October 13, 2009

This Week In Business: New Product Analysis, And Free Code

We tend to focus on marketing concepts that have nothing to do with merchandise ... things like strategic use of the word "new", or "free shipping today only", or the colors and fonts and links and placement thereof on our homepage.

And there's no arguing that you can improve conversion rates by focusing on this stuff. It works. All of it.

I'm here to tell you that you can generate far more profit by focusing on two things.
  1. Understand customer behavior over time, not just within campaigns.
  2. Understand the merchandise that different segments of customers prefer.

This viewpoint is considered heresy by many in marketing. And that's amazing, because the job of a marketer is to get a customer to purchase merchandise, right? It's not the job of the marketer to manipulate the customer into a complex web of conditions and criteria surrounding a free shipping promotion that extracts short-term profit at the expense of a long-term customer relationship.

Allow me to ask you two simple questions.

  1. Do your best customers prefer the same merchandise they've always purchased, or do they love to purchase new merchandise?
  2. Do your prospects, customers who have never purchased from you, trust merchandise you've always offered, or do prospects buy merchandise from you that you haven't previously offered to customers?

If you can honestly answer each question, based on data you've mined from your customer database, then you are well down the path of dominating the competition.

It turns out that every company is different, and that every company has a new/existing product "DNA" that dictates their success.

I worked with an established brand. This business had a loyal customer base that only wanted to purchase the same product, over and over and over. This company would not use "new" in an e-mail subject line, it would be pointless, the customer loved the same stuff they bought ten years ago. Every time this company tried to "modernize the brand", the customer rebelled.

I worked with another established brand. This business had a loyal customer base that craved new product. The more loyal the customer was, the more likely the customer was to deviate from the "same old same old", always desiring the latest and greatest innovations from the brand. Conversely, the new customer, the prospect, the customer who had not purchased before, would only buy the most conservative, best selling items. Clearly, the prospect only "trusted" a small subset of merchandise that the brand had offered for more than a decade. It's not hard to envision a marketing strategy that immediately boosts productivity, based on this knowledge, is it?

Identifying this Customer x Product x DNA issue is critically important. Success in e-mail marketing to existing customers is entirely dependent upon understanding this phenomenon. Success in paid search marketing is entirely dependent upon knowing if new or existing customers use paid search, coupled with merchandise preference (new or existing product), resulting in a properly merchandised landing page. Catalog marketing is entirely dependent upon putting new/existing merchandise in the first twenty pages of the product based on customer preferences.

A fundamental knowledge of these issues allow you to generate more profit than does a mastery of subject line headers, for instance. In fact, anytime you thoroughly understand how customer segments interact with merchandise, you are in a position to dominate the competition. The past decade saw us walk away from this fundamental understanding of our businesses, as we searched for mechanical e-commerce tweaks that didn't increase customer loyalty.

Now it's time for the free (and very simplistic) code. If you want to do a quick-and-dirty analysis to see if new or existing customers prefer new or existing product, then convert the simple SPSS code below to whatever software tool you use. The dataset I'm using has one row for every item the customer purchased over the past ten years. I'm simply analyzing the average "age" of an item the customer purchases (if the item has been available for 10 years, then the item is assigned a value of "10", if it was only available in 2009, it is assigned a value of "1"), by customer spending levels over the past twelve months.

Or send your data to me, and I'll run the analysis for you. Either way, you get the information you need to understand this dynamic!

get file = 'c:\datasets\Kevin_Hillstrom_MineThatData_ItemDetail.sav'.
select if (demand gt 0).
select if (quantity gt 0).
aggregate outfile = *
/break = item_no year
/demand = sum(demand).
aggregate outfile = *
/break = item_no
/year = max(year)
/years = n.
select if (year = 2009).
save outfile = 'c:\datasets\dummy.sav'.
get file = 'c:\datasets\Kevin_Hillstrom_MineThatData_ItemDetail.sav'.
select if (demand gt 0).
select if (quantity gt 0).
sort cases by item_no.
match files file = *
/table = 'c:\datasets\dummy.sav'
/by = item_no.
compute hst_demd = 0.
compute fut_demd = 0.
if (year ne 2009) hst_demd = demand.
if (year eq 2009) fut_demd = demand.
compute yrs_2009 = $sysmis.
if (year = 2009) yrs_2009 = years.
aggregate outfile = *
/break = household_id
/hst_demd = sum(hst_demd)
/fut_demd = sum(fut_demd)
/yrs_2009 = mean(yrs_2009).
if (hst_demd gt 0) hst_demd = 1.
if (fut_demd gt 0000) and (fut_demd le 0100) fut_demd = 0100.
if (fut_demd gt 0100) and (fut_demd le 0250) fut_demd = 0250.
if (fut_demd gt 0250) and (fut_demd le 0500) fut_demd = 0500.
if (fut_demd gt 0500) fut_demd = 0999.
formats hst_demd(f1.0) fut_demd(f4.0).
aggregate outfile = *
/break = fut_demd hst_demd
/cases = n
/years = mean(yrs_2009).

No comments:

Post a Comment

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

Alex Morgan (Yes, This Will Be About Merchandise)

When the USWNT Olympics Team was announced, all-time great Alex Morgan wasn't on the team. Nor was she on the four player alternate list...