Building startups at EF6: the first half

EF office

Entrepreneur First (EF) is a startup accelerator with a twist. They recruit individuals instead of teams, purely based on talent. The programme takes 6 months, with a breakpoint at three months into the programme. The first part is focused on forming teams. The second one, provided that you make it (as a team now), is based on growth and getting you ready to raise some capital.

I came into the programme straight after finishing (while finishing) a freelance project. I can say I was unprepared. I did not have an idea that was well researched, just some vague interests. I thought I’m interested in healthcare, but quickly realised it would take me too much time to get up to speed in anything related to healthcare. Also kept two old ideas that never got properly put into practice, namely an app store for smart homes (was a hackathon project, lots of people influenced the idea) or buildings and an end-to-end travel booking system. I didn’t push them much at the beginning as most of the cohort was much better prepared in terms of ideas.

In the first days, my reaction was generally what am I doing here? These people are good at their things! I’m a programmer, but almost everyone else here can code, so there’s nothing unique that’s my thing. People here have Ph.D.s or at least master degrees or lots of industry experience from which to come up with ideas that solve real problems. And no, they’re not those kind of people who have degrees just for the paper — they built impressive projects.

At EF we describe what we’re really good at, what makes us special, as our edge. There are three types of edges:

  • Technical edge is when you know some tech really deeply. Have done a Ph.D. in it or at least a masters degree or have worked a significant amount of time in the area.
  • Problem edge is when you know lots about an industry, or problem and how to solve it. Generally, people who have worked in an industry for some time and have figured out a recurring problem that they can solve, and have contacts and domain knowledge to be able to quickly iterate and test solutions.
  • Smart Generalist is when you don’t have a deep tech edge or a problem edge. Usually full-stack programmers and enthusiastic business people who’d like to start a business but have no defensible idea. EF strongly recommends that generalists pair up with either problem or technical edge people as the skill sets are likely complimentary. Ideas formed by tech or problem edge people, based on their edges, are more likely to perform well compared to random ideas.

I joined as a generalist and thought it made sense to work with somebody with an interesting idea. Started working with Abhishek at the beginning of the programme. His idea was initially a music recommendation app based on context, and in the first day, we pivoted to a mobile context SDK that fuses data from the sensors in your smartphone to understand your high-level context. Exciting idea, productive team… We released the first version to a small number of developers, we were fixing stuff until 3 a.m. after someone reported an issue, exciting times.

Five weeks in, we realised our visions weren’t aligned. I see context-awareness in the way we were doing it as a nice to have whereas Abhishek sees it as a game-changer for mobile user experience. Decided it’s best to split up.

Confusion ahead. Didn’t know what to work on and I didn’t really find others’ ideas as something that I could work on for the next couple of years. Pitched my two old ideas about IoT and travel to other sole founders, had some conversations with a few people and tried to figure out what it is that I really want to work on and I’m good at.

The idea that started to stick is a communication or speech coach app that analyses your voice as you practice for or give presentations, sales calls, meetings, etc.

Had a co-founder for a week, and unfortunately, we weren’t very productive together so we decided to split. It was a hard moment, as I knew splitting would very likely mean missing the EF funding committee as a team, thus not getting funded for the next three months. I didn’t have a plan B if EF didn’t work out.

Productivity went down the next period. I wouldn’t wake up in the morning, I didn’t want to admit it but was quite stressed and didn’t get much done. Felt horrible for a couple of reasons:

  • All the sole founders at this stage seemed to find the speech coach idea interesting, but not enough to drop whatever they were experimenting with. I felt the same about their ideas — not only I didn’t think I would be very useful, but also didn’t find them as exciting. Besides, I didn’t want to drop the speech coach thing before even giving it a fair chance.
  • I felt like an impostor building my speech coach thing as I don’t have experience in the domain. I’m also an awful speaker. I can hack things up, but I felt that’s not defensible for the funding committee.
  • Was worried that the market for speech coaching apps wasn’t big enough, and didn’t have anything concrete to prove otherwise. Not even a prototype to go and do lots of serious customer development.
  • Also was worried about my own progress. Didn’t manage to get a demo up and running — that would’ve boosted morale.

Decided I need a cofounder. Spent some time confusingly looking for confounders. The clock was ticking to the funding committee presentation days. Two weeks left, then a bit more than a week left, then the schedule was presented. I wasn’t scheduled to present. Was called in for a chat, EF offered to help me find a job and make sure I’ll be alright. It was a nice gesture — I didn’t fully follow up with that.


My biggest mistake was not preparing for the programme. Three months to make a team and prove team productivity and feasibility of an idea seems a lot of time. It’s not. To really start working in a team it takes a week or two. That’s 16% of the time. I worked 41% of the time in my first team, and was confusingly looking for what to do for the next 16+%. Then the stress of not passing kicked in slowing me down even more.

This leads to the second biggest mistake: after splitting up, not fully focusing on one thing — in my case, building a demo. Stress and doubting the idea and my skills kept me busier than actually working on the idea. Strong beliefs, weekly held is something that the EF team suggests. Keep your beliefs all the way, stand by them strongly (strong beliefs). When you learn something new, or things change, adapt quickly (weakly held). Don’t waste time having doubts.


EF is a great accelerator and I highly recommend it to anyone who wants to be a founder. Also a bit of advice, do prepare before the programme and know what you’re there for. Know your edge, understand what you’re great at and what you’re not so great at.


What next?

I wrote the first draft of this post a month ago. Back then I had no idea what I’ll be doing. But now I sort of have figured it out. It’s very likely I’ll be starting a Ph.D. in machine learning in the fall, and I’m really excited about it. I am also busy for the summer, but that’s for a later post.