Most companies try to segment their markets, and most do it badly. So many work for months to create a "universal" segmentation scheme, only to make things worse. Here is a better way to think about segmentation.
First segment customers differently than prospects. It makes sense to segment customers based on factors such as 1) total revenue for the account, 2) higher versus lower price plans, 4) products purchased, 4) potential revenue growth, and 5) length of time as a customer. These factors may reflect the different needs or value of customers. These types of segmentation can help the customer service team discriminate by the level of service -- phone versus email-only support, phone center hours, phone wait times, etc. Segmentation for prospects should help target the right messages at the right buyers, so separate prospects based on needs. Prospect segmentation can also help define which marketing channels to apply to which targets, which is more a measure of the revenue potential.
When thinking of needs-based segmentation, this article and video from Clay Christensen offers great advice. He calls his talk "Milkshake Marketing", and poses the question "for what task does the customer want to 'hire' the product or service"? This approach requires some fundamental thinking about the role of products, and goes way beyond traditional demographic analysis. It can lead to new insights about customer needs, and can target messages more accurately.
At the same time, marketeers must ensure their segmentation schemes are actionable. This may run counter to some of the Milkshake Marketing insights. But in the end segmentation needs to drive action programs, or else it is purely an intellectual exercise.