I recall introducing pair programming to a team years ago. During the first round of pairing it just so happened the the most senior person on the team (20 years working on the same code base) got paired with the most junior (recent graduate from school on her first job). Mostly likely everyone assumed that the senior would be hand-holding the junior through the exercises, and yet when we got to the debrief, there was a surprise.
The senior couldn’t stop talking about how much he had learned from the junior, in that short time working together. He stressed that together they came up with a much better solution than either one of them could have done on their own.
While this is extreme example, this kind of result when we truly collaborate is typical. Together we come up with much better results than when we work independently, no matter the skill levels of the people involved.
So why don’t we collaborate more? I think there are several reasons.
1. Most people have never experienced real collaboration and they’re convinced that what they already know is as good as it gets. It’s not.
What most teams call collaboration is actually coordination. You divide the work up, each person goes off and does their piece, and then you review each other’s output. That’s not collaboration. Collaboration is working on the same problem together, at the same time, building on each other’s thinking in real time. It’s uncomfortable at first, especially if you’ve spent your career with your own task list and your own sense of ownership. But what I’ve seen again and again is that the discomfort fades quickly once people experience what’s actually possible.
2. We’ve been trained from childhood that individual accomplishment is the only valid way. All the way through school, it’s called cheating when you share your answers with others. And having been trained that way, it now feels uncomfortable to actually collaborate.
We spend 12 to 20 years having that framing reinforced, and by the time we enter the workforce it’s deeply embedded. Most corporate performance management systems reinforce it further: your review is about your contributions, not your team’s contributions. Even when the work itself calls for collaboration, the incentive structure is still measuring individuals. That tension doesn’t disappear just because you’ve decided to work together.
3. There is a shared psychological bias that believes that two people working together will take twice as long to get the work done as those same two people working independently. This is despite decades of research on pair programming that shows the opposite. The bottleneck has never been typing speed, it’s always been thinking speed and we think better when we collaborate.
The math seems obvious: two developers, one task, surely you’re cutting your throughput in half. But it’s wrong, because it ignores where the actual time goes. Laurie Williams’ research at the University of Utah in the late 1990s and the studies that followed consistently showed that pairs take roughly 15% longer to produce code, but that code has significantly fewer defects and requires less rework. When you factor in the bugs you don’t ship and the rearchitecting you don’t have to do six months from now, the economics look very different. The cost you’re seeing is real. The savings are invisible because they’re in futures that never happened.
So here’s my challenge to you: before you decide collaboration isn’t worth it, actually try it. Not someday. This week. Pick one piece of work and do it together: same screen, same problem, both people present and contributing. Not dividing it up, not reviewing each other’s output. Actually working on it simultaneously.
Then pay attention to what happens. Notice what comes up that wouldn’t have come up if you’d been working alone. Notice where the other person catches something you missed, or takes the idea somewhere you wouldn’t have gone. The experience itself is the evidence.
If you’re not sure how to get started with real collaboration, or you’ve tried it and it hasn’t clicked, let’s talk.
See also: