June 28, 2012
In a fascinating recent article in the Wall Street Journal, Dana Mattioli reported on how the travel website Orbitz recently began using “seemingly innocuous information,” like whether their customers are on a Mac or PC, to segment them based on their likelihood of booking more expensive hotels. Using powerful statistics and analytics, their analytics team identified patterns in the data that would have been simply too difficult to find with Excel and hard work.
Because Mac users were more likely to book nights at some of the more expensive hotels, Orbitz is experimenting with presenting the list of hotel choices in different ways. In some cases, the more expensive hotels didn’t appear on the PC list, or were shown further down the list. You could look at this as price optimization, or rather, a clever tactic to upgrade their product mix with customer segments where they think they can. It’s important to note that any user could simply “sort by price” when they wanted to, but this subtle change could mean that 1 or 2% higher realized price that can drive even larger gains in profits.
While the Orbitz example is a B2C situation, it illustrates the same concept that is applicable in B2B: different customers value your products & services differently. And if they value your products & services differently, then you should attempt to manage pricing differently as well. This is commonly referred to as price differentiation.
I know this works in B2B because at Vendavo, we’ve helped companies in some B2B sectors like chemicals, industrial manufacturing, and high-tech manufacturing do essentially the same thing. The first step is evaluating all the potential attributes for developing the segmentation scheme (e.g. whether the user is a Mac user or PC user… a customer attribute). Once you’ve developed segments based on price differentiation, then you can begin to truly optimize your pricing.
So I think the question we should be asking ourselves is: “What data do we have today, that may help in identifying pricing segments in our business?” Yes, you probably have a very basic type of segmentation which uses two or three attributes like: region, size, and industry. But do you really know empirically that these are valid and predictive of differentiation? And what attributes do you have available to you today that are also valid, but you’re not currently utilizing?
With so much potential profit improvement out there – and with your competitors reading the same articles you are – can you really afford not to investigate the possibilities?