Understanding the ‘Always On’ Customer Buying Journey

Today’s plethora of social channels and the capability for instant feedback has transformed the way we generate sales leads and cultivate customer relationships.

Enter Big Data analytics, which now plays a major role in better understanding our customers and their buying process. A Big Data analytical environment is critical to producing analyses that integrate a wide range of source data, enabling marketers to synthesize data into a high-level view for strategic marketing management, while providing the ability to drill down to the tactical execution level at the same time.

Accenture pointed out in Serving the Nonstop Customer (October 2012): “A customer’s path to purchase used to be linear. Now the journey is dynamic, accessible and continuous. Marketing executives need a new model that can help them become and remain relevant to their customers in this uncharted environment.

“Even casual observers of the world of commerce know by now: The traditional marketing ‘funnel’—the model that described a customer’s path to buying goods and services as linear… has lost its relevance. It’s too slow, too static and too generic to be used as a foundation for companies’ marketing, sales and service strategies, and as a guide to their execution.”

Given this shift in thinking, my team has been using high-performance EMC and Pivotal technology, such as Isilon Big Data storage hardware and Greenplum data computing appliances, to quantify our customer relationships and understand the most relevant way to communicate (and generate sales).

There is a lot of variability among our contacts and how they respond to different marketing vehicles. This chart was built by Greenplum and Tableau and represents the audience reached through six months of marketing programs initiated by a single EMC division.

Taking the variations into account, our data science team broke these companies out into value-based customer segments, increasing our precision by grouping them based on shared characteristics.

This calibrated our marketing mix to maximize effectiveness. Based on the unique characteristics of each population (in terms of decision-making, media mix, products purchased, purchase cycle and spending with EMC, as well as other demographics and firmographics), we can target different segments with marketing tailored to their specific needs.

For example, a segmentation scheme for one of the largest Asia Pacific countries enabled us to produce a campaign that targeted the segments most responsive to outbound telemarketing. This in-country telemarketing campaign is in flight now, and it has become the pilot for a large global campaign that we are developing and rolling out through field marketing.

Further modeling using association and time-series analysis produced a list of target companies with the five products that each company was most likely to purchase. We are using this approach for cross-sell and up-sell field marketing campaigns globally, based on customer and market potential.

The chart below shows market potential based on modeled propensity to buy for campaign targeting (colors represent market segments).


The nonlinear nature of the marketing funnel is even more evident when we drill down to the customer level. For each segment, we combine marketing, sales and firmographic data to create marketing “blueprints” representing a typical customer profile. These help us learn about how that customer behaves by graphically depicting all the contacts, job titles and marketing vehicles that the customer responds to during the course of a purchasing cycle.

Our next step will be using machine-learning techniques to identify patterns of collective contact behavior within a company so that we can predict when they will be most receptive to our communications. As we work towards a deeper understanding of the “always on” customer buying journey, we will blend in unstructured social data using Greenplum/Hadoop DCA for an even richer view of our customers.

About the Author: Mike Foley