2006: Online Generates In-Store Sales

Five popular songs from 2006:

• "Bad Day" - Daniel Powter
• "Temperature" - Sean Paul
• "Promiscuous" - Nelly Furtado featuring Timbaland
• "You're Beautiful" - James Blunt
• "Hips Don't Lie" - Shakira featuring Wyclef Jean

In 2006, people started asking good questions.  Here's one:

• "If we invest a dollar in paid search, what is the short-term online payback, and what is the long-term in-store payback of our investment?"

Good question!

This question isn't easily answered by attribution algorithms, which basically did not exist in 2006 and wouldn't be able to handle the problem in 2013 in most cases.  Too often, attribution experts look backward ... they'd be better served by looking forward.

The problem was addressed via Online Marketing Simulations!

We simulated how customers acquired via search would evolve over time.

Customers, it turned out, would evolve favorably!

We learned that a dollar of paid search would yield:

• $2 of online demand in the next two weeks.
• $2 of in-store demand in the next two weeks.
• A 3-1 ratio of in-store demand to online demand in the next two years.
• A nearly retail-only customer within five years.
• 50% of search customers were existing customers.

The combination of short-term online/retail demand, coupled with a huge long-term retail payback, allowed us to invest much deeper than we otherwise would invest.  Dollars formerly spent on break-even catalog activities were reallocated to paid search activity that drove future in-store business.

Retail folks like online activities that grow in-store sales.

The CFO likes online activities that have a handsome payback.

And the online division likes getting more money to spend on marketing to grow the total business.

It's 2013, and few retailers have the sophistication that Nordstrom had back in 2006.  Online visits were linked to in-store purchases.  The path from paid search to an online visit to a store purchase over three days could easily be tracked.  The information gleaned from an integrated database in 2006 set the stage for what folks now call "omnichannel growth".

Similarly, we have challenges in 2013.  We have to prove that mobile/social, to an 18-39 year old audience, causes changes in e-commerce, and changes in-store performance.  Who, to your knowledge, is good at illustrating this dynamic - not looking backward, but by simulating the future?  Leave a comment in the comments section.

By the way, you might wonder how we did this kind of work, back in 2006?  Well, we had a really solid information technology team.  This team, and members of my analytics team, wired data from our web analytics platform, our in-store clientele software system, order-entry systems, and retail point-of-sale systems.  In 2006, mind you.

We knew that our good customers visited the website three times a month, visited a store two times a month, and purchased one time a month, with 85% of purchases happening in stores.  When you know that customers behave like this, why in the name of all that is good would you try to MAXIMIZE CONVERSION RIGHT NOW BY OFFERING DEEP DISCOUNTS?  We knew the customer would visit our website every ten days, and had virtually no intention of buying online.  When this is your customer relationship, then you know that almost everything the punditocracy is selling is fundamentally wrong.

Now, if your customer has a 20% chance of buying again next year, then yes, you have to get everything you can out of every visit.  But when you know that an online visit will lead to an in-store sale (almost monthly, mind you), well, you market differently.  

Even back in 2006.