Note: The implications of MakerFleet have changed since this was first written months ago. It’s now not only a tool for startups, but slowly becoming a tool for schools, architects, and much more. But I keep this for posterity. 

Making to Manufacturing

When I was in college, I spent every waking moment building things or learning how to build things.

It was ironic that I learned the most not in school, but waking up at 6:30AM to be the first one in the machine shop, and then camping outside of the Electrical Engineering shop waiting for the EE professor to drop by. I found a certain joy in taking an idea that was previously just a handful of my neurons and then transforming it into a physical product. For some reason, that was more interesting than studying for exams or doing problem sets.

A wooden word watch

Fresh out of college, I realized I needed a challenge. I had quickly gotten tired of designing one-offs. Designing one of something was fun, but I wanted to get into manufacturing and share one of my designs with the world. So I took one of the birthday presents I built for a friend, marketed it, and put it up on kickstarter and reddit.

I started the watch company for two main reasons
1. To learn how to start a company
2. To provide a unique gift to each person who believed in me and my idea

I figured that no one in their right mind would believe that a 22 year old fresh college graduate could pull off a ME and EE manufacturing design, so for those crazy people that did, I wanted to reward them. Heavily.

I decided to make each watch personal. I wrote code that personally inscribed each watch face with an infinite combination of choices. Every watch I sent out was unique for the user, from the unique wood grain to the actual customized engraving. But as I grew, I knew that the scalability of customization would destroy me, and after the initial batch of 300, I stopped offering the customization feature.

Four months later, I fulfilled 400 watches. It wasn’t easy, but it was one hell of a learning experience. Taking something from an idea to manufacturing by myself was probably the best learning experience I could ask for.

But after I finished my initial batch, two of the main reasons I started the company were no longer interesting. I had learned how to start a company and I could no longer provide unique gifts to each person.

Hence, I found myself in a fork on the road. Do I continue making watches or move onto something else?

I told myself I would only sell 150 more watches and then I would shut down the company to start the next thing. But what was the next thing?

The pressure got my brain racing again.

A Mission

I realized that my biggest motivating factor in life was helping people in obscure, indirect ways:

• I started learning engineering by building birthday presents for my friends.
• I took an old idea called Datamatch, which was Harvard’s annual Valentine’s day dating survey, and went door to door asking restaurants if they would sponsor dates between top matches, making Valentine’s Day something everyone looked forward to.
• I took some of the money from the Harvard Computer Society and created an initiative that funded students up to $300 to start their own personal projects.

So I spent months thinking “How can I help people?” I figured that the knowledge I had gained from manufacturing and starting my own company needed to be shared, but there were already hundreds of blogs out there. And instead of just offering “advice”, I wanted to play an active role in the success of my customers.

So instead of a blog, I needed to build a tool that actually brought value to my customers, and helped them take their dreams into the real world.

The Idea

There’s a certain joy in sharing something you’ve built and seeing how people interact with it. I still remember the euphoria and sheer happiness I felt after I received my first watch customer email:

Harnek!!!!

We received the watch in the mail. This is a pure work of art! We are SOOOO thrilled! You have created an absolutely beautiful design and we are so happy with it.

THANK YOU so much for working hard to get it here for Christmas.

Happy Holidays!

When I first got that email, I probably spent an hour laying in bed, smiling my ass off. Months of hard work and late nights, and someone actually liked my product! I have that email screenshotted and as a shortcut on my desktop.

I wanted other people to feel that. That feeling you get when you share your idea with the world and people actually like it. The best is when what you love becomes a sustainable business, but unfortunately that rarely happens, even less so with hardware companies. So I decided I was going to make it my mission to help other people get there; I was going to build a company that cared deeply about moving people to manufacturing.

But what do those companies need to succeed? We already have hundreds of CAD programs, thousands of manufacturers, and prototyping is easier than ever with 3D Printing. So why were more than 99% of hardware companies failing?

I had a bold hypothesis: They weren’t building the right product. 

We all live in a bubble where we’re the main protagonist. We all believe that we’re important and that our product is going to revolutionize how people think. But until it actually adds value to the customer, you’ve done nothing. You may have deluded yourself with one raving customer review, but any statistician will tell you that extrapolating a world impact from changing one person’s life is plain stupid.

So you iterate. You listen to that first customer, build what they want, and then sell to 25. You listen to the 25, build what they want, and then sell to 200. So on and so forth, until you’ve reached thousands of customers that love your product.

With software, iterations are cheap and fast. You can send your program or website to 25 people, get instant feedback on how they use it through the hundreds of monitoring apps, and then changing the program can be done overnight. A software company can be built and sold in a year.

With hardware, iterations are expensive and slow. You can send your prototype to one or two people, get feedback over a period of a month, and then changing the product requires conversations with the entire team, from product design to manufacturer. A hardware company takes years before its first product eventually releases.

But what was the largest pain point for hardware iterations? I broke down the process and analyzed the time and cost:

1. Customer Feedback Time: Getting valuable customer feedback time in hardware is usually around 1-2 weeks.
2. Design Time:  With CAD programs nowadays, you can change a design within an hour of getting feedback.
3. Prototyping Time:  A day to a week. 3D Printing and rapid prototyping has fixed this for one unit, but it’s not really scaleable for multiple units.
4. Manufacturing Time:  This usually takes a month to two months, and a lot of money.
5. Iterate: Back to step one for the next iteration.

Looking at this, it seemed obvious that if you could delay going to full-scale manufacturing until you absolutely needed to, you had a better chance of building the right product.

So you need something cheap and reasonably scalable.

You needed to prototype at scale. You needed to make 200 of something and send that to customers. You shouldn’t change your manufacturing methods until you absolutely needed to.

Then, you could get rid of Manufacturing Time and have people go through steps 1-3 at scale. Send iterations to customers, and then manufacture it.

Companies and Makers could iterate on the right product up until they reached a couple hundred customers, and then once they ironed out the issues with those customers, move to larger manufacturers. The early adopters of your company would be ecstatic that they had such an impactful voice, and you can move to the early majority confident you would succeed.

If only there were a way to prototype at scale…

The Online Factory

Let’s say we had a bunch of manufacturing equipment directly connected to the Internet, with no people in-between and just software. What would be that factory’s pros and cons?

+ Quick at-scale prototypes

+ Modular factories that could be placed all around the world and provide engineers with same day, at-scale prototypes

+ Allow people to build projects that require various manufacturing techniques without needing to pay a high capital cost

+ The software could organize the manufacturing process so small iterations could be easily and quickly done

– Software can’t replace experienced machinists

– Inexperienced people using the printers would get frustrated easily

– High capital cost for the business

– High mix manufacturing scales well with software, but not with people. One person changing between assembling hundreds of different prototypes would not work. So it would only work for up to 200-300 unit volumes, unless someone rented out the factory 24/7.

But if this factory did exist, it could help hardware companies build the right product before scaling. They could have 25-50 versions of their prototype in a week and obtain customer feedback faster than ever before. They probably wouldn’t use MakerFleet for large scale manufacturing, as a high mix system can’t scale well, but we could help companies find the right product faster.

The more I thought about it, the more I realized that this was how I could help makers and companies share their designs with the world. It needed to happen, and with my experience of IoT and manufacturing, I could build it. It was exactly the challenge I needed, with a mission I could get behind.

But like every investor will say to a 22 year old fresh out of college going into a field he doesn’t have a degree in:

Cool idea kid. Prove it.