Three types of work

Solution-certain, solution-uncertain, and problem-uncertain work

There are 3 types of work — solution-certain, solution-uncertain, and problem-uncertain.

In startups, a lot of the work you do as a founder or early employee is in those last two buckets. When people ask me for career advice I often tell them that a great reason to join something early-stage is precisely because you get to do work that isn’t solution-certain. Let me explain with some examples.

Solution-certain work is implementing well-understood solutions. If you join a big tech company like Google or Facebook, most projects have this structure. More senior people will have already done the work of outlining the solution in detail and your goal is to implement it. While you'll have the opportunity to refine your craft, you are robbed of the opportunity to explore the full solution space. For example, you might be tasked to design or implement a new page in an onboarding flow using existing components.

Solution-uncertain work happens when the problem is clear and you have to figure out the best solution. This work is more difficult, more creative, and requires more persistence because there are more unknowns. You probably won't get work like this at big tech companies until you've been there for many years, but you might get it at some late-stage startups. An example of a project like this is to improve customer onboarding. The solution can take many shapes — make it faster, make it simpler, provide more guidance, etc. You have to figure out which path to take to best solve a given problem.

Problem-uncertain work happens when the problem itself is uncertain and your objective is to first define it rigorously and ensure you're asking the right questions. Since perfectly executing a project that solves a non-existent problem is worse than making progress towards solving an important problem, this is the most important and highest-leverage work. Continuing with the onboarding example, perhaps the state of the onboarding experience is indeed poor, but poor onboarding is not the best articulation of the root problem. Instead, maybe the root problem should be reframed as bad retention, and you should therefore consider spending your time on other initiatives like acquiring different users in the first place, improving your positioning, or changing your pricing.

If you want to start a startup, get comfortable with work that requires you to think about problems rigorously and be open to solutions that take many shapes. Indeed, one of the first decisions you'll make when you start a company — what idea to work on — is the perfect example of a situation where understanding and defining the problem really matters. The reason most startup teams fail is not because their solution is sub-par but because they're solving an insufficiently important and misunderstood problem.