What I would tell myself in 2021

When I started OEG in Beirut in 2021, I spent most of my worrying on whether anyone would want it. That turned out to be the easy part. The things that actually cost me time were all things I had not thought about at all.

Sell the outcome, not the mechanism

For the first stretch I explained what OEG was: a way to hire vetted people from the region at a fraction of local cost. Accurate, and much less persuasive than I expected.

What people actually respond to is the outcome. Nobody wants to hire from a region. They want a role filled by someone good, without spending three months on it, without the compliance risk, and without discovering in month two that it was a mistake. The geography is how it works, not what it is for. I spent too long leading with the how.

The systems work is not a distraction from the real work

Early on, building the CRM properly felt like the thing I did instead of selling. It was procrastination with a spreadsheet, or so it felt.

It was not. Every week I ran the business out of my own head was a week I was making the eventual transition harder and more expensive. The pipeline eventually had to move into a system, and doing it at thirty deals would have taken a day. Doing it later, with history to migrate and habits to unwind, took considerably longer. What that system actually has to contain, I set out separately in the revenue system most startups skip.

The rule I would give myself: build the system one size before you need it, while it is still small enough to build quickly and while you still remember why every part of it exists. That rule is most of what changes on the way from startup to enterprise.

Hire slowly for judgement, fast for execution

I got this backwards at least once. I moved fast on a role that needed judgement, because the person was available and enthusiastic and I wanted the problem solved. And I deliberated for weeks on a role where the output was clearly definable and a competent person would have been productive in a fortnight.

The asymmetry is straightforward once you see it. A poor execution hire produces work you have to redo, which is annoying and bounded. A poor judgement hire makes decisions that shape the next year, and you often do not find out for months.

Three businesses is not three times one business

The thing nobody tells you about running several companies is that the cost is not linear, and it is not mainly time. It is context switching. Moving between a staffing problem and a video production problem and an edtech problem in one day means arriving at each one slightly worse than you would have been.

What makes it workable is not discipline or a better calendar. It is that each business has someone who owns it day to day, so I am the person they escalate to rather than the person doing it. Getting to that structure took longer than it should have, because handing over work you are good at feels like negligence right up until the moment it obviously is not.

Distribution is the constraint, almost always

Every business I have built has been limited by distribution rather than by the product. That is nearly universal and nearly always ignored, because building is more satisfying than selling and it feels more like progress.

If I started again I would spend the first month on how people will find out this exists, and only then on what exactly it is. Not because the product does not matter, but because a good product with no distribution is indistinguishable from no product, and the reverse at least generates feedback.

Everything I learned the expensive way in five years was available to learn cheaply. I just did not believe it applied to me yet.

The one I would keep

For all of that, the original instinct was right. There was a real gap between capable people in one place and companies that needed them in another, and it existed for no better reason than that nobody had built the bridge. Trusting an absurdity you can see with your own eyes, over the consensus that it must be that way for a reason, is the only part of 2021 I would not change.