How to fail at starting a startup

8 min readJul 7, 2021

Second long-form update. Start at When To Jump off a $500 Million Rocket Ship.

One of the fantastic things about starting a startup in 2021 is the incredible abundance of startup resources freely available. Y Combinator and Startup School, vocal and successful entrepreneurs on the Twitterverse and Medium, and VCs like First Round regularly sharing their own insights from their portfolios have made it such that you could easily spend your 9–5 day just consuming content. It might even feel right to dive head first into Zero to One, The Lean Startup, or whatever Naval Ravikant might be promoting that week.

“I’m equipping myself with a mental model of how to succeed so I don’t make the same mistakes!” I probably said to myself.

All of these resources are great — fantastic even. But I think it’s too easy to fall into the trap of trying to form some grand unified theory of startup formation, product development, or growth. Unlike math or physics, operating a startup is more like the product of every school subject combined, and then some. In science, the larger the system we try to describe, the more we tend to teach trends, observations, and generalizations, so that people in these careers can spend most of their lives finding the caveats and the exceptions. Whereas a single particle’s motion could be precisely described in a vacuum, we struggle to simulate how proteins fold, and we vaguely interpret how our entire body responds to stress.

I find it more helpful to view navigating the startup landscape like running a government rather than modeling a particle or protein. Initial conditions aren’t clearly defined and we don’t consider spherical cows; Everyone has opinions, no one is ever right, and what happens in practice is almost always disappointing. Moreso, building a successful company from scratch operates on such a long time horizon that it’s really only something that most people can go through in its entirety once in a lifetime. Maybe someone will make it twice if they’re lucky, and three times if they count their elementary school lemonade stand operation.

When I think of general startup advice as primarily based on small sample sizes and with ill-defined initial conditions, 95% of it is irrelevant. And of the relevant 5%, I find that the most useful exercise is to do as scientists do: examine them for caveats and exceptions. “Why isn’t this the case for what I’m working on?” (Oh the irony of this advice itself!) This exercise has often been more revealing than the advice itself when hardening my understanding of something.

All this is to say: focus on the work. Stop trawling Twitter, scrolling Hackernews, and reading about how we need to get to Internet 3.0 or 4.0 or 42.0. Instead when you focus on the work, and you focus on a problem, you can take to the great resources the web has compiled, match advice as best you can to your system’s conditions, and leverage what you find as inspiration to work hard in creating your best solution. With that, I’m telling myself to just start, then focus. I hope I’ll get somewhere.

I’m trying to time-box my writing (of which I’ve already significantly exceeded!) — so I’m forcing myself to put down my keyboard for now and the remainder of my thoughts will be a big brain dump of thoughts I’d like to remember from the past 3 months (AKA not advice, but take this as inspiration… if it is inspiring). If any of it’s extra interesting, or I run into the same things repeatedly in the future, maybe I’ll come back and write some more. I think of this as serving more as a time capsule of where I was at the time (July 7th, 2021) than anything else. As of now, Passionfruit is long dead, and Starview is on the chopping block.

Context

Solo founder + first venture, full-time, technical background
~1 year of runway
Operating for ~3 months, no real product + users, still largely operating in the idea generation + validation phase
2 years of work experience at a SaaS startup (Vanta) going from < 10 employees — 80 employees
Graduated from Brown University
Worst outcome is to live with my parents, and find some job at a FANG/equivalent company

Building or getting to a working product sucks for early learning, but is worthwhile to feel excited.

Only build when that’s what’s required to validate a hypothesis. After working on Passionfruit and Starview in a pretty significant capacity, I think that in all cases, building a working product is something that has not helped me validate my most critical hypotheses in any real capacity. Oftentimes, I think that I need to build a working product in order to get my foot in the door, but I should try to opt for easier ways to get there i.e. ask to interview them for some piece of writing.

Building has been largely a part of a selfish innate engineering desire to be able to touch and feel something for myself. This has been helpful though to have clearer thoughts around the product, and what a long term vision could be. Perhaps most importantly though, building has been helpful for getting me excited. On one hand, I feel like I could have learned all I learned without having built a working version of Starview, but I also might not have stuck with it for as long without getting to play with something that I loved every day.

With these in mind, I’d like to be more intentional about why I’m building. Build to keep my energy alive, but don’t substitute it for the best path for me to maximize my learnings in the shortest amount of time. (If I need the energy to even keep going with something, then certainly building is still preferable to an otherwise less expensive non-building alternative.)

What’s worth building during idea validation?

I’m finding that what’s worth building is to build just enough so that someone can understand when they would use your product and for what purpose. This could just be a sentence like: “Pay per view for Netflix shows,” or some clear UI/UX mockups for something more unfamiliar. The gut check I’m going to be working with is:

Are the people (assuming they’re the right people to talk to) you’re talking to able to easily answer: “What would you do with this?” and “When?”

How do I know if I’m on track?

At this stage when I don’t have invested users, simple gut checks have turned out to be fairly high signal. What’s helpful for me is keeping something that’s easy for me to assess quickly and clearly. More specifically, I’m operating on two “sheesh” metrics (call them whatever you want, but “sheesh” is the best thing that’s come out of me “researching” TikTok):

Am I “sheeshing” when I work on this idea each day?
How often are the people I share this with “sheeshing”?

When I’ve found myself in a new sheeshing position, or when I’ve lost that position, I’ve historically been at some significant inflection point worth starting some serious reflection.

Grit, focusing on the work, and focusing on focus

Each time that I’ve found myself facing some significant lack of response from my peers or my cold outreach, or I was just feeling like I wasn’t moving as quickly as I’d wanted, man has it sucked. Not only because I wasn’t getting the results I’d wanted, but because I’d want to continue doing that work even less. The more I got caught up in “why don’t people respond” — the more I’d tend to try to shy away from reaching out more, and think about what else I should be doing. This usually is good — reflect on what’s not working so you can iterate. The problem arises when mental wavering and low morale gets in the way of getting the work that needs to be done, done. When approaching cold outreach, I have to internalize that it’s a numbers game. When I suck at Figma, I need to roll with the punches. When I’ve made a well-reasoned plan for what I need to do, and how much effort it will take to pay off, I need to have the grit to believe in the plan and execute. If there isn’t significant new information that clearly changes how my plan will play out, stay focused on the work. Stay focused on focus. Here’s one checklist I’ve put together to help:

Do better work checklist:

Do I need internet?
Do I need to be at home?
Do I need my computer?
Do I need my phone?

Hodgepodge of things I want to keep written down and aren’t clearly contextualized. Blabla grains of salt.

Be disciplined and rigorous about walking through the foundations of what I’m putting together. Make the 1 pager. Write down a few critical hypotheses and then get going talking to folks.
Really cheaply go through and interview people who I suspect might have the problem that I am interested in — when they’re people I know. Just DM them quickly!
When evaluating ideas, another point to evaluate is how quickly can I validate hypotheses around it? What’s my access to the people who have these problems?
I’m pretty freaking slow at testing and validating hypotheses so far. I should weigh my ability to validate hypotheses/access to the right audience significantly. Access to this = distribution = feedback = progress.
Hypothesis based approach is really powerful! Stay focused on this.
Go through a socratic method on WHY I think things should be the way they are, or will get me to where I need to go. Think about why not and be my own devil’s advocate.
Sitting and thinking a lot has usually been my best friend. It’s hard to feel like there’s a right time to do this though meaningfully. You have to gather critical thought mass, where you think something is working well or not working well to make the most of this. I think a regular check in where you do some light version of this, that may reveal that I should do a heavier version of this — where I’m naturally more unsure, and naturally having more thoughts running wild — is reasonable.
Know that focus and capacity for “hard work” are seriously limited.
Don’t underestimate how hard it is to deliver something valuable. Even incremental value requires work. It doesn’t mean it has to take a lot of time though — just don’t think it’s easy.
Twitter is poo. Spending time here is not worthwhile.
Iterate on a 1 week time scale. Believe in myself and believe in the work. Thrashing around too much on a shorter than 1 week time scale makes for not a lot of work getting done so that there’s a very low probability that any of it will be meaningful.

Memory garden

Hodgepodge of documents that I’ve been using

1 pager template: Write this up when I’ve got an idea I’m excited about
Problems/ideas list
Hypotheses sheet for Starview, some Passionfruit stuff mixed in there

A few thousand words

Something something follow me on Twitter for increasingly infrequent updates, or here on Medium, for rarer but perhaps more thought through updates in my journey!