Using Cluster Analysis to Identify and Convert

Using Cluster Analysis to Identify and Convert
Adobe Products Featured

I can’t wait for this year’s sold out Adobe Summit. A horde of marketers and analysts will gather in Salt Lake City for one of the most fun and stimulating weeks of my year. If you’ll be there, let’s be sure to fully engage and take advantage of this opportunity for industry growth. If you can’t be there, be sure to follow all of the great content and coverage coming from the event. We won’t keep all the insights to ourselves.

I’m particularly looking forward to a session provocatively titled “CRO on Steroids.” My colleagues, Kendra Jenkins, Senior Consultant, Adobe Target; and Trevor Paulsen, Product Manager, Adobe; will co-present four strategies to help organizations prioritize visitors based on likelihood to convert. Year round, I work on optimizing the user journey with Adobe Marketing Cloud, and I’m excited to hear these experts address personalization and content marketing targeted to high-value visitors.

Kendra and Trevor will show us the following tips and tricks:

  1. Using cluster analysis to identify relationships and define attributes of visitor groups
  2. Automated Personalization: Using machine-learning to discover profitable segments and prioritize targeting activities based on insights
  3. Explaining how predictive modeling can determine the likelihood of conversion
  4. Using profile look-a-like modeling to expand the reach of identified segments

In anticipation of this session—and all the great conversations Summit promises to spark—I want to get us warmed up with a few thoughts on Tip #1: Using Cluster Analysis.

Clustering Is Segmentation on Steroids

The term clustering is often mistakenly used interchangeably with segmentation. While this essential step to effective targeting and personalization is very similar to segmentation, it has some important distinctions.

Segmentation, at its most basic level, is the practice of dividing your users, customers, or subscribers into groups of individuals based on similarities that may be relevant to your marketing. Brands commonly segment their audiences by age, gender, industry, geolocation, and so on, and then target each segment with unique site content, emails and promotions.

Clustering is sort of like advanced segmentation; it takes advantage of Big Data and powerful analytics to let algorithms — rather than the marketers — define highly refined segments. Clustering makes it possible to easily create segments based on multiple complex variables. It’s possible to separate individual visitors by 10, 20, or 30 customer dimensions, taking into account demographic data, location, past and real-time behavior, and more.

Clustering Helps Brands Find Their Most Valuable Users

Creating the most useful and meaningful clusters for your organization doesn’t happen all at once. An Adobe Marketing Cloud analytics white paper, “Big data, big opportunity for wireless operators,” shows how brands may start with “segmentation based on readily available primary data” to gain initial, broad-strokes insights into various user groups. McKinsey & Company performed an analysis of Western Europe mobile consumers and first separated users according to voice and data usage and spending patterns. They then used demographic, technographic, and life-event data to reveal unique user personas, which they even gave descriptive names, like “Traditionalists,” “Practicals” and “Omnivores.”

The goal of rich cluster analytics is not only to gain a clear, accurate view of your audience, but also to enable marketers to predict user behavior and make their data and insights more actionable. Customer-engagement efforts are much more effective when marketers can anticipate what visitors are searching for, where they want to go next, what problems they are trying to solve, and what promotions will be of value to them.

By clustering, McKinsey & Company discovered that:

Although certain groups might have been dominant in size … data usage and hence, value to the operator, did not align directly along these lines. For example, one group, the “Omnivores” was the smallest group of mobile and smartphone users but utilized 80% of data services.

This was a crucial insight because it revealed that mobile operators who aren’t segmenting according to data usage will fail to optimize their services and price plans appropriately, and thus, effectively target their most valuable user base.

Clustering Is a Cornerstone of Cutting-Edge Marketing Efforts

Cluster Analysis can lead to the identification of valuable sub-segments that you previously didn’t even know to look for and engage. With advances in customer data collection and management, along with immediate access to customer profiles and behavioral data, adapting content and experiences in real-time is within reach for organizations across industries.

Sophisticated tools and platforms, such as the Adobe Marketing Cloud ecosystem of analytics capabilities, are making once impossibly complex and time-consuming marketing tasks a fluid, real-time reality. With Adobe Analyticsadvanced clustering, brands can analyze many variables simultaneously to quickly turn visitors into actionable audiences for further analysis, targeting, and personalization.

Recommended Articles