Advanced Marketing Segmentation: Behavioral & Predictive Modeling

The last 3 years have seen a dramatic shift in marketing that have made it highly technical.  So much so that many companies shifted significant marketing related functions over to their IT team.  This was, and continues to be, a big mistake.

YES – Marketing has become extremely complex.ted-goff-woman-stands-before-two-charts-on-wall-the-complicated-one-says-finding-cartoon

What needs to happen in order for today’s CMO to be successful is simple.  The CMO needs to embrace technology, and skillfully blend it with creativity.  She/he needs to help their teams grow more technical skill-sets, and understand the power of data to the success of their marketing efforts. Why is this necessary? Because popular models of segmentation have become so commonplace. Chasing leads by standard demographics and KPI’s is the quickest way to achieve mediocrity, find the spam box, lose valuable customers – and ultimately have to answer for sliding revenues.

Let’s start by examining the most popular segmentation models:

  • Standard Demographics: Segmentation by Gender, Age, Location
  • Acquisition Based: Segmentation by Products Purchased
  • (RFM) Recency, Frequency and Monetary Value: Segmentation based on How Recent, How Often and How Much is purchased.

The primary reason Amazon has been so successful is that it committed early on to Drilling Down into the shopper BEHAVIORS in order to PREDICT shopper NEED.  The concept of “Customers Who Purchased This, Also Liked…” is a lot more complex than it looks on the surface. Product searches, purchases, navigation, clicks, trends in buying and situational interaction is collected and analyzed to build behavior trend models that lead to PREDICTIVE models.  This is what keeps Amazon ahead of most everyone else.  Don’t believe for a minute that you cannot achieve the same objectives – even on a tight budget!  I promise you, it can be done and I have first hand experience seeing it and implementing it.reveal-listen

Let me give you a few simple examples of how to more effectively break down data that you already have – and lay the groundwork for more productive interactions with your prospects and customers alike.

Behavioral (Needs Based) Segmentation:

  • Motivational Behaviors: Filter your list by – purchasers that only buy product “on sale“, purchasers that buy on the “last day” of a promotion, purchasers who only buy with coupons, customers that return items…keep thinking about this…what motivates interaction??
  • Categorical: Go BEYOND product-based upselling and start grouping or clustering your segments by categories.  For example, a customer who has purchased Jimy Choo dress shoes is not only interested in Jimmy Choo dress shoes. In fact, they are more likely to be interested in similar products or complimentary products – so broadening the appeal by category, but tailoring to the customer (by item specifics ie size) will encourage MORE purchasing.
  • Brand: Taking ‘categorical’ a step in a slightly different direction… If the customer has purchased more than one pair of Kate Spade heels – then don’t just send an email pushing more Kate Spade heels…that would be VERY short-sighted.  This customer has given you fantastic information about themselves a) they know you for the heels you sell and b) they love kate spade product.  Your job just got exciting – because you have a great opportunity to present them with a full array of {Kate Spade} products – from Handbags, to Wallets, to Scarves, Jewelry and More!

Predictive modeling takes groups of data even further – to develop a cadence that walks in step with the needs of customers – and increases the likelihood of invoking a purchase.

The FIRST RULE that I communicate to clients is this:

START with an end in mind.

Segmentation should not be about creating marketing buckets.  Successful Segmentation is All About SOLVING A PROBLEM. So – create segments based on what question or issue you are trying to solve on behalf of your customer!

Here are your simple steps to effective (behavioral and predictive) segmentation:

  1. Define the Problem or Issue that Your Customer Needs Solved
  2. Develop a Hypothesis About the Solution You Will Suggest
  3. Determine the Segment Variables needed to Identify the Customer Need Accordingly
  4. Collect, Filter, Normalize and Pre-Process Your Data
  5. Create Appropriate Segments that Will Benefit from Your Solution
  6. Execute Actionable Steps to Communicate the Solution to those Customers in Need

problem-solvedSubject for another day: Marketing Automation Solutions – there are many SaaS tools available that can help to automate the tasks of behavior monitoring (ie activity scoring) that can be setup to automatically send dynamic content to customers.  Timely and meaningfully tailored content has been proven over and over to be highly effective.

Behavioral & Predictive Modeling will help you meet customers where they are & help to cultivate loyalty, breadth of interactions, frequencies of purchases, and, oh….let’s not forget…a healthier bottom line for your business 🙂

Advertisements

About ecommadvisor

E-commerce enthusiast, evangelist, speaker and consulting leader. Founder of several successful e-retail businesses including grapevinehill.com - #1 footwear retailer in the world on eBay and Amazon. Now a full-time consultant, helping small to mid-sized businesses worldwide achieve their multichannel sales and marketing goals.
This entry was posted in Advanced Segmentation, Beyond e-Commerce, eCommerce "Best Practices", Marketing Automation, Site Conversion, Uncategorized and tagged , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s