“Should I join a start-up?”
I’ve received this question a handful of times. While I don’t have a clear and direct answer, I’ve found the following metaphor to be helpful.
Imagine an unexplored piece of land. Possibly rich in resources. Maybe even treasure. Or it could be barren. There could be unknown dangers lurking. Or hostile forces that arrived earlier.
To assess the potential gains from the land, we need to establish an outpost. In come the commandos. Commandos are speedy, scrappy first-strikers. Need someone to set up a foothold in unknown/enemy territory? Call them in. They will set up camp and explore the area, reporting back on potential gains, dangers, and hostile forces.
If the commandos report the land is rich, we send in soldiers next. Soldiers are second-wave troops who expand your territory. They build on the know-how gained and maps drawn by the commandos. The goal? To quickly swarm to capture as much land as possible. They also start to reap the gains from the land.
After some time, there’s not much unoccupied land left. Hostile forces—hearing about your success—have moved in, trying to wrestle a piece of the pie. Your commandos and soldiers are now exploring new lands or combating enemies on the frontlines.
You need police to maintain peace and order over your vast territory. Police establish rules, patrol the area, and keep things in order. They maintain stability and efficiency. If small groups try to invade, they can defend and repel on their own, or at least minimize territory lost.
That piece of land above? It’s your next project/product/market.
There are many new lands to explore, and many lands in various stages of being explored and/or exploited. Each requires a different mix of commandos, soldiers, and police.
At the beginning, the potential gains are unknown and the risk is high. Commandos need to get from zero to one—build a working prototype and find product-market fit. Speed is critical.
Here’s how it could look like:
Imagine trying to introduce e-commerce in South-East Asia back in 2012. How would it work when there’s no widespread credit card usage, no reliable logistics, no proper home addresses? Should your tech be on-premise or in the cloud? What product categories should you focus on? Where to locate warehouses? Commandos tackle these questions.
As a commando, you get to build from the ground up, learn how to learn to figure things out, and move fast. It comes with high autonomy, iteration speed, and learning. Nonetheless, it’s high pressure, high risk, high uncertainty. The lifestyle is the harshest of the three roles.
If you enjoy this (and it’s hard to tell without actually experiencing it), perhaps you’re a commando. You can find commando roles in pre-Series A/Series A start-ups, innovation labs/skunkworks teams/new projects in established companies, etc.
Sometimes, the prototype finds product-market fit and gains a bit of traction. Soldiers then scale the prototype from one to 100. Speed is still important. Nonetheless, scalability and stability becomes more crucial as the product, market, and user base grows.
Here’s how it could look like:
You have a critical mass of buyers and sellers in one country. How would you replicate this in other countries? How about trying different business models (e.g., sell own inventory vs. marketplace)? How would you scale for holiday seasons and big sales days (e.g., 1111)? Soldiers solve these scalability questions.
As a soldier, you’re given a prototype and some paying customers. You need to build on it, refine and scale, ensure production stability, and attract mainstream users. You still have high autonomy, get to move fast, and do impactful work. There’s likely less pressure and uncertainty, and the lifestyle’s better.
If you don’t want the risk that comes with being a commando yet still want to have a big impact, then consider being a soldier. You can find soldier roles in hyper-growth startups (Series B/C/D) with fast-growing product and market, or a new project team that’s taking off.
After some time, you’ve cornered a piece of the market. The product is established, the team is established, the brand is established. Police maintain and grow market share. (The commandos and some early soldiers have left for new challenges). Convenience and reliability (and price) are now key to gaining mainstream users and earning their trust.
Here’s how it could look like:
A Chinese e-commerce/fin-tech giant acquires you. Resources are secure for the long term. There are many competitor platforms and switching costs are low. How do you maintain, and continue to grow, market share? How do you update legacy infrastructure and software in a stable manner? How do you reduce costs and keep them low?
As a police, you inherit infrastructure, processes, and code. You have to maintain, refactor, and make incremental improvements. Ensure everything runs smoothly and doesn’t fall apart. The work is low risk and stable. Of the three roles, the lifestyle is the best. Nonetheless, there’s significant bureaucracy slowing things down, leading to less learning and impact.
If your focus is on stability and lifestyle (e.g., you have family depending on you, you have low risk tolerance), then a police role suits you. You can find such roles in established companies (e.g., IBM), the government, or academia—their first and second wave troops have mostly moved on.
This metaphor also helps with other questions.
“I have two offers. One is a pre-Series A start-up, another from FAANGMULA. Which should I accept?”
“There’s a new team set up to work on project X. It seems like a long shot but the idea’s innovative and high potential—should I join them?”
There’s no one correct answer to those questions. Your answer also changes with your stage in life. Nonetheless, I’ve found thinking about Commando, Soldier, Police to help guide my decisions. I enjoy being in a commando-soldier role—innovating and building things from scratch and then scaling it for high impact.
Paul Jarvis, author of Company of One, also wrote about a similar paradigm: Starters and Operators. (Paul identifies with being a Starter). He shares his decision to sell a product as the support cost had gotten too high, and he didn’t want to manage a support team.
What other approaches have you found helpful in thinking over career choices? Reach out via Twitter (@eugeneyan) or share in the comments below!
Thanks to Xinyi Yang, Gabriel Chuan, and Marianne Tan for reading drafts of this.
I write about data science, machine learning, and career. Join 1,000+ readers. Weekly updates.
Welcome gift: 5-day email course on How to be an Effective Data Scientist 🚀