Is Entrepreneur First Worth It? Part 2: Bad Ideas
Why my first startup ideas sucked and what I learned from them
After a few weeks at EF, I had developed a suit of armour.
Each week, I generated and tested a new idea by speaking with dozens of prospective customers. The level of action that EF expected from me resulted in a pace that I became addicted to. The momentum was exhilarating. EF forced me to conquer my fear of failure, because speaking with so many prospective customers required facing rejection on a daily basis. As it turns out, hearing “No,” or feeling stupid for a minute or two isn’t so bad.
Beside conquering my fear of failure, learning how to test ideas quickly was the most valuable skill that I learned at EF.
By the end of the program, I’d have a cheat sheet for evaluating ideas, consisting of the lessons learned from each of the ones that I had validated. The ideas that I tested and the lessons that I learned from each of them are the subject of this and the following post.
Welcome to Part 2 of my 4-part series about my experience at Entrepreneur First. If you’re new here, start with Part 1.
Idea #1: Automated Mortgage Document Processing
“Whatever you decide to work on is going to be hard. But it doesn’t have to be small.” - EF Staff Member
My cofounder and I were in our first 1:1 with the EF staff. The feedback that we’d received on the problem that we’d been working on for the past two weeks wasn’t good - the market was too small.
We wanted to improve the mortgage application experience by automating the document review process that lenders use to verify assets and income, which is manual and time-consuming. A number of startups are solving this problem for US mortgage lenders, but their solutions don’t work in Canada. This problem was “on-Edge” for me - I had an unfair advantage in solving it due to my expertise and network in the mortgage technology industry. I had found a customer for whom the problem was “hair-on-fire.” But unfortunately, there weren’t enough other customers like the one that I had found.
The Canadian mortgage industry is large enough to support a VC-backed startup, but the b2b market is too concentrated for one. There are roughly 9,000 banks in the United States. In Canada, there are 35 (and 6 of those 35 have 93% market share). So, gaining enough market share to build a VC-backed startup would require working with at least some of the 6 largest banks. And I know from having worked with them before that their sales cycles can take years. With open banking on the horizon and Plaid increasing its presence in Canada, it’s possible that an alternative to using documents to verify assets could arrive soon. The juice didn’t seem worth the squeeze.
Any b2b problem that I identified in the Canadian mortgage industry suffered from the same problem. I didn’t want to build a consumer mortgage business because I knew a number of well-funded, smart entrepreneurs already doing that in Canada. We could have gone looking for problems in the US mortgage industry, but I felt like I didn’t have the time to do that because my insights and network were unique to the Canadian industry.
It was time to pivot, or more accurately, to start from scratch.
Lesson #1: map out the Idea Maze
The Idea Maze has saved me more time than any of the other concepts that I learned at EF.
The Idea Maze is the collection of all of the paths that result in your idea succeeding. EF encourages you to visualize the successful execution of your idea at the end of the maze, and to imagine all of the paths - including the false starts and dead ends - that could come before it.
Once you’ve validated that the problem is in fact a real problem, you need to understand why it hasn’t been solved yet. If the problem is real and big enough, you are almost certainly not the first person to have identified it. So, if a solution to it doesn’t exist yet, there’s probably a good reason why it doesn’t exist. Your job is to discover what that reason is and to determine whether or not you can overcome it through your answers to “Why us?” and “Why now?”
The most efficient way to map out the idea maze is to speak with founders who have tried to solve the problem you’ve identified. You’d be surprised at how willing they are to share their experiences.
I didn’t map out the idea maze early enough with my document processing idea. If I had, I would have realized that all of the paths that could lead to the successful execution of my idea included working with most of the big 6 banks, which would be nearly impossible. The solution I had imagined didn’t exist because there wasn’t enough demand for it; the big 6 banks didn’t have a strong incentive to innovate due the lack of competition.
Idea #2: Product Management AI
For my second idea, I wouldn’t make the same mistake again. I would find a problem with a larger, less concentrated market.
After deciding to leave the Canadian mortgage industry behind, my co-founder and I oriented around our mutual experience in product development (his as a software engineer and mine as a head of product). I recalled a painful experience where I had missed a pivotal customer insight that was hidden in our trove of qualitative data. This insight would have changed the course of our product roadmap, but it was missed because I didn’t have enough time to analyze all of the available data. A quick search of my favorite Product Management Slack community revealed that I wasn’t alone - many heads of product struggled to aggregate and process the onslaught of data received from customers, customer-facing teams and analytics tools. We imagined a solution that would aggregate all of the data sources and use AI to automatically surface the most important insights to product management teams.
I spoke with heads of product at small 10-person startups, at scaleups like Hopin and Omnipresent and at public companies like Asana to understand who had this problem and what steps they had taken to solve it.
My conversations with heads of product confirmed my hypothesis that detecting the signal from the noise was mostly a problem for large, rapidly growing scaleups.
But there was a catch: well-funded competitors were emerging everywhere, and the incumbent product management system of record was about to launch a feature that could solve this problem. I wasn’t afraid of competition as long as we had a unique insight about how we could solve the problem better than our competitors. But we didn’t. Ultimately, this was an obvious problem that any product manager turned founder could see, and we didn’t have a unique perspective on the solution.
After two dead ends, our energy as a team was drained, and we couldn’t find another idea to get excited about. We decided to break up.
A note on breakups: EF goes to great lengths to normalize breakups, but nothing can prepare you for your first one. In our personal lives, when the going gets tough, we don’t give up on the relationship - we work through it. But at EF, because your opportunity cost is so high and your time is so limited, you’re encouraged to break up as soon as you start to doubt your team’s productivity.
Lesson #2: avoid the tar pit
The product management AI idea was a Tar Pit Idea.
Tar pit ideas are common ideas that attract a lot of founders.
Tar pit ideas are deceiving - they look like good ideas, but they’re not (just like tar pits look like freshwater ponds to thirsty animals but are actually sticky death traps). Tar pits tend to be great places to find fossilized remains (failed startups). Animals (founders) step in, get stuck, die, and decompose, attracting more animals who don’t see the tar pit’s victims hidden below the surface.
To paraphrase Dalton Caldwell and Michael Seibel, tar pit ideas tend to be consumer ideas that lots of founders try. They’re tar pits - and not just hard ideas - because they feel sexy, and you receive lots of encouragement to work on them, so they pull you in. They capture the imagination.
My idea wasn’t a consumer idea, but it felt like one amongst product leaders because everyone had experienced it (much in the same way that a consumer idea is relatable to most founders). Every product leader I spoke with had a visceral reaction to the problem, and of course, many product leaders turned founders were working on it!
The takeaway is to recognize the supply and demand dynamic underlying a given idea.
If the idea is obvious and sexy, there is probably a large supply of founders who want to work on it and who are qualified to work on it (in this case, every product leader turned founder). According to Dalton and Michael, tar pit ideas will have the largest oversupply of startup founders who want to work on them relative to market demand. To increase your odds of success as a startup founder, avoid the tar pit by looking for something for which there’s a low supply of founders and high market demand.
So where should you look? A good rule of thumb for finding ideas that satisfy the supply and demand criteria is to look for boring and complex problems.
“If you hang out in audit software, then you are in an area that’s complex, but no one wants to boast they do it at a dinner party. In boring and complex, the spiritual reward of the industry is lower, and so you just get fewer entrepreneurs. Therefore, the chance of succeeding in boring and complex is significantly higher.” - Charlie Songhurst
It was the end of October, and I was newly “single” as a cofounder. Reflecting on my first two ideas, I wanted to make sure that whatever I worked on next would have a global market (unlike my first idea) and be less obvious and ideally boring and complex (unlike my second idea).
I was approached by another single cohort member, an analytics and data science leader who had built the analytics teams from the ground up at multiple unicorn startups. He shared a surprising insight: data teams - the teams responsible for measurement - can’t measure the impact of their work. I was intrigued. The market seemed big enough (any tech company with an analytics function) and the problem seemed esoteric enough. We formed a team and got to work setting up interviews with analytics leaders.
That’s the end of Part 2. Tune in for Part 3 next week where I’ll cover the other ideas I worked on and the conclusion of my EF experience.