How many times has this happened to you? You have a conversation with people on your team about building something. You arrive at an agreement about that vision. Then they go off to build that thing. A few weeks later, they come back to unveil what you agreed on--except it looks absolutely nothing like you had discussed.

Your first impulse might be to think they weren't listening. But chances are, that's not the case. Most likely, they heard you loud and clear, but they were seeing something different. It's a symptom of a problem I've come to call the illusion of agreement.

Here's the problem: You have something in your head. I have something in my head. Both of us think it's the same thing because we're agreeing out loud, but inside--in our own minds--we're seeing different scenes. I can see only mine, and you can see only yours. This isn't a public stage--it's everyone's personal backstage. It's like we're each in our own dressing rooms looking in the mirror and thinking we're seeing what someone else is seeing as well.

I first recognized this problem early on, back in 2007. We were building Highrise, a small-business CRM. We split up into a couple of teams and started working on individual pieces of the product in isolation. The idea was our team would work on our part, and the other team would work on their part, and we would eventually meet in the middle.

But that would require each team knowing where the middle was, and what that middle would look like. Since the middle didn't exist yet, both sides had to imagine it. And that's where the problem began.

Rather than building toward a clear, shared understanding, the two teams built toward an illusion of agreement. Rather than looking at sketches together so both sides had common visual ground, we had access only to written descriptions of the direction. Words--even carefully written words--can be perceived in vastly different ways. Drawings are more definitive. It's why architects gather around looking at blueprints rather than reading descriptions of how a room should be designed.

So, as both sides plugged away, it seemed like progress was being made. But it was the wrong kind of progress. As the teams got closer and closer to each other, the two sides didn't match up. They were like pieces from different puzzles. What a lesson that was. We ended up having to scrap many months of work. Back to the drawing board we went. This time, we created sketches. And that allowed us to build toward the same goal.

It was a bit of a gut punch. We'd built our first product, Basecamp, by sketching out everything first. It worked great. But then we had the sophomore process slump, thinking we could do things differently this next time. We figured we could skip some steps, take things for granted, and end up with the same outcome. Turns out not. Making sure everyone could actually see what we were working on together was critical to the process. We haven't made that mistake again.

Now we actively work to avoid these situations. How? Whenever a few people are working on something together, we always try to look at something visual, even if it's rough. It doesn't need to be a fully fleshed-out design, but it at least needs to be a sketch. Something everyone can see, understand, and internalize. As Jonathan Ive, chief design officer at Apple, said in a recent interview: "Things are exceptionally fragile as an idea--entirely abstract--but once there is an object between us, it is galvanizing."

When you look at something real, everything gets much clearer very quickly. You can freethink, you can brainstorm, you can make decisions. When you finally shatter that illusion of agreement, you can arrive at an actual agreement.