In the Enterprise Portals Report I outline three different ways to segment the market:

• By scenarios or target portal type (e.g. web development or collaboration portal)
• By scope and complexity (enterprise vs. departmental)
• By tier (infrastructure, specialized, open source)

In particular I've advocated using scenarios as the most meaningful approach to find out which product is the best for you.

A recent workshop alerted me to a fourth approach, which struck me as a useful way of describing how many enterprises perceive the market today. Let's call it segmentation by complete vendor offerings. What happens is that enterprises do not always look at the 15 major portal products in isolation. Rather they frequently compare the entire Microsoft stack (or BEA, IBM, SAP for that matter) and consider how the vendor can help them across the board. While SharePoint may be weak in integration and e-commerce, to be fair Microsoft has other non-portal offerings that can help you in these areas. But on the whole, the weakness of this approach is that you may end up with an ill-fitting portal solution in your quest for a single infrastructure provider.

Nevertheless, while I still favor scenario-based analysis, just remember that this often leads directly to a difficult discussion of the potential need for multiple portal products. My advice: you may indeed require multiple products, but try to start only with one and then move forward keeping an open mind.