Lessons from Y Combinator

At the end of 2013, I started SketchDeck with my co-founder David. We set out to solve the pain of creating presentations. A month later we were accepted to YC, and in January 2014 we were on a plane to Silicon Valley. In this post (a year after that flight) I share the lessons I found most valuable from this experience.

The post is biased towards b2b (all the startups I’ve worked on fall into this category), and apply to startups from conception to the early growth stage. My guess is much applies to b2c and slightly later stage startups too, but I don’t have the experience to confirm this.

1. Make something people want (or, if you don’t have users, you’re failing)

The day after we arrived we met with Kevin Hale, one of the YC partners. He remembered our application and asked how things had progressed. After five minutes of probing we admitted (to ourselves as much as him) we had no real users.

The first goal of a startup is to make something people want. If you don’t have users you’re failing.

We defended our position by saying we were still building what they want. We needed a few more weeks. This, Kevin explained, was a trap. We would likely spend months or years improving the product and people would continue to not use it. Without users, you’re shooting in the dark with what to build.

You have to get to something that at least one person is using as quickly as possible. You should get there in days or weeks, not months. Here’s how…

2. Concierge your product or service (or, do things that don’t scale)

Google, Facebook, Twitter are all built on incredibly scalable technology, so therefore shouldn’t startups do the same?

The short answer is no. At least, not at the beginning.

Your goal is to make something people want (the test being something at least one person is using). There are a couple of options to get there:

Option 1: Take a guess at the perfect product or service, figure out how to build it. build it, then find users that want it. This will likely take a few months (speaking from personal experience). When they don’t use it (they likely wont), ask them why not and iterate on the product (a few more months work when repeated).

Option 2: Intimately understand the problem you’re trying to solve (preferably one you have experienced yourself). Find someone who has this problem, and solve it for them manually. Figure out what they found most valuable and repeat (on them and another person).

Here’s an example to solidify these options. Lets say you are solving the problem of busy people spending too much time in their email inbox. Using the first approach you might build an email prioritisation plugin for gmail. Using the second approach, you find a busy person and convince them to give you access to their email inbox. For the next week, you sit in their inbox and keep it clean for them — trying different things each day and collecting their feedback.

Both options can work. However, they have significantly different implementation times. Option 1 takes weeks or months for each build, test and iterate loop. Option 2 takes hours or days. I’ve now tried option 1 twice, both failed (along with 8 months of “wasted” time). I used option 2 during YC and got to a viable problem-solution in a just few weeks.

This lesson doesn’t stop applying after you’ve found a viable problem-solution. It should continue to be use for new products, features and services as you grow.

3. Pick a metric and focus all your energy on making it grow

By now, you have made something people want (we got to this stage about a month into YC). So what is the goal of the startup now?

Growth.

First, how not to do this. You’ve learnt that people want an email auto-responder. You shut yourself away and build it. It takes you two months. You then take it back to your first user and try it out. He’s lost interest and doesn’t use it.

Here’s what you should do. Pick a metric which represents the success of your startup. For example, number of active users, revenue, events completed, etc. Do what ever you can do to get this number to grow fast. At first, you should be doubling or tripling this number each week. After a few weeks of hyper-growth things will settle down. Set yourself a target of around 5% to 10% growth per week. Now, do anything you can to hit this growth rate. If you hit it, you’ve succeeded. If you miss, you’ve failed. As simple as that.

By focusing your energy on growing one number, you will automatically optimise your time to doing what’s most valuable. You will scale what needs to be scaled, and leave what doesn’t. You will continue to concierge most of your product/service for months. This is fine, in fact it’s good (more on that in the next lesson).

4. Startups are scrappy and broken, don’t worry

Many startups look incredibly shiny from the outside. They have great looking websites, impressive founders, millions of funding. You therefore assume that they must be operating seamlessly on the inside. You look at your own startup and notice everything that isn’t working and feel disheartened.

Don’t worry.

All startups are broken (something you realise first hand when surrounded by 70 other startups during YC). What matters is that you’re growing and delighting your users. Things will get fixed as you grow.

Here are some examples of things that were broken for us until recently (lots of other things are still broken which I won’t share with you!):

  • We had no way of validating invoices from our designers — we just paid what they asked
  • All our orders were traced with pen and paper — we lost the paper on one occasion!
  • Many of the buttons in our app just fired an email at our inbox which we had to manually react to

The scrappiness has some unexpected benefits too. For example, you can change anything quickly!

5. Delight users

On the face of it, a startup is in an incredibly weak position compared to an established company. But there are advantages which you should exploit. One of these, is your ability to deliver exceptional customer service. The delight this creates can more than make up for defeciencies in your product or service.

Here are a few of the things we do:

  • Communicate with your users personally — users love emails from founders
  • If anything goes wrong, personally fix it for them
  • Call new users and ask if there’s anything you can do to help
  • Over deliver on promises

Delighting users should also become a guide for day to day startup life. Frequently ask yourself, will this delight our users? If it does, great. If it doesn’t, think carefully before doing it.

6. Ultimately, startup success comes down to the speed and quality of the execution

It’s easy to confuse a startup for a job, where showing up and making your boss happy is what counts. People go far in the corporate world with this approach!

A startup is different. If your product sucks and your users are unloved, your startup will fail regardless of how much time you spend in the office. If you do things too slowly, you won’t be able to retain talented employees, obtain the funding you need or beat off faster, better startups.

Good execution is hard work. By its nature, it’s hard to describe in words, but I’ve tried to capture a few ideas that may help:

  • Create an place where you can work effectively. Work hard when you’re working. Relax when you’re not
  • Spend the majority of your time on activities that grow your metric in the short/medium term1. Minimise everything else
  • When you first start, talk to users and build your product. Ignore everything else
  • Don’t build perfect features first time round, they will likely change
  • Have a clear idea of your goals, and the next activities needed to achieve them. Don’t spend time creating business plans, gantt charts or similar
  • Only build stuff you know users want (or you need to hit your growth target). If you’re unsure, ask them
  • Look after yourself: exercise, sleep and spend time with your friends and family

I hope you found this useful. Writing and re-reading it certainly has been useful for me. If you have any suggestions or comments, please get in touch.

Footnotes

1 Short / medium / long term optimisation deserves a post of its own at some point. I’m still figuring this balance out. My current philosophy for our startup stage, is heavily weight for the short term (next month). Keep an eye out for things that are going to significantly help or hurt you in the medium term (next 2–6 months). Ignore the long term (6 months +), except for motivational/investment/marketing purposes.