October 28, 2007


Over the past decade, householding logic changed from a set of geeky algorithms to a core competency that multichannel catalogers need to have.

Today, I received a direct mail offer that I shouldn't have received. I have two accounts with this company, one at one address, one at another. The two addresses have the same billing address, and use the same credit card number.

In other words, the direct marketer should have known that I already subscribe to the product they were offering.

There are many elements of householding that need to happen in a multichannel catalog business.
  • Multiple computer cookies linked to one identifier. Customers use multiple computers --- where possible, the cookies should be linked to one unique identifier. Your web analytics team should analyze data by individual cookie, and by a unique web usage identifier.
  • Multiple e-mail addresses linked to an individual customer. The data should allow "many-to-many" relationships --- so that kevinh@minethatdata.com can be linked to myself, and be linked to my wife. In addition, Kevin Hillstrom should be linked to kevinh@minethatdata.com and kevinh@zzzzzz.com.
  • E-mail addresses and web usage identifier linked to customer (i.e. Kevin Hillstrom).
  • Customers linked to households (Kevin Hillstrom and Wife linked to our home address).
  • Households linked by credit card numbers, mail-to/bill-to/ship-to addresses, and phone numbers.
From a marketing standpoint, you're free to execute campaigns using any degree of granular data you measure as being effective.

From a "running the business" standpoint, having data linked together properly is an essential competency. I'd rather have my data householded properly than have a glowing multichannel marketing strategy.