Web projects are similar to all other projects in that they involve the three basic parameters of project management- cost, schedule, and quality/specs.

A primary concern in any project is scope creep (aka feature creep), which is the tendency to continue to grow the project beyond the point of satisfying requirements. Another essential concern is to identify and clarify the requirements of the project before work begins. It's always cheaper to fix problems early (rather than later) in a project.

Web site projects are interesting because the product is aesthetic. You're not buying three pounds of iron, or anything that can be weighed and counted. Web sites have a subjective feel and a tone.

Web projects are non-linear projects, as described in the classic project management book The Mythical Man-Month. If you're working on a linear project like ditch digging, you know that x people can dig y feet of ditch in z hours. So if you're behind schedule, you can add people and time to the project and you can calculate exactly when you'll catch up and what the cost will be.

But in an intellectual project, where all the people are connected in a network containing 2n-1 nodes, adding people adds more connections to the project, and will actually hurt project performance.