My Corner of the Web

Contrarian Insight: You Don’t Need Soldiers in Your Startup


Today, I want to share a personal story. In the past I led an “activation team” that taught me the importance of scouts:

My “failure” as an activation team lead

During my time leading, we aimed to improve our conversion rate. Activation was the % of users that installed our mobile app and then funded their account.

We failed to move the needle over six months. But we learned a ton.
In hindsight, I had the wrong framing. Instead of “improve conversion,” it should have been “Why is conversion what it is? What are the problems with our current experience?”

We had too many soldiers and not enough scouts.

What do I mean by that? Keep reading.

High-growth startup leaders have a flawed default assumption: “I am right.”

Most startups, especially after Series A, make the fatal flaw of assuming they know what they’re doing. They think they need to deliver, not discover. That couldn’t be further from the truth.

In reality, their map of information has gaps. They should invest more in discovery and less in delivery.

Don’t believe me? Why else would 35% of startups fail after they reach Series A funding and before they can raise a Series B round? (source)

This arrogance is what happened to my activation team. We had money in the bank, so we assumed we could iterate outputs until we found the solution. It didn't work.

So what’s the best way to update our maps? We need scouts.

The Scout mindset values finding the truth and understanding, over defending existing beliefs.

Julia Galef popularized the concept in her book of the same name.
I recommend checking it out.

In the book, she contrasts scouts and soldiers.

The Soldiers’ job is to exploit existing information.

A soldier is good at focusing on a single outcome.

For example:

  • Mission: take the hill
  • They create a strategy to take the hill
  • If something goes wrong, iterate
  • Continue until the hill gets taken

Soldiers enjoy rapid iteration methodologies like the OODA Loop.

The Scouts’ job is to get new information.

A scout remains impartial toward the enemy. They must update the map with factual information.

For example:

  • Their mission: Where is the enemy?
  • They create a strategy to search in 6 different places
  • They learn that the enemy isn’t in any of the places
  • They update their map with the positions the enemy isn’t

Soldiers make terrible scouts.

Imagine if you gave the soldier the mission to find the enemy.

How would they feel after failing to find the enemy many times? Because of their soldier mindset, they’d likely feel discouraged.

After a few months of optimizing our onboarding process, my team felt exactly like that.

It’s your job as a leader to make sure your team adopts the right mindset is critical.

Mindset is the essential precondition to being an effective scout.

Next, we design experiments:

  1. Figure out what you already know
  2. Confirm how you know it
  3. Identify what you need to know
  4. Identify the best way to get that information

Spoiler: Rarely is building software the fastest way to build knowledge

You can bring in the soldiers once you’re confident you’re building the right thing.

I mean confidence in the face of evidence, not overconfidence.

Soldiers excel when the goals are clear and workable. Yet, many startups engage them far too early.

Based on my experience, it’s because Big Tech influences them. They read blog posts about how big teams do things and then adopt those practices.

Avoid the temptation to copy the big guys. Instead, focus on the fundamental principles and build your knowledge map.


  • The best learning processes use reinforcement learning (plan -> do -> reflect -> repeat)
  • Don’t mistake output for learning.
  • The job of a startup is to take large contrarian bets. You can either gamble or manage the risks. Either way, you have to do it fast.
  • Soldiers are for exploiting existing information. Risk is garbage in -> garbage out.
  • Scouts are for acquiring new information. Use them to create your advantage.
  • Soldiers make terrible scouts.
  • You don’t need soldiers if you’re a Series A startup.