Getting Inside the Enemy's OODA Loop

So you want to join a Startup?

When I talk to college students looking forward to graduation and thinking about joining startups, I always ask the same question - why you and why startups - and invariably get the same response: “I’m good at figuring things out, getting really close to the problem and iterating on a prototype. I manage risk and ambiguity well, like independence, and hate boredom” so on and so forth. In conversations oriented around “what next,” a laundry list of desirables is great but hard to work with. Luckily, these students aren’t really saying they value 15 different things, but really one thing. They want to live for the OODA loop.

Many things in life are OODA loops - in a way, you yourself are at a very low level. Show me an agent in a partially observable Markov decision process and I’ll show you an OODA loop. What differentiates people - and organizations - is the speed at which they loop.

For most people and organizations, observation, orientation, and decision are rare, haphazard, and painful. Many jobs entail an action set that can fit in a 3 paragraph job description. Plenty of corporations run the founder’s playbook and maybe pay McKinsey for a few units of decision-making here or there. Most people inherit their major life choices from mass media (or, in the past, from mentors and relatives). A single OODA loop might last decades.

Startups are different. Their environment is *very* partially-observable. Any company’s early growth is highly path-dependent - which customer segment do you target, how do you sequence product development, which roles do you hire for when and whom do you take? For tech startups, add market and product risk to the equation and suddenly you know so little that acting for very long without OODing becomes a death sentence. Observing without ODAing becomes myopia, Orienting without DAOing becomes sophistry, Deciding without AOOing becomes folly. Embodying the tao of OODA is table stakes for survival.

Good startups, then, are differentiated by the quality and speed at which they complete loops. Both matter, but mathematically speaking it’s clear there’s interchangeability. If your loop reveals twice the information mine does but is twice as slow, who eats whose lunch begins to depend on model details. Moreover, vertical context and competitive landscape matters - competitive SaaS might need speed where moonshot hard tech demands quality. Tech rivalries, like dogfights, are won when one contender gets inside of the other’s OODA loop.

How does a startup loop? The founders probably split up the lion’s share of observation and decision amongst themselves - almost every CEO is dedicating a major share of their blood glucose to customer and stakeholder communication + resource allocation. But orientation and action are a team sport. Working at a startup is like pulling a bandit lever and getting a nugget of never-before-seen information, every minute, over and over. If you can’t answer the questions “what does this mean to me” and “what does this mean to everyone else” well and fast, well, the second O in your team’s OODA loop just got that much worse. And, consequently, so did the D, A, and first O. Uh-oh.

If employee information processing limits loop quality, employee action time limits loop speed. Ok, everything has come together and you know what the best next step is. Now the easy part - doing it. Can you do it fast? Like, really fast? If you can’t, the team’s OOD apparatus is going to have to stall under-utilitized for a two week sprint before anyone can move on. Uh-oh.

So you’ve got to be good at orientation and action, clearly. But actually that’s not quite enough. Any founding team that’s moving as fast as possible almost by definition will need to delegate some decision-making and observation to key team members - proof by parallel processing. This means that you’ll be running your own OODA loop under the hood, and will be responsible for orienting it with the group’s and those of your teammates in one hell of a mean field game. The quality and speed of your actions and orientations will begin to depend on your decisions and observations.

What founders mean when they list “self-directed” in the job description is they need to hire an OODA loop. Still excited to join a startup? Great. How operational are your O, O, D, A, and what are you doing about it?