In 2004, 37signals, the software company I co-founded, released a Web-based project-management and collaboration tool called Basecamp. At the time, we mostly did Web design; Basecamp was a side project that we developed in our spare time to make it easier for us to work together.

Back then, project-management software was mostly about charts, graphs, statistics, and one-way broadcasts. Basecamp was different. It provides team members with a consistent place to work on projects and tools to swap ideas, share feedback, make revisions, and deliver the final project online. Millions of people across nearly every industry have used Basecamp to manage more than eight million projects; 96 percent of users say they would recommend the software to others.

That can mean only one thing: It's time to start over.

Why mess with something that has proved so successful? There are a couple of reasons. For one, eight years is a long time. Consider the ways in which the world has changed over the past eight years. We've learned a lot about collaborating in that time. We've received tons of feedback from users, many of whom have shown us the ways in which they work. Plus, there are technologies available that didn't exist back then.

But that's only part of it. About a year ago, we began discussing how we might improve our best-selling product. The more we talked, the more it became clear that the only way to significantly improve Basecamp was to start over.

Think about a product's life span. When something new is released to the public—and this is especially true of software—it's hardly set in stone. You get feedback from customers and make modifications. You add features, refine existing ones, and make things better over time. If you really listen and do it right, the product earns its success.

But paradoxically, that success makes it harder to change. As time goes by, people get used to things the way they are. And the more someone is accustomed to doing something a certain way, the harder it is to ask him or her to change. When it comes to introducing ideas, the years have a way of boxing you in.

And that's where we found ourselves with Basecamp—a successful product that was tough to change in major ways. Of course, it has evolved; over the years, we've made thousands of incremental improvements to the software. But now we have ideas that are more revolutionary than incremental. We think these ideas will dramatically enhance Basecamp's speed, power, and flexibility.

The problem is that we cannot make these kinds of changes in the existing product. Over time, software builds up legacy. The old technology is baked in, and the roots of the product are so knotted that simply unwinding them becomes a massive undertaking. Think about trying to uproot a 250-year-old oak tree versus a two-year-old one.

The easy thing to do is nothing. But continuing on the current path is a time-tested formula for complacency.

Of course, customers have a way of building up legacy, too, and there's bound to be some grumbling. We'll deal with any such issues as they arise. But one thing is certain: Starting over doesn't have to mean forcing change on existing customers. We'll have two versions of Basecamp—the Classic version and the new version. Users will be able to switch to the new Basecamp or stick with the Basecamp they are already comfortable with.

After a year of hard work, this is all set to happen soon. How will our customers receive it? In an upcoming column, I'll let you know.