In January of 2015, I was sitting in my room, bored out of my mind. The semester hadn’t started yet, and I had nothing to do. So, for some reason, I decided to do something silly: make a banana costume.

I went to the fabric store and bought 3 yards of yellow felt. (Yes, I know felt is not a good material for clothes.) I spent the next few weeks cutting and sewing and gluing until I had a wearable banana costume. Once it was done, I wore it around my apartment complex and had fun talking to people.

I then took pictures in front of various landmarks in Boston and posted the pictures on Facebook. My friends thought it was funny so they shared the pictures with their friends who also thought it was funny so they shared the pictures with their friends who also thought it was funny so they shared the pictures with their friends who also thought it was … you get the idea.

The next day, I woke up to hundreds of messages asking where they could buy a banana costume. That’s when I realized that there might be an opportunity here: bring joy to millions of people through costumes that look like food.

I am starting a company called Banana Industries and I will be manufacturing banana costumes. The banana costume is a one-piece jumpsuit resembling a banana that covers the entire body from the ankles to below the neck. The costume zips in the back, has yellow sleeves, and contains arm holes.

I came up with this idea while on a trip to London, England with my parents last month. While walking around Piccadilly Circus we stopped in a police gift shop. There were T-shirts that said “Scotland Yard,” mugs with the British Royal Guard on them, and other souvenirs. On the second shelf I saw something that made me laugh: a bright yellow, one-piece banana costume with an opening at the top for your face. I walked over to my parents to show them, and they laughed as well.

We continued our walk through London but I kept thinking about the banana costume. Then it hit me: Banana Industries should manufacture these costumes! I have always dreamed of starting my own company and here was my opportunity.

I’m a programmer, and I’ve worked at several startups. I like to write code and build things. In recent years, my day job has been with a company that makes custom-printed t-shirts. We make our living by (among other things) taking orders from customers on the web, printing t-shirts, and coordinating shipping.

A few months ago I started thinking about creating a banana costume for myself. It was a very simple idea: just print an image of a banana on a yellow shirt, wear it over my head, and cut out holes for my eyes and mouth.

It would be funny. And if it was funny enough, I might be able to sell them to people who thought it was funny too.

In 2011, I was a student at Virginia Tech studying Software Engineering. At the time, one of my favorite things to do was to make friends laugh. In fact, I would often waste hours in Photoshop just playing around and trying to come up with funny graphics that would make people laugh.

One day, I was laying in bed thinking about how much I love bananas and how funny it would be if my friends were to wear banana costumes during Halloween. Then, out of nowhere, an idea popped into my head: “What if I made a banana costume for myself and wore it to class?”

I thought this would be incredibly funny and immediately began working on a design for the costume. The next day, when I showed my roommate Drew (who is now Co-Founder) what I had been working on, he told me that he would wear the banana costume if I could get 100 people to like the Facebook Page that we setup together.

So, we setup a Facebook Page called Bananadon’s Banana Costumes and began sharing it with all our friends. Amazingly enough, 100 people liked the page in less than 48 hours after we created it. Drew kept true to his word and wore the costume to class that next week while I finished making mine

When I was in college, I had an idea for a new type of costume. It was an inflatable banana suit with a fake pair of legs sticking out from the bottom. I made a prototype and wore it to a party. People loved it and word spread quickly.

The next morning, my phone started buzzing. I had a bunch of texts and missed calls from people who saw me at the party and wanted one too. Some kids on my floor were even willing to pay me $40/each to make them a suit.

I didn’t think much of it at first. After all, just because people like the idea doesn’t mean they would actually buy one. But after telling everyone who asked that they could get their own costumes, I decided to give it a try.

My friend Luke and I launched BananaStandCostumes.com where we sold these inflatable banana costumes for $79 each. We made our first sale on Halloween night and eventually sold over $100k worth of suits our first fall season in business.

It’s hard to start a company in an area where you’re not an expert. I’ve been lucky: I learned about starting companies by working for startups, and I’ve had a chance to try my hand at product development, marketing, sales, customer support, and raising capital. All of those skills were absolutely necessary to build out the business since it was just me doing everything.

Still, I was in uncharted territory trying to build a full-time business on top of something random like banana costumes. No one starts a company because they think it will be easy. But there are plenty of good reasons not to start a company. The biggest is that it might not work out, and you could end up losing money or wasting time or missing out on opportunities.

If you’re going to make a bet on the idea that your startup might succeed, you should at least know why the odds are against most startups. In fact, we can look at the history of Y Combinator itself for some clues about why startups fail.

My story begins in the summer of 2013. I was a rising Junior at Dartmouth, interning at a startup in San Francisco called Thalmic Labs. They were building an armband that let you control technology with gestures. It was an exciting time to be there – they had just secured $14.5 million dollars in funding.

I wanted my own armband, but I did not have $150 to buy one (they are now called the Myo). Luckily, my cofounder and best friend from Dartmouth, Lee, had a 3D printer. Lee printed out a version of the Arduino-powered Thalmic armband that had been designed by someone on Thingiverse. We soldered some components together and it worked!

We thought other people might want their own armbands as well. So we created a website to sell them for $30 each (we later raised the price to $40). At first, our customers were mostly friends and friends of friends; we only sold around 20 or 30 armbands in total. But then we ran out of bands!