Without a system for making decisions quickly, speed turns into guesswork, waiting, and small conflicts that eat up time we don't have.
I'm Ilija, a full-stack developer. My colleague Stevan, also full-stack, and I recently decided to build a startup: Kosmonaut, a platform and pass (similar to Fitpass) for coworking spaces. We quickly learned that if we wanted to be both fast and effective, we had to define that system.
Before Kosmonaut, we spent two years working together at the American startup EvolveWell. There, we had the chance, brief but important, to work with our former CTO, Marc Fletcher. He taught us the principles I'm about to walk you through. He was a great manager and an even better person. He worked with us for a short time, but he taught me a lot about communication, conflict management, and startup culture in general.
I try to apply the same approach in every organization I'm part of. Here are the three principles that impacted our speed the most: two-way and one-way decisions, DRI, and just do it. All three, at their core, come down to one thing: trust.
Two-way and one-way decisions
One of the most important concepts is the split between two-way and one-way decisions.
A two-way decision is one you can easily reverse. Like changing a button's color. Anyone can just do that on their own, no review, no meeting, no five people in a thread.
A one-way decision is one that's hard or expensive to reverse. That needs revision and review from other people on the team. A good example is pricing. If you lower or raise the price, that's not "let's just change the color back tomorrow." It affects users, partners, expectations, positioning. You don't go in blind there.
This concept matters even more in the AI era. As Stevan likes to say: we are just one prompt away from implementing this feature.
That's why we treat around 70% of our PRs as two-way decisions. It's more of a gut feel than exact math, but that instinct matters. Most of what we ship can be fixed or removed. And that's exactly what lets us move fast and ship new features almost every day.
If you treat everything as one-way, you're definitely moving slowly. If you treat everything as two-way, there's a real chance you'll make a mistake that costs you. The real point is trust — you trust people to tell the difference themselves, instead of slowing them down by reviewing every little thing.
Directly Responsible Individual (DRI)
The second concept: Directly Responsible Individual, or DRI. One person is directly responsible for a decision.
For example, I'm responsible for marketing. I make silly, goofy videos to promote the startup and myself. Most of the time I fail. But I like to think of it as part of the learning process.
Sometimes I ask Stevan for feedback before I post. And believe me, Stevan loves giving feedback. Sometimes unfiltered. Sometimes it even hurts my feelings, sometimes he flat out offends me. What can you do, people make mistakes, and that's okay.
Here's the important part: sometimes I disagree with his feedback. But since I'm the DRI, I get the final say on whether the video goes out or not, and whether anything changes. Input is welcome. Veto isn't.
This principle runs entirely on trust. I trust Stevan in his areas. He trusts me in mine. He trusts that even when he disagrees with my decision, I did the research and gave it my best shot. The same goes the other way. I trust him completely.
That's how we avoid getting into heated fights over small stuff and burning hours we don't have.
We also have a rule: feedback and discussion shouldn't take more than 10 minutes. If we notice a thread starting to heat up, we stop writing. We hop on a call instead. Text tends to escalate. Voice and context calm things down.
Just do it
The third concept I simply call just do it. Like Nike says. Just do the thing. Don't wait for someone else to do it for you.
A teammate recently felt like we were working from home too much and it was pulling us apart as a team. At a big corporation, that would mean scheduling a meeting to debate why it's happening, with a pile of "etc, etc" explanations. Instead, the teammate just found an open space, asked everyone when they were free, and booked an in-person work session. Simple, easy, fast. The point isn't to think about how and why, and especially not to wait for someone else to implement your idea — the point is to just do the thing.
It doesn't have to be perfect, and in most cases it won't be. But looking at it objectively, fixing at least one mistake next time, and improving the process along the way — that's what actually gets us to success.
At Kosmonaut we have the freedom to do most things on our own (as long as we're not putting the company's public image at risk). If I want to try something new, I make a plan and execute it. If Stevan is thinking about a feature, he writes it and submits a PR.
And this, again, comes back to trust. I fully trust the people I work with. I trust they'll make a good decision and do what's best for the company or the community.
Waiting is usually more expensive than being wrong about something you can easily fix, or something you're not even sure will have any impact at all, but that could genuinely make a huge difference.
Thank you for reading this far.
If you find this useful, let me know what you think. What are your core principles as a team? What helps you move fast without burning each other out?