an “RF Grid.”
(For those who haven’t read the book, an RF Grid is the most advanced (and highest ROI) implementation of the Drilling Down method. It combines the Recency-Frequency behavioral measurement with customer LifeCycle information generated by RF Scores over time).
If you have a website or telephone “self service” interface, falling use of it might mean customers are getting ready to defect, or it might mean they are satisfied and are going to stay long term. There’s no way to tell in advance, but the customer behavior will “speak” and tell you which it is.
Here’s what you might be able to do:
- Make sure you understand all the data points available to you
- Isolate “best customers” – those who signed up and stayed signed up for the longest time, with the least cost (variable cost to you – installation, marketing etc, not in terms of total calls to the center).
- Run RF profiles over time (LifeCycle) on each piece of “action-oriented” data available to you, and determine which provides the highest correlation to best customer behavior.
For example, high RF of calls to the center might be highly positively linked (good service
leads to better customer) or might be highly negatively linked (billing problems create repeated calls = mad customers who disconnect).
A falling RF score might be good – less recent and frequent calls = higher satisfaction – or may be bad – less recent and frequent calls = customer “apathy” or indicates they are looking for an alternative service. In service businesses, you generally look for sharp changes in behavior – a drop of 30% in usage, and increase of 50% in calls. These are good targets for automation since they’re quite clear cut.
Also, as you probably know, bundling, if available, usually results in longer term customers. The reverse is also true – customers who reject bundling tend to be short term customers.
And finally, source of customer is absolutely critical in this kind of business, especially since
your “markets” may be geographically constrained. Since you are an electronically driven, data-dependent acquisition shop (SIC codes, size, number of computers, Geo ….) you have the luxury of looking at RF by customer acquisition source. Good customer retention starts with proper customer acquisition, and it should be relatively easy to look at LTV by customer source (even without using any RF, at a simple level – a “quick take”).
Here’s what I mean. Pick a start date, say one year ago, and take a quick look at your highest value customers (gross billings?) over this time and see where (what campaign, data element, etc) they came from. Then look at lowest value (disconnected?) customers from the same start point, and see where they came from. If there are differences, you’re on your way to finding the answer you’re looking for. In addition, once you determine their is a difference, survey a subset of each group and try to find the commonality in the groups and differences between the groups. This links the data to the emotions and provides a backdrop for improving acquisition technique.
Don’t try to do this starting from a “micro” level and looking up. Start with macro ideas (geography?) then “drill down” (couldn’t resist) a layer, then another. When you get down to the level where their appear to be no sizable differences between groups anymore, you’re
done. Going any lower is just “noise.”
Hope this was helpful! If it was, and you think the book, site and newsletter are valuable, please consider sending me your “review” of these tools for publication on the Drilling Down website.
Download the first 9 chapters of the Drilling Down book: PDF