YC Startup School - Week 2

Startup School 2018 continues into week 2. This week we are building product and dreaming of product market fit.


This week's lectures:

Building Product - Michael Seibel

This was a fantastic talk. Damian Canetti-Rios from GrantDocs has turned it into a set of questions for you to answer for your start-up.


Michael discusses how having his ego invested in Twitch was a positive factor. This was a pleasant change from the common advice to do the opposite - its not your baby, don't over-identify with it, maintain a healthy distance etc. which I think falls into the "nice idea but fraudulent" category of advice. I don't believe the people saying it follow it themselves.


I was troubled by some of the discussion of "don't be an artist". Art is being used as a code word for irrelevant but if you look at successful artists, you will find single-founders with technical chops. Each work is a product and they are the salesman. When successful, they build a studio, start a movement, make a ton of money and earn a place in history.

You can read more here A rebuttal of don't be an artist.

Good customers

To find good customers, Michael asks:

Whose business is going to go out of business without using you?

This seems to exclude almost all successful businesses. Even if a product is critical to collecting revenue e.g. Stripe, vanishes one day, so what? There will be an awful few days migrating over to a competitor but any well run company is not going out of business. Even Apple sells commodity products with a competitor in every vertical.

The only type of business it really applies to are those with very unhealthy relationships with the customers e.g. platforms that are middlemen between businesses and their customers. I would be gravely concerned about being so vulnerable to the whims of another business.

Sean Ellis has an attempt to define product market fit as "40% of your customers would be very disappointed if you vanished". Fairly arbitrary but its a bar Apple passes so maybe more useful!

Bad customers

This has given me pause for thought - bad customers are those that are not suffering the problem you are solving:

It's almost never the user of a product that comes into YC. And so ignore your investors, ignore your friends. They will lead you 100% astray. Out of good intentions, they'll try to be helpful. Well like, "I've never lived in the Sahara. I've never been thirsty. But maybe it should would like this," right? But it's horrible. Run away. Run away. Once you start having customers, I think it's a very helpful exercise to try early to identify bad customers.

Its interesting because they may be lovely people saying kind things about the product but you can't listen to them. I have been having discussions with my beta-sign-ups. When asked "What problems do you want to solve?", they fall into two categories:

The first category produce a vast list of decision making issues they have faced including issues with communication and collaboration. These are the people in pain who I can help. They are often senior developers, team leads and project managers.

The second category are attracted by its appearance, think it looks cool and intrigued by what it can do but not really sure what they would use it for. I should be cautious. They are saying nice things, but they don't seem to be suffering. Could this class of user lead me astray? Possibly. I suspect another factor is that I have not done a good enough job explaining the problem I want to solve. Its possible that this category can be moved to the first.

Product Management

This was a good discussion for how teams can plan features including an appeal to write specifications:

We would have a product meeting, and we wouldn't write anything down, right? You're an idiot if you can't remember a conversation four people had, right? No.

This is a problem Thorny helps solve. If you write a specification, you are capturing the outcomes of decisions. You are writing the "what". It's a lot harder to capture the "why" because it means explaining why you are not doing any of the alternative solutions. This is time consuming to capture in prose and hard to maintain as your thinking evolves so we rarely bother.

I believe our reasoning is more valuable than the actual decision. If facts change, the old solution has to be discarded but most of the reasoning still applies. It might only be a small tweak to your reasoning that can take you to the next solution. Thorny makes it fast to capture rationale while you explore ideas, helps you make the initial decision and then makes it fast to revisit that decision in light of new information e.g. the inevitable implementation hurdles and new requirements you discover.

Charging money

I really needed the discussion of never-discounting and overcoming the resistance to charging money:

The way not to do it is, "I'm afraid no one's going to use and so I have to not charge any money."

It's making me think I should bite the bullet early and only have a paid beta. It would ensure that I am listening to the right customers. David Rusenko in the next talk also touches on this:

I would say initially, don't worry about monetization either. You want to focus on making sure that the product experiences right. Don't shy away from monetizing your products. Don't shy away from charging for them. But initially, you need to make sure that your product is working for people.

So maybe not. If a paid beta reduces what I can learn then that would be a bad idea.

Finding Product Market Fit - David Rusenko

David references Clay Christensen's Jobs to Be Done and the Hubspot growth framework. He also advised against spending your time writing blog posts which makes this article a little embarrassing!

Optimise to learn

This is fascinating - at an early stage, your goal is to find discontinuous improvement. You should not chase incremental improvement but explore unknowns that might make big leaps. This means you might actively choose against building a better product in favour of a product that teaches you the most.

Renewal rate vs. Churn

This was interesting:

I don't like the churn metric for looking at this. A lot of people use that because it's easier to calculate. But basically, churn is not cohort based renewals cohort based. So churn is just looking at number of subscribers lost divided by total active subscribers in any given period. I don't like that because if your denominator is changing, let's say you're growing really, really quickly, your dominant is changing faster than your numerator. The numerator is based on last year's denominator. So, basically, it could be deceptive as to what your actual churn is. But your renewal rates are great, because that's looking at the percentage of people who are eligible to renew, and what percentage of those people actually renewed, and that's cohort based. So, I like renewal rates a lot better.

A Conversation with Ooshma Garg

Possibly the best name ever. It reminds me of the word delicious in Japanese - ooishi. I think the 'sh' sound appears in the word delicious in many languages because it happens when you salivate so it makes total sense she runs the meal-kit startup Gobble!

The interview covers a few topics that come up frequently e.g. issues facing non-technical and solo founders. Its an uplifting chat with someone gritty but I can't say that I understand meal-kit start-ups. Y'all need to learn to cook. Can they survive an economic downturn?



The forum has been running some popular Ask Me Anything's from YC alumni. Topics have included hardware, growth, games and Darby from Clerky has patiently handled a 600 post thread on legal issues.

Landing pages

A big discussion on the forums was landing page critiques. I had some nice feedback and it emphasises a few things that have been mentioned before:


Around 1000 of us have added ourselves to a map created by Stephen Bussard.

Its only a sample of the total SUS audience but a few numbers caught my eye include that Lagos must be heaving with start-ups!

  • US - 300

  • India - 127

  • Nigeria - 58

  • UK - 46

  • China - 40

  • Holland - 20 (more than most European countries)

Office hours

This week our 5 minutes were spent on round-robin updates. There was one no show. You can miss Office Hours just once before you are kicked out so fingers crossed they can make it back next week.

The group is remarkable - there is no "gee, I don't know where the time went" I would hear in a corporate stand-up, instead everyone is showing progress. This is one of the reasons to join Startup School - its a forcing function. The prospect of telling someone you haven't achieved anything is enough to jump a psychological hurdle and do something you normally wouldn't.

From quite a shy development process, I have sneaked a little mention of Thorny into a Hacker News post. Despite being quite hidden it attracted a few hundred hits and 40 sign-ups to the beta which was a pleasant surprise. I individually emailed every sign-up to learn a little bit more about them and what they are looking for and boy, that takes a lot longer than I would expect.

The backgrounds of the people signing up are interesting and from all around the world and includes developers, project-managers and interestingly, biologists.

One thing that stands out is that the people who have signed up are friendly and interesting. This is perhaps the most encouraging thing of all. I imagine there are many business stuck between middlemen where they don't really like their partners be they banks or estate-agents. That is going to be depressing. Its a different ballgame when your audience are people you are naturally inclined to help. I already feel indebted to them and hope to give them a fantastic tool that materially improves their lives.