I want to write about delegation. This is based on a podcast conversation, edited and expanded here.

In most startups and early-stage companies, the founder or CEO owns a wide range of responsibilities — which concentrates a lot of authority in one or two people. At 10X, I ran a single-top company from founding through early 2020. From around the summer and fall of 2020, I began progressively delegating. Now, roughly a year on, I feel we've moved into a new stage. I want to try to organize the model — the purpose, the principles, and how I think about it.

The Purpose of Delegation

I see two main purposes:

  1. Maximize the total volume of leadership exercised across the organization
  2. Create discontinuous opportunity for the person who delegates

For example, when I handed off product authority, it freed up time to search for new business opportunities — to explore, to look for ways to grow 10x or 100x. The person doing the delegating finds a trigger for their own discontinuous change.

Delegation is both a goal and a means for non-linear company growth. I can't think of a more effective tool.

Three Key Points in Delegation

SmartHR CEO Miyata-san wrote an excellent post in December 2019 called "The Technique of Delegating Authority." I read it and filed it away at the time. But after about a year of actually doing delegation at 10X, the model has become much clearer in practice.

Here are three things I focus on:

  1. Hire with clear expectations and hand off issues that are half a step ahead
  2. Onboard toward autonomy
  3. Enable back-to-back collaboration

1. Hire with Clear Expectations and Hand Off Issues Half a Step Ahead

First: the right goal framing and aligned strategy understanding are the foundation.

For example, with Stailer's business model: if someone asks "is maximizing GMV the ideal state?", I'd say "that matters, but it's not the whole picture." GMV growth alone doesn't guarantee that 10X's business is in its best state, or that partners and their users are in their best state. Even the most elegant KPI can't fully represent the ideal state in isolation.

This is why we need to place goals using precise language — language that doesn't just capture the metric but the intent. And beyond the goal, we need to align on how to pursue it — the approach, the thinking process, the decision-making protocol. Without a shared protocol, each person uses different "climbing routes" to reach the same summit, and the company can't leverage its accumulated assets or provide coherent support.

Once that foundation exists, the question of "hand off issues that are half a step ahead" becomes meaningful.

The right framing of "half a step ahead" is critical. If it's not clear where the boundary is — "how much can I decide on my own?" "how much freedom do I have?" — you end up in a system where the person being delegated to constantly checks in with management. That doesn't maximize anyone's leadership. What you need is: "here is the level of decision you own; here is where we remain back-to-back."

The right issue bracketing for the person's current potential and expressed value is what matters. This is something I've failed at in practice — I've handed off issues with poorly defined scope, and people struggled not because of inability but because of unclear boundaries.

Observe the person carefully. Understand where they are now. Define "what state do I want them to reach?" and "what does half a step ahead look like?" Then bracket the issue appropriately.

Also: communicate this approach during recruiting and onboarding. If expectations shift over time, realign explicitly.

As for how the person achieves the goal — that's for them to figure out, and neither they nor I fully know in advance. So encourage experimentation. Creating conditions where people can try many things is the delegator's responsibility. If they can't find opportunities to experiment, help them find or create those opportunities.

On the psychological side: if failure feels too scary, that's a problem the person above is responsible for. Unconditional encouragement of challenges and experiments is important. That said, not all challenges deserve celebration — challenges that deviate from the company's values should be corrected. At 10X, if someone took an approach like "just try all 100 ideas in order," that would be a values misalignment: "Work back from 10x" and "Be autonomous" imply that you should prioritize within that 100, not work through them serially.

2. Onboard Toward Autonomy

For someone to act autonomously, they need access to information without friction.

If there's a gap in the quality or quantity of information between me and the person I'm delegating to — if I'm getting inputs they're not — I can't reasonably expect them to make the same judgment calls. So creating openness — removing information barriers — is the delegating party's responsibility.

With an open information environment as the base, you can begin to align context. But information access alone doesn't create context alignment — context needs to be built deliberately.

The most important thing is "why" — why we're doing this, why this matters. So for any output, I try to reinforce consistently: "the why is most important." If someone can't articulate the why clearly, I help work through it — facilitating the wall-bouncing process until they can stand on their own.

When handing off an issue, be careful not to hand off tasks — handing off a task hands off a "how." An issue lives at the "why" layer. What you're handing off is a "why to bracket and own."

Most professionals are experts at tasks, at how, at what. Moving to the "why" layer is a discontinuous challenge. That's precisely why helping people raise their operational layer — from what to why — is one of the most important parts of onboarding toward autonomy.

3. Enable Back-to-Back

As delegation progresses, issues get handed off and outputs start coming back. Initially, you need to give direct feedback on those outputs.

Outputs that go out into the world become the company's face. The quality level required — what "good" looks like — needs to be aligned on. Direct feedback on outputs is effective for calibrating that.

But more important than the output is the process and protocol — how it was built. The delegator is ultimately responsible for building a system that can reliably produce adequate output. If that system doesn't exist, that's a delegation-side failure.

Once the system is in place, shift the focus of feedback from outputs to process and judgment. Then gradually withdraw from direct involvement in the team — letting the team own the quality calibration and decision-making fully.

The progression:

  1. Feedback on output → feedback on process
  2. Feedback on thinking and values/context
  3. Feedback to team management rather than directly to the individual

This sequence is the smoothest path toward real back-to-back collaboration: a team that can make decisions, set quality standards, and deliver — and that you can fully trust your back to.