News

Wired Innovation Insights: The Science of Segmentation for Sales Effectiveness

Wired-logo.png

September 25, 2013- 

By Neil Biehn

Take it from me, a data scientist – segmenting effectively is a science. While many marketing, pricing and sales professionals know and use segmentations, many really don’t understand just how vital good segmentations are to increasing sales effectiveness. Furthermore, we might believe that one really good segmentation is all we need, but that couldn’t be further from the truth. Every sales lever we pull has varying forces at work, including a different type of measurement. Consequently, segmentations can take various forms and sizes depending on the questions for which you need answers.

To a marketing or sales analyst, this is a daunting task. It’s hard enough to create a single segmentation, let alone multiples. Spinning up multiple surveys, talking with product managers and analyzing the competition requires a Herculean effort. And in today’s ever-changing markets, your research findings can become stale in a matter of months or even weeks. On the other hand, your data already has a ton of information just waiting to be used. Each customer interaction – offer, sales transaction and purchase order – can be used to predict how customers will respond to all of your sales moves. Here are four suggestions to build effective segmentations for your organization:

  1. Pick the question you are trying to answer. Limit yourself to just one question at a time. For example, the reason deals will close this quarter can be very different than the reasons customers buy additional products or add-ons.
  2. Determine the Key Performance Indicator (KPI). Using data to understand your markets implies you have a way to quantify different market responses.
  3. Find the factors and attributes that reveal changes in the KPI. A good segmentation reveals how different market segments act differently. Uncovering attributes that reveal a need for different strategies exposes untapped value. We’ve seen many attributes that change customers’ perceptions on offers, products and services. Some of these will be very unique to your business, while others are more ubiquitous. Be sure to check your data sources to identify which attributes could be used, including your CRM database, customer master data, product master data and sales transactions. With data in hand, take a look at charts and graphs that show the KPI by attribute value, including customer size, purchase frequency, geography, customer industry, group affiliation, lifecycle/loyalty, product category, product velocity, commodity/specialty and product cost, just to name a few. Finally, using statistical methods and/or machine learning algorithms will also help determine potential attributes.
  4. Combine the attributes to form a segmentation. Each segment can be used to prescribe a specific set of actions with respect to the customer profile and the products that are offered or sold. Statistical and machine learning algorithms can further group attributes to create a more predictive segmentation.

With each new problem we address, a new segmentation is in order. While most analysts shiver at the prospect of creating multiple segmentations for multiple purposes, there really is no other alternative, provided you are measuring a new KPI. There are new lists of attributes to test and validate our hypothesis. What works for up-sell or cross-sell may not work for promotional campaigns. To manage this complexity of new segmentations, sales and marketing analysts should look at technology. From testing new attributes to automatically updating information inside a segment, big data technology for sales can make your scientific approach to sales easier to manage, while generating incredible value.

Neil Biehn is vice president and leader of the science and research group at PROS.

Previous

Manufacturing Digital: Getting a Big Impact from Big Data: Part One

Next

The Pricing Advisor: Using Big Data to Sell More Effectively: Increase Profit with Fenced Offerings