February 19, 2013

Weighted Channel Pairs

In this image, distance matters.

Mail and omnichannel are far apart.

I recently read an article, suggesting that mail is one of the best ways to drive a mobile transaction.

Maybe.  But you can answer this question, using your own data.

In your database, create what I call "channel pairs".  In other words, weight transactions based on historical importance.  Transactions in the past year count for a weight of 1.00, transactions 13-24 months ago count for a weight of 0.50, transactions 25-36 months ago count for a weight of 0.25, all older transactions count for a weight of 0.125.  Multiply the weight by the dollar value, and you have something!

After doing this by channel, identify the most important channel, and the second most important channel (by weighted dollars).  Run a query that identifies the most common channel pairs.
  • Customer #1 = Retail First, Email Second.
  • Customer #2 = Tablet Commerce First, Catalog Second.
  • Customer #3 = Retail First, Mobile Phone Second.
  • Customer #4 = E-Commerce First, Search Second.
Now, think about your strategy for each of the four customers above.  Does the strategy include "doing everything"?  Probably not.  Just as important, you'll learn which channels are aligned with each other.  This type of data leads to the distances illustrated above.  Overlay customer demographics, and you have something!

However, you probably won't find much of a link between mobile and direct mail.  And that's ok.