Sunday, December 4, 2011

Duplication - A four part series

Part One: Intro and Visual Options
Whether you’re trying to calculate the number of shared customers between two stores, the number of readers between multiple publications, or customers using multiple products, duplication can have an extremely valuable impact on understanding market trends, threats, opportunities and an even bigger impact of strategy development. 
It can get real complicated fast, so I’ve decided to break this post into a series.  I’m an extremely visual analyst, so in Part One, I’m going to keep things basic and show a few options for demonstrating duplication in analyses.  In Part Two, I’ll break down how to calculate duplication given specific data sets and provide an introduction to its analytical utility.  In Part Three, I’ll give multiple examples of how to use this for insight and planning.  And in Part Four, I’ll give some trade secrets on calculating duplication amongst three and even four different sets of data.
Let’s start with a hypothetical market and competitive data set. 

In Mid Atlantic Metro we have 1,000,000 adults (18+).  In this market, we have to department stores:  Bull’s Eye and Wal-Co.  We also have access to information that tells us how many people bought from these stores in the last week:  430,000 from Bull’s Eye and 610,000 from Wal-Co.  Since 1,400,000 adults, calculated from adding Bull’s Eye and Wal-Co customers together (430,000 + 610,000), clearly exceeds the size of the market, there is obviously customers buying from both stores.  The question becomes, how many bought from both? (I’ll give greater depth calculations in Part two of this series.)  With the right data sources, we are usually able to decipher how many adults bought from both in the last week (typically referred to as “AND logic”) or we are able to extract how many bought from either of the two (typically referred to as “OR logic”).  To get us to the visual options let’s just say we have access to the number of customers that bought something from Bull’ Eye and Wal-Co in the last week:  260,000 adults (this represents our “duplication” number).
Here are a few options for visually demonstrating duplication:
  • Traditional Venn Diagram (Exhibit 1)
  • Proportionate Venn Diagram* (Exhibit 2)
  • Pie Chart (Exhibit 3)
  • Bar Chart (Exhibit 4)
  • Percentage Bar Chart (Exhibit 5)
*Note: I’m still working on calculations for the accurate proportionate area for the duplicated area between the circles.  So for now, it’s a best guess in distance apart from one another.

In Part Two, I’ll give more detail on the meanings of different numbers and how they can impact strategy development and additional analyses.