Starting from the questions the team answers again and again is the right test for a portal. Where is my order, can I get a copy of the invoice, how do I report a problem. If a proposed feature does not match one of those, it can wait.

Most portals I have seen fail the same way: version one tries to cover everything, customers try it once, cannot find the one thing they came for, and go back to email. The portal then becomes another place staff have to update.

At Ascendra Ventures we scope a first portal as one problem with a measurable result, usually fewer repeat emails or calls for three or four named tasks, and the acceptance cases are those tasks done by a real customer without help.

When you launch with a small group, what do you measure first: repeat logins, or the drop in the email questions the portal was meant to absorb?