“You don’t sell a product or service—you sell value.” Salespeople often hear that piece of advice, and for good reason: The value that each customer sees in a product or service is different—and so is his or her willingness to pay for it. Segmentation reveals how much various types of customers (eg, by geography, sales channel, etc.) will pay for your offering.
Segmentation plays an important role in pricing: The obvious advantage is the ability to analyze how various segments fare when pitted against each other. You could also set prices and business constraints by segment.
But the real benefit comes when your salesforce makes deals with customers: You can use segmentation in real time to suggest what discounts to offer or what products to cross-sell. In effect, you use history to illuminate how your salespeople can navigate a deal to win it.
When we talk about pricing, we need to look at transaction segmentation, and that means reviewing your invoices. Based on this segmentation, you can determine which segments are profitable and which are not. Also, within each segment, you can see the transactions that outperform the mean score, and the transactions that lag and pull the whole segment down.
This knowledge can help you make your deals smarter and your salesforce more intelligent, because they have data science backing the discounts or prices that they offer. How do you do it?
First and foremost, you need a rock-solid yet easy-to-use infrastructure to capture as much information as possible from your invoices. If your business isn’t capturing that data now, it’s time to start. This is imperative because invoices are a goldmine of information.
Next, you need to sift through the transactions. A good place to start is breaking down the transactions by the various levels of product, customer, sales and geographic hierarchies. All the parameters may not apply; choose yours based on the nature of your business.
Before proceeding with the segmentation process, decide on the parameters that would measure the maturity of the segments that you’re going to create. Examples include:
- Number of transactions per segment
- Number of customers per segment
- Standard deviation and residual variance of a price point per segment, etc.
This is a very important step: These parameters will help you decide whether you are getting the true picture of the segment or whether that picture is skewed.
Segments are created by combining the different levels across the various hierarchies; this is an iterative process, which will eventually lead you to a segmentation model. The segmentation model is a collection of different segments within the definition of how a segment is defined. This model will give you a fair idea of how your transactions stack up.
Each segment now represents a group of transactions that pertain to a level in the product hierarchy, customer hierarchy, distribution channel, geographic hierarchy, etc (based on a combination of one or many of these attributes).
If your deal does not fall into any existing segment, go to the next closest segment. This may happen if you are catering to a new customer segment, or selling a new product or selling in a new geography. So, the closest segment would be the one which satisfies most of the parameters in your deal.
Every segmentation model has a default segment that contains all outliers; this segment should contain as few transactions as possible. If there are many transactions falling into this exception segment, there is an issue with your segmentation model, or your model is outdated and it’s time to create a new one.