January 05, 2016

Are You Terrified By A Personalized Merchandise Assortment In A Catalog?

Yup, you think to yourself, "well, that's too complicated and it's fine that we send a 128 page catalog and a remail of that catalog two weeks later and a second remail of that catalog two weeks after that ... so what that the strategy was conceived in 1989, that means we have ample history to develop an accurate forecast from. #tradition"

Why not start simple?

Start with email.

Don't you think your favorite email vendor (MailChimp is displayed above, you obviously have your favorite) would fall all over themselves to help you develop personalized merchandise campaigns, based on prior purchase history and website visitation activity?

I promise you, your favorite email vendor hired a Data Scientist, or ten Data Scientists. They're just sitting there, CPU idle, waiting for somebody to give them a meaty challenge. These geniuses probably don't even find your challenge challenging. How's that for a sentence?

Anyway, run a test.

Test design.
  1. Pick a campaign ... maybe an Easter campaign.
  2. Randomly split your email audience in half.
  3. Half #1 receives your typical campaign ... one version, 20% off or whatever, brand appropriate message ... guaranteed to generate $0.08 per delivery. Blah blah blah blah blah!
  4. Half #2 is split into ten different merchandise-centric treatments. Based on prior customer purchase history and visitation activity, the customer is assigned to one of ten different merchandise-centric treatments. Your email vendor works with your creative team to conceptualize what the ten versions are. Your in-house analyst works with the Data Scientist at your favorite email vendor to align customers to one of the ten different versions, based on prior purchase history.
  5. Deliver the darn email versions!
  6. Measure results.
  7. Salivate over the possibilities.
  8. Ask for a promotion when sales come in +20% to your typical campaign.
Use email marketing as a "proof of concept".

For crying out loud, my team at Nordstrom, in 2001 ... in 2001 mind you ... assigned customers to an online marketing strategy where 10 unique email creative/merchandising treatments were delivered in each campaign. In 2001. Bill Clinton was still President in January 2001. We easily popped a 20%+ lift in campaign performance by delivering targeted merchandising campaigns to each customer. By 2006, we were sending hundreds of unique campaigns to customers, versions tailored to customers with specific store employee preferences. In 2006. If we did this stuff 10-15 years ago, there is no excuse for not doing this stuff today.

If the tactic works, and you get a 20% lift in email marketing performance, then there are two logical conclusions.
  1. You should be executing this type of strategy in every email campaign, not just one campaign near Easter.
  2. If the strategy works for email marketing, it is certainly worth testing in your catalog marketing program, right?
What have you got to lose?

Contact your favorite email vendor, right now.

Get busy!

Let me know what you learn.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.

When Winners Aren't Quite Winners

It's common to measure winners via total demand generated. It's an easy calculation. But it's also the wrong calculation. It'...