2007 was a banner year for Multichannel Forensics (book, paper). The methodology was applied to more than twenty-five brands and more than one hundred products/channels. There are many insights and key learnings from spending a "year in the field".
E-Mail Marketers: The dynamics of your business are more sophisticated than your metrics allow you to measure. Your software tools and vendor community provide insights that don't allow you to see the full impact your programs have on the brands you manage. Overall, you provide more value than your measurement tools allow you to quantify.
Search Marketers: There are rich dynamics/interactions between your channel and other advertising channels (e-mail, catalog, other online marketing). If you work for a cataloger, make sure you get proper credit for the sales you generate!!!
Catalog Fact #1: Matchback algorithms are flawed among the audience of customers who purchased in 2006/2007 using both the mail/phone channel and the online channel. All of the rhetoric you've heard over the past decade is less applicable to today's research-oriented multichannel customer. For these customers, catalogs are only a "component" of the purchase process, not the "reason" for a purchase.
Catalog Fact #2: I'm convinced there is a fraction of your customer file that will always use catalogs to place orders over the telephone. Expect this audience to get progressively smaller over time, creating significant pressure on profitable catalog marketing activities.
Retailers: Websites plays a major role in the development of a healthy retail customer file. From a Multichannel Forensics standpoint, consider online advertising "channels" in your simulations, to understand which advertising "channels" foster long-term retail-loyal customers.
Segments: One individual told me that he ran Multichannel Forensics analyses at a "segment" level --- meaning that he ran individual analyses for different RFM customer segments. The insights are more enlightening than the high-level work you share with your CEO.
Forecasting: Surprisingly, most folks prefer using the simulation tool to build five year sales forecasts. The actual dynamics, for many, are less important than the end-point each product/brand/channel are evolving toward.
Websites: The most brilliant work I observed this year is in the use of Multichannel Forensics for key product classifications in the online channel. There are rich dynamics and interactions occurring between customers and online product classifications that almost all brands fail to see --- brands simply don't have the tools necessary to understand and capitalize on the complex interactions that are occurring.
New Customers: There continues to be an ongoing battle between customer retention/loyalty advocates, and customer acquisition proponents. Multichannel Forensics repeatedly indicates that customer acquisition is the "secret sauce" that fuels successful businesses. Don't listen to that "it costs eight times as much to acquire a customer as it does to retain an existing customer" stuff. Keep acquiring customers.
Control Groups: Multichannel Forensics provide brilliant insight for folks who execute year-long catalog or e-mail holdout groups.
The Square Root Rule: Multichannel Forensics combined with The Square Root Rule provide fantastic insight into the interaction between a healthy business, and the differing rates of advertising required to achieve a healthy business.
These represent a fraction of the learnings from 2007. What did you learn about your business, after conducting a Multichannel Forensics analysis?
Winners are obviously important, but your middling items ... those performing "in the middle" of your assortment, well, they are i...
It is time to find a few smart individuals in the world of e-mail analytics and data mining! And honestly, what follows is a dataset that y...
When introducing the concept of Marketing Budget Experiments ( click here for pricing details ), I shared thoughts on payback windows. If yo...
As usual, my summer schedule will dial back just a bit ... maybe three posts per week instead of five, sometimes four, sometimes more. And y...