For years recency, frequency, and monetary value (RFM) was king. Even catalogers that shied away from lavish database models segmented their buyers according to how recently they’d made a purchase, how frequently they purchased, and how much they spent. Some marketers even got fancy and added a “P” to RFM to track the type of products purchased.
But now industry professionals advocate introducing another consonant to RFM: C, as in channel.
“What marketers are beginning to understand is that buyers responding in different channels have different characteristics,” says Steve Tamke, vice president of Hackensack, NJ-based list services firm Mokrynskidirect. “A buyer who receives a catalog and buys on the phone is going to behave differently than a consumer who buys from you after getting an online promotion.”
In short, if you’re not segmenting your buyers according to channel as well as by recency, frequency, and monetary value, you could be making faulty prospecting and circulation decisions. That’s what Erik Martinez, vice president of marketing at Arlington Heights, IL-based home decor merchant Design Toscano, found. When the company began comparing Internet buyers with phone and mail buyers five years ago, Martinez noted that Web buyers spent about 5% less per order than offline customers. But Martinez also discovered that Design Toscano could reduce the number of print catalogs sent to them.
Instead of mailing 10-12 catalogs to Internet buyers, Martinez says, “we might send six or seven catalogs.” In lieu of those skipped catalog mailings, Design Toscano will send the Web buyers e-mail messages instead, at a savings of nearly 50%.
“Our sales decrease by 37% when we substitute e-mail for a catalog contact [to this degree], but we almost double our profits,” Martinez says.
Design Toscano’s strategy of cutting back, but not eliminating, catalog mailings to Web buyers is optimal. As Jim Coogan, president of Sante Fe, NM-based consultancy Catalog Marketing Economics, notes, the print channel is often the tool that spurs online buyers to head to the merchant’s Website.
According to Broomfield, CO-based direct marketing data and services provider Abacus, on average 65%-85% of Website sales come within four weeks after a prospect receives a print catalog, which suggests that the mailpiece was indeed a driver of the Web traffic.
Getting started
Before you can segment your database by RFMC, you have to be able to identify the “C” by which customers purchased. The first step is to flag Web transactions as they occur, Coogan says.
Casey Carey, Abacus’s vice president of marketing, advises clients to look at offer channel and response channel separately. Likewise, Coogan suggests segmenting online customers into “pure Web buyers” vs. “catalog-driven Web buyers.”
The offer channel, Carey says, is the one that matters in terms of RFMC. “If you separate out catalog responders from e-mail responders from retail-only buyers, you can start to design more-effective contact strategies for each group.”
To flag those who bought from your Website following a print mailing, you can match up the Web buyers against mail files. And by examining Web traffic logs, you may be able to segment out the pure Web buyers by source, such as price-comparison sites, search engine optimization (SEO), or affiliate marketing.
Segmentation in action
Once you have the segmentations intact, Mokrynskidirect’s Tamke says, you can put together a circulation or contact strategy specific to that particular channel.
For example, you might want to pull out a segment of three-month phone buyers who ordered $100 in merchandise and one of three-month pure Web buyers who’d ordered the same amount of merchandise. Test a few catalog mailings and e-mail messages among both segments and monitor the performance. More than likely the Web buyers will respond less favorably to the catalogs. If that is indeed the case, you might begin to send them fewer catalogs but more e-mail.
And assuming that the phone buyers respond profitably to the print catalogs, continue to mail to this segment as you had been. As for e-mails to this segment, “you should keep an eye on the opt-out rate on the e-mails for the phone buyers,” Tamke says, “to make sure that you’re not e-mailing them too often.”
If you participate in a cooperative database, Coogan recommends having your provider model your Web buyers and your traditional buyers separately — you’ll get different and better responding prospecting names than if you model the databases one big homogenous group. What’s more, he says, the co-op databases will optimize the Web buyers on your house file and eliminate those households that are not mail order responsive. You can drop 10%-40% of Web-buying households from your print mailing list this way, he says, which should definitely save you money and boost the profitability of your mailing.
Back to matchbacks
The last step is to have a matchback process in place so that you can allocate Web orders back not just to their RFMC segment but also, in the case of catalog-driven Web buyers, to which list.
Source code capture for Web orders is typically the hardest part of getting an accurate measurement of the effectiveness of print mailings as an online traffic driver. Shopping carts are generally poor at forcing the capture of catalog source codes, Coogan says, so Web sales can be underreported if circulation managers rely solely on response reports showing those source codes captured when orders are taken.
For instance, call centers can capture 90% of catalog source codes; Web shopping carts capture maybe 50% of catalog source codes.
Coogan advocates that marketers provide an incentive, such as a percentage or dollar-off discount, to encourage online customers to input source codes. It’s well worth the effort and the incremental expense, he says. If you flag your Web transactions, segment by RFM and channel and use matchbacks to get full tracking of Web response to catalogs, you’ll have accurate sales-per-book data to use to plan future mailings.
RFMC…G
Segmenting by geography as well as by channel has helped Winchester, VA-based Rugs Direct identify where to open its stores, says president Randy Kremer. Rugs Direct mailed its first catalog in 2004. And now with the cataloger opening two brick-and-mortar stores in 2006, segmenting is even more critical.
“Segmenting has helped us plan our growth model by looking at where the buyers are coming from,” Kremer says. “It’s critical in our growth. It’s also helped us with budgeting to try to determine how we spend our marketing dollars.” The biggest challenge Rugs Direct faces is prospecting aggressively and successfully. Now with the store outlets, Kremer says, “we’ve been mailing into those store markets to prospects as well mailing to house file buyers who have bought in the store.”
— MDF