Why hardware is so hard

It takes a certain kind of crazy to get into building a company around hardware. To invent something from scratch that people find valuable is hard enough at the best of times, to do it by creating a physical manifestation of that is near impossible for many reasons. Then to find a way to distribute it, support it, update it…is an Everest style challenge without oxygen or winter clothing.

Hardware is hard.

Look around you right now and think about the hardware that you use regularly. It isn’t hard to point them out. You wear a watch (or smart watch), you carry a smart phone, you type on a computer, you talk to Alexa, you watch TV, you make coffee or cook your food. Hardware is everywhere and building it as a core to a business seems easy enough. Right?

So many things go into building the right hardware for the right reason. And that is the challenge to overcome before going down this road.

The last company I worked for had an incredible product. It was the company that I had wished I had started. It was a hardware company that focused on the gym market. The system was designed to lead and track a workout for a gym member. The product was actually multiple products including a wearable, multiple types of sensors, native mobile apps and a web component. The sensors were installed on each piece of gym equipment, the wearable was worn by the member of the gym and the apps pulled it all together in a programmed and directed workout.

This was a very complicated but beautiful challenge to solve and the company did just that but here’s why it was hard to build and, eventually what led to a product switch.

Challenge #1: Too many moving parts

The company was in 3 different businesses which often happens with hardware companies as they start. We were building hardware, software (mobile app) and content (workout routines). We had to build it this way to make it all work together. Until we had enough momentum and our partner channel was well stocked and we built our API on top of a reliable hardware stack, we needed to do it all ourselves.

Gym equipment is not easy to attach a sensor to either. There are so many different styles and types of machines — from free weights, to selectorized machines, to plate loaded — so there would never be one sensor, there always needed to be many.

The key was making all the hardware and software work together and this was done by adding beacons (yet another moving part) to each machine to help identify it as unique. We then had to build something that would allow us to register that beacon (yet another moving part) which ended up being a wearable in order to be the conduit that paired the app with the machine. Incidentally, for those asking in your head about timing, this was before the smart watches emerged. Being a little ahead of a curve is also a great risk with hardware.

Software played such an important part in what we did. We needed app developers to work on building native apps for iOS and Android. We needed data scientists in order to differentiate between a dumbbell arm curl and a tricep extension. We needed software engineers to build the operating system that connected the app to the machines and we needed it all to work without a hitch.

Counting reps

Challenge #2: Manufacturing is slow

It’s pretty easy to build a prototype with a 3D printer these days. We can jimmy up a quick test piece of hardware made of plastic and make it work. But once you are ready for full production for sale — or in our case installation on gym equipment — time is what kills you. Some of our installs were for 300+ pieces of equipment at a time, split between all those different styles mentioned before. Manufacturing took weeks, sometimes months depending on component availability. Things don’t happen quickly in hardware. The plan needs to incorporate some give in order to make sure the supply and production chains have adequate time for slippage. This will happen unless you control the chain or have the clout of Apple. Which you don’t so expect delays.

Challenge #3: Costs are higher than you think

It costs a lot to build a product. There are no easy workarounds. Committing to building a piece of hardware needs money and time. Double whatever money you think you need to build and, while you are at it, do the same for time. Even the best designed hardware on paper will run into adversity when you hit play on manufacturing. Very rarely (if ever) does the product leave the screen to be produced and distributed as it was originally designed or with the same component parts.

To test the product in the wild, companies will typically do low volume production runs — we did ours at a local manufacturer — and this increases cost significantly and is often not scalable at this cost. It’s necessary because in order to test it, you need the hardware. It could be an interative process — build, test, optimize, build, test, optimize. This has to be done until it is done right.

Challenge #4: Distribution

Our challenge with distribution was that it was a complicated installation. The decisions we made to have different sensors meant that it was not an easy thing to drop ship and have set up in a moment. We had magnetic coils that needed to be installed into selectorized machines, end caps for the barbells, RFID rings that needed to be glued onto dumbbells, bluetooth beacons that needed to be affixed in visible locations on each machine, wearable charging stations that needed to be prominent when members walked in…the list was endless. Distribution for new products is never going to be easy — especially something as far ahead of its time as this was. We knew that and it was part of the vision to make a new standard for gyms. The workout operating system.

Most companies suffer from distribution challenges. Today you can set up an online store and sell your product — creating a new challenge of rising above the noise of the crowd. It’s a different expertise needed to distribute an unknown product. It takes time, relationships, luck and a beautifully designed product that consumers or businesses understand immediately.

Challenge #5: Defects

Let’s face it, stuff breaks and when you are in the hardware business things are bound to stop working. There will be a small percentage of product that simply fails out of the box or something in the environment makes it stop working. Failure is inevitable.

One of our early designs for counting the amount of weight someone was lifting was also something that was a magnet for dust. Eventually enough dust gathered on the sensor and it stopped doing what it was supposed to do. Corrosion, humidity, the pounding of heavy weights all lead to potential defects. It could be something as simple as a battery slipping from its contact. We even had a percentage of our components that reached us already defective.

Challenge #6: Capital

Hardware is not software. When it comes to funding it always needs more. More time, more resources, more research and way more money than originally thought. That may seem simplistic and obvious but undercapitalizing a hardware company will seal its doom. Our company raised a seed round on the large size but it really needed 10X that amount to actually build what it set out to build and stay alive long enough to see it through.

When raising investment for a hardware business, raising too little or selling too big a vision is a non-starter. Sell the vision but insist on the right raise or the gap will be too wide.

Challenge #7: Knowing when to pivot

Sometimes hardware startups bite off more than they can chew and are facing one or many challenges on this list. It might be time to pivot the product in order to simplify and to live another day. This is a hard compromise to make but the alternative may be giving up altogether.

The key to a pivot in hardware is time. It can’t happen near the end of life for the business. There needs to be enough money in the bank that gives enough time to design, build, test and deploy the new product.

Our company had to pivot for all these reasons but we did it with enough runway to test the idea on customers, validate requirements and then move into production and deployment without raising another round. The next product was closer to commercialization and would start earning immediately upon launch. It wasn’t the one we had set out to create but it was on a path towards that vision.

Hardware is not for the feint of heart. Version 1 of any new technology will take longer and cost more than initially thought. The vision of the product may need to be recast many times during the early days in order to get it out the door. Then, slowly, it is refined release by release. Sometimes component parts get better, sometimes a different approach is discovered. Regardless, getting that first version out the door is crucial. That’s where the real work actually begins.