View this article as a PDF file
Libey Incorporated
Economic Outlook
Secrets of the Catalog Master
Vol. MMV No. 5 August 2005
Advisors and Intermediaries To The Catalog Industry
Philadelphia, Pennsylvania
Donald R.
Libey, Editor
MeritDirect Press to Release First Book
by Donald R. Libey
Libey and Pickering on RFM and Beyond
MeritDirect Press, the book publishing endeavor of MeritDirect, will release its first book at the September 2005 DMA Conference in Atlanta. Written by Don Libey and Christopher Pickering, the book is titled Libey and Pickering on RFM and Beyond. It explores optimized RFM plus product, channel, position and market. The 360 page, hardbound book will be distributed in the U.S. through Amazon.com, in Europe through Synergy Partnership Ltd., and in Canada through Lloydmedia, Inc. MeritDirect Press publishes books that assist multi-channel direct marketers to become more successful and profitable. Established by the partners of MeritDirect, the mission of MeritDirect Press is to bring knowledge and wisdom of multi-channel direct marketing to clients and the professional direct marketing community. Here are a number of advance excerpts from selected chapters of the book.
Chapter One: Recency
Recency and the Customer Database: Libey
Database technology is neither mystical nor arcane. There are those who would have the database seem mystical and arcane because they have a vested interest in keeping it that way. Actually, databases are nothing more than shoeboxes.
Years ago, peddlers scribbled customer information notes on 3 x 5 cards and filed them alphabetically by customer name in a shoebox. Every purchase was recorded, including the date, the product, the price, personal details to trigger future sales, gossip picked up from neighbors, and other details supporting future sales and recency. Every time a sale was made, the card was pulled and updated and re-filed in the box. Before a sales call was made, the card was consulted and the intuitive knowledge brought top of mind for the crafty peddler. Children’s names, color preferences, sizes, birthdays, returns, income and all manner of useful information was recorded to assist the peddler in making another sale now, or on future trips. The hearts and souls of the customers were contained in those shoeboxes.
Now, it should be recognized that not every peddler used the shoebox method; only a few mastered this technology. The rest proceeded from place to place guided by the seat of their pants, relying on serendipity or luck. But the shoebox peddlers were successful where others failed. They employed knowledge rather than charm, and knowledge always wins. They were actually the first true merchandisers because they knew what the customer wanted before the customer knew and they knew where, when, why and how to present those products to each customer to assure a sale.
Today’s advanced database technologies are basically file cards in a shoebox. For all of our sophistication and technological advancement we still scribble notations on the customer cards whenever we make a sales call and whenever someone buys from us. And, we are interested in the same information that the 19th century peddler captured. It is interesting that the peddlers experienced the same technology demands we experience today. At some point, they found they needed more memory for their shoebox databases. Their solution: they doubled the memory by moving up to 5 x 6 cards. Does anyone remember the early Remington Rand Cardex systems used for customer files?
The method of organizing the customer database is, essentially, alphabetically by name. We use phone numbers, customer numbers and other identification notations, but we find a single customer by sorting in one way or another. The end result is still the same: We get the information we need to make another sale. One component of that information is recency.
Where the peddler had one shoebox, we have multiple shoeboxes. We may have one for catalog customers, one for retail customers, one for web customers, one for e-mail customers, one for infomercial customers, and so on. Where the peddler made notes on one card, we may be making notes on five cards, as well as cross-notes on all cards. When we get a catalog order, we note it on the customer’s card in the catalog shoebox. We also note it on the customer’s card in the web shoebox, because last time that customer bought through the web site. And we note it on the card in the infomercial shoebox and the one in the retail shoebox, because the customer has bought in those channels also. It’s important. We need to know everything. It just might come in handy. After all, we’re obsessed with this stuff.
The peddler, preparing the wagon for a swing through the territory, would go through the customer cards and redistribute them into a logical route format–the circulation plan–that ordered the trip and the information about the customers to be contacted on that trip–or the contact strategy. Perhaps only those customers who had purchased every year for the past five years would be called on during that year’s trip. Perhaps only those customers who purchased something two years ago but had not purchased anything during last year’s trip would be called on. If the peddler was trying to expand the business, perhaps people who had never bought, living close to customers who had bought regularly, would be squeezed in between customer calls. A number of choices were available to the peddler depending on the business strategy and the accurate knowledge of customer history found in the shoebox.
Once the peddler had selected the customers to be visited, the cards were put in traveling order. Villages and settlements would be scheduled and allocated against the available number of days for the trip. Customers and prospects would be added or dropped to make the route efficient, particularly with regard to expected sales and profits. No successful peddler ever traveled a random route trusting to serendipity or luck; successful peddlers built marketing plans, predicted product purchases, calculated predictable sales and earnings, and managed inventory, just as we do today.
The peddler made a relationship between customer purchases and trips. Over time, patterns emerged that could be relied upon to repeat. One customer might buy every trip regardless of the trip timing. Another customer bought only on the fall trip after harvest when cash was in hand. A third customer bought only every other trip, sometimes in the fall and sometimes in the spring; and a fourth bought only every other trip in the spring. Thus, the peddler understood seasonality. The intuitive knowledge of these patterns was built into the customer cards and ordered in the mind of the peddler through the intimate and tactile understanding of the shoebox database.
The peddler’s objective was no different from the objective of the database marketer of today: To optimize sales and to generate the greatest possible profit with the least possible expenditure of resources. The shoebox system worked then, and the database system works today.
Key to the success of the shoebox and the database, however, is the recognition that the primary purpose was and is marketing; that is, selling more stuff to more people. The prospecting and customer databases exist first and foremost to create additional sales and to keep customers buying. They do not exist for the convenience of accountants, for generating invoices, for managing inventory, or other purposes, although these are secondary by-products of the database. Too often, too many diversions interfere when data becomes organized into information; too many parasites attach themselves to the database and obstruct the primary purpose: Selling. CEOs who embrace a formulaic approach to marketing, supported by sophisticated and accurate database practices, must first swear an unwavering allegiance to marketing to their customers and reject all other unessential, non-selling noise interfering with that one dominant purpose.
The Master Marketer having control of multiple channels also has control of multiple sub-databases that accurately portray the individual channels. It must be possible to fully understand database information for the catalog channel separately from the on-line channel and separate from any other channel. It must also be possible to fully understand the database information for the combined databases of all channels and to be able to see changes and trends intra- and extra-channel; otherwise, how do you know where to put your money?
Recency and the Customer Database: Pickering
Database technology is neither mystical nor arcane. So says my writing partner. I would make one small modification: Database technology should be neither mystical nor arcane. In the everyday business world, however, your database workings are sometimes mystical or arcane.
To be certain, there are more choices out there for affordable and powerful database technologies than ever before. A new cataloger can get a reasonably powerful software package for a few thousand dollars. The disconnect generally comes in the use of the systems.
We are marketers, so we think that the systems should primarily support marketing. Unfortunately, our peers in order fulfillment and accounting think that the systems should work for them. Middle ground can be found.
Looking at recency, it seems obvious that it applies to the date of most recent purchase by an individual (or a site). What happens, though, when an existing customer calls to place another order and the customer service representative (CSR) cannot quickly find the existing customer record? Often it is easier for the CSR to create a new record.
This duplicate record creates several problems. First, it overstates the number of customers. Secondly, it misstates the recency of buyers on your file. It also creates additional work in setting up a new account, slows down the ordering process by making the customer give data that is already resident elsewhere in the ether of the database. The duplicate record problem is frequently exacerbated by the compensation structure for CSRs. It is typical for CSR compensation to be based on volume of calls handled. If their perception is that it is faster to set-up a new account, then that is what they will do. To address this problem, many marketers work with their CSR group to develop ways to improve the ordering system. Using customer numbers and finder numbers are effective ways of addressing order input issues.
You likely assign your customers a customer number of some sort. This allows look-ups, enables your relational database to access their item loop (history of products purchased), and endless other useful things. And most enlightened direct marketers print their customer number on their catalogs, frequently in a colored box.
On the prospecting side a similar process can occur. A non-permanent number can be assigned to a prospect mailing piece, much like a customer number, that will allow the database to populate name and address information into the order screen when a call is received. The problem is that all too frequently the CSRs have not been trained to ask every time for the customer or finder number. Once trained, or incentivized, CSRs can capture these numbers, which yield many benefits:
- Faster, more accurate order entry.
- Fewer duplicate records added to the database.
- Higher key code capture rates, particularly on new customers.
In addition to trying to ensure the accuracy of recency, marketers today have more channels to track, analyze, and follow. In some systems (generally newer) it is easy to add fields to track recency by channel and in legacy systems this is sometimes quite difficult.
If you have a system that can keep a recency by channel, by all means, do so. Also contemplate what other similar data you should keep:
- Database recency–the recency associated with the most recent purchase irrespective of channel.
- Channel of first purchase and channel of most recent purchase.
- Date of first purchase.
The database recency will be the easiest to use recency and will drive much of the segmentation. Database recency should be the recency used when measuring overall database performance, such as house file counts by recency.
The channel of first purchase and last purchase will be critical to see how your customers like to purchase from you. Many catalogers with web presence find that they are most effective acquiring new customers via the mail, but once the initial purchase is made repeat buyers prefer to buy via the web. To confirm that theory simply look at recent two time plus buyers whose first and most recent purchases were through the same channel versus those that migrated from one channel to another. You cannot asses this, though, unless you have your database set up to capture and report on this vital information.
Date of first purchase is a little used nugget of information that can be unbelievably useful in determining house file circulation strategies. Drilling into customers by the year of first purchase allows you to do ‘vintage analysis’ to see if, for example, customers acquired in a given year perform differently than those acquired in another year.
Consider 1998 buyers versus 1999 buyers. There was no large scale change or shock to the system in 1998, so we can treat that as a ‘normal’ year. In 1999 we were on the cusp of a change, the New Millennium. As discussed, this led to some accelerated buying. It also led to some different buying habits. In 1999, those catalogers who sold goods like camping equipment, freeze dried foods, water purification and survival equipment, fuel storage, first aid, and power generation products had a better than average year. Some had their best year ever.
As these customers progress and attrite through the life-cycle, they cause anomalies. Many of the purchasers were making a one-time purchase: they aren’t campers by nature, but they bought a few lanterns and sleeping bags. You won’t see them again. In 2000 and 2001, as these buyers progressed down through the Recency cells, it looked as though the company was doing a poor job of retention. In 2002 and 2003 it looked like there were large pools of re-activation candidates. The problem was there was the ‘glut’ of event driven buyers.
By using date of first purchase and creating recency reports for each year, the Master Marketer can see these anomalies. Taking this information they then begin to develop house file circulation plans that treat those buyers differently, perhaps adding additional selection criteria or redeploying circulation from these segments to more productive ones.
These examples, I am hopeful, show that a big key to understanding your database technology is getting familiar with how the data is entered, rolled-up and updated, and extracted. The Master Marketer has an understanding of all of these.
Chapter Four: Optimized RFM
Optimized RFM Plus North American Industrial Classification System (NAICS): Pickering
The NAICS codes feel like the metric system: it is simply easier to use than the incumbent SIC (Standard Industrial Classification) system and there is tremendous resistance to it.
The fact is that most major providers of business-to-business information offer SIC codes—not necessarily because they feel that it is a better system (it certainly is not), but in response to demands of the market. Most business-to-business marketers have SIC codes embedded in their systems and it is easier to keep in a BAU (Business As Usual) mode. That being said, using either NAICS or SIC is the most logical place for business-to-business marketers to look for the X factor to add to their RFM segmentation. The power is strong, perhaps only equaled by employee size.
Both codes, NAICS and SIC, go to at least four digits (NAICS actually goes to six and there are proprietary SIC classifications that go to eight digits). For all but the very, largest business-to-business marketers this is not practical because it creates too many segments of too small a size. Paralysis by analysis will inevitably set in. To avoid this, use the codes at their two-digit level or even at a super-two digit level—there are some industry standard groupings that collapse eighty plus two-digit groups to about twenty groups.
When you have the level of segmentation you can easily use, pockets of much higher than average and much lower than average value (defined by lifetime value, lifetime dollars, monetary momentum, etc.) will appear. Rather than allowing this simply to become ‘gee whiz’ information, ask the following questions:
- If my best segments defined by RFM + NAICS accounts for forty percent of my value and twenty percent of my customers, am I targeting an appropriate and corresponding percentage of my house file circulation to these segments?
- Am I looking at these high value RFM + NAICS segments and targeting them in my new customer acquisition, even suffering a lower response rate because of the higher payoff?
- 3. Am I looking at these low value RFM + NAICS segments and avoiding them in my new customer acquisition, even foregoing a higher response rate because of the lower payoff?
The Master Marketer has asked and answered these questions and continually re-visits the question to see how the customer base has changed. It is all about making a closed loop system: use the intelligence gained through analysis to drive better decisions.
Next Page
Pages: 1 2
3
