Founders of hardware startups quickly get tired of hearing the phrase “hardware is hard”, especially when voiced from someone that has never worked for a hardware company and has very little appreciation for what the phrase actually means. That’s the purpose of this article. I’m not trying to scare off anyone that just started, or is thinking about starting, a hardware startup. But I do very much recommend they go into the venture with eyes wide open.
About two-thirds of my professional career as an operator was spent at hardware companies. Many of the gray hairs on my head and battle scars on my body come from mistakes we made that are unique to hardware companies. I’m hoping to save you some of that pain.
Two Companies in One
The first reason it’s hard to build a successful hardware startup is because it requires building two successful companies: a hardware company and a software company. That’s because practically every hardware solution also includes software. Sometimes the software is only firmware that’s embedded inside the hardware, but more often it’s also application software that users directly interact with.
This means that, in addition to the various trials and tribulations I’ll describe that are hardware-specific, hardware startups enjoy all of the trials and tribulations that software companies experience. Yikes! But, as you’ll learn later in the article, the hardware-specific aspect of the solution can carry a broad range of complexities and implications.
Defensibility
I’ll start with some positive news. There is so much work and risk involved in designing, developing, and delivering a successful hardware solution, that very few entrepreneurs would dare go even think about going down that path. It’s so much easier to write a few thousand lines of code and throw an MVP version out into the market to see if anyone is interested.
This gives you some defensibility. Surely, you will encounter competition. But you won’t need to worry about a couple of recent college graduates spending a few months to copy what you’ve done or quickly try to make it better.
The Concept of an MVP Isn’t Very “Minimal”
It’s almost always recommended to first produce a minimum viable product (MVP), because it helps provide some validation of desirability (the world suffers from the stated problem enough that they’ll give your solution a try) and feasibility (the proposed technical solution can be built).
In the hardware world, an MVP is usually referred to as a prototype and there are usually several iterations along the way to a production version that’s ready for sale. The early prototype versions are usually, ugly, clunky, buggy, costly and much larger than the ultimate desired solution. But they operate. The problem is that the concept of “minimal” isn’t near as extreme as what’s often possible for software companies. This means a lot of work goes into just producing a prototype.
Time Distortion
Everything related to the product development cycle takes a LOT longer. Software startups operate in bug fix cycles of hours or days, and feature release cycles of weeks or months. Hardware startups operate in cycles of months and quarters, respectively. Just think about that dramatic difference.
Shortcuts Rarely Allowed
Mostly because of the slow cycle times, taking shortcuts that don’t work out yields extreme pain. An undo and redo follow the same month/quarter cycle time. This means that hardware startups have to measure twice and cut once on all sorts of decisions. You would never hear a hardware startup founder say “I don’t know, let’s just give it a try and see what happens” for any decision outside of the lab.
This factor also means that hardware startups quickly learn they are taking on big risks if their first electrical and mechanical engineers don’t have much experience. They are more prone to make mistakes due to inexperience which, in turn, are very impactful.
Costs
Software startups can bootstrap all the way to a product launch, if they’re really scrappy and creative. That’s virtually impossible for a hardware startup. Designing, testing, and prototyping various iterations of the product requires considerable spending on specialty equipment and hardware components. Service providers for regulatory certifications and legal compliance also cost quite a bit of money. Hardware startups usually have to raise a minimum of $1-2M just to reach a product launch, and $3-5M in required funding is not uncommon.
Getting through the product launch doesn’t mean the end of excessive cash consumption. Hardware startups have cash tied up in raw materials (ie – components) and inventory that don’t yield revenue for weeks or months until manufacturing runs are made – and those manufacturing runs cost serious money too.
Fundraising
Angel investors and venture funds that are interested, and willing, to make investments in hardware companies represent a small minority. It’s already hard enough to raise funding (check out my bestselling book on fundraising here) for any startup, but when the pool of investors is dramatically smaller, it makes it that much more difficult. That, combined with the longer period of needed funding before reaching the point of revenue traction, causes an excessively high rate of hardware startup death in the early stages of evolution.
Once hardware startups reach the point of making production runs, they need a different source of funding for materials, manufacturing, and inventory. That’s because selling equity for those purposes is a very expensive source of funding. Lines of credit and similar debt financing options are better for this.
[But wait, there’s more!!!]
Now let’s look at the unique issues and requirements hardware companies face as a result of having a physical product. If that product is not fully-designed and developed by the company, but rather components and subsystems are sourced from others and finally assembled by the company, some of these issues and requirements are less impactful.
Supply Chain Management
Much like a full-stack developer or software architect is the first most significant hire for a software company, so is the supply chain professional for a hardware company. They are worth their weight in gold. The more experience they have, the more mistakes they’ve made and are likely to avoid making while working for your company.
Don’t underestimate the value of this role. Of course, they are responsible for inventory management, component forecasting, shipping and the like. But they also serve as a critical check-and-balance for the hardware engineering and product design teams. That’s because the supply chain professional cares about manufacturability, product cost, and packaging.
Left unchecked, the hardware engineer might just design things in a way that’s easiest and most logical to them. If your seasoned supply chain professional eliminates two prototype iterations per year and one wasted injection mold every two years, you’ll cover half of their salary just in avoided costs.
Multiple Design Requirements
Hardware companies have to design for more than just user ergonomics. They have to design in such a way that the product can reasonable (easily is even better) be manufactured. If a mold is required (ie – for injection-molded plastic parts), the design will need to take that into very close consideration. Molds can easily cost tens of thousands of dollars and only certain types of design changes can be accommodated without having to pay for a whole new mold. And once orders start to scale, they need to design for volume manufacturing. That’s a special skill that many EE’s and ME’s don’t have.
Manufacturing
Virtually no hardware company manufactures their product in-house. This means they’ll need to select a manufacturing partner, which also means selecting a country to manufacture in. Manufacturing in the home country simplifies quite a few things, such as payment currency, language, and shipping logistics (no customs and important needed). It also enables easy visits to the manufacturing location if there are issues that need to be investigated. But usually there are other countries that can manufacture the product for much, much cheaper. Big trade-offs are involved with this decision.
Testing, Certifications and Compliance
This is one of the categories that gets completely ignored in the early days, but becomes so critically important in order to actually ship the hardware product to customers. Testing and certifications related to safety and country-specific regulatory compliance consumes a lot of time and requires specialty service providers. And the more countries where the product will be sold, the longer the certification list and the higher the certification cost. The bigger bummer is that with many product revisions, the testing and certification needs to be repeated.
If you think I’m kidding about all the different types of certifications that can be required, just look at the collection of logos below.
Warranty
Most hardware products come with some warranty, but the length of that warranty carries huge implications. Maintaining compatible spare parts is just one issue to consider. The logistics process for handling returns (called reverse logistics) is a specialty trade that very few know well. For example:
- How much troubleshooting is involved before authorizing a return?
- Who should pay for the return shipping?
- Does the returned product need to be inspected first to confirm a failure before a replacement product is shipped to the customer or is a “spare-in-the-air” policy more appropriate?
- Can customers pay money to extend the length of the warranty and/or get an expedited replacement?
- What will you do with the returned products? Whether you decide to repair them or not, they probably can’t be re-sold as new.
Labeling
I predict you’ll be amazed at how much time you spend on two different types of labels. One is associated with regulatory-related requirements for labels that must be affixed to the product. That doesn’t involve design work but rather placement decisions. The other is branded labeling for the outer package, which involves quite a bit of design work if the packaging will be visible to the end consumer.
Packaging
Like with labeling, I predict you’ll amazed at how much time you spend on the physical dimensions and design of the product’s packaging. The objective is to be efficient with space utilization while also having a positive unpacking experience for the customer.
The other consideration with packaging has to do with shipping and storing multiple packages inside an outer box. Especially if you have multiple products that might ship to a distributor or customer together, and each has a different size package. You want to be able to mix and match different combinations of products into the same outer shipping box.
There’s No Single Definition of “Hardware”
Many investors and others often throw around the word “hardware” as if its means mostly the same thing across all companies. That’s definitely not the case. The diagram below is intended to show some dramatic differences, using the amount and significance of custom-designed and developed hardware as the key benchmark for comparison.
Many software startups could simply provide instructions and prerequisites related to the hardware their customers should purchase to install their software on. But sometimes that’s too risky and uncontrolled, causing the company to bundle with and/or pre-load their software onto off-the-shelf hardware for a more predictable configuration. You’ll see that depicted on the left side of the graphic.
Other startups need to design and produce custom hardware components that are unique to their solution. You’ll see that depicted in the middle and right side of the graphic, and to varying degrees of significance.
All hardware startups care about things like component obsolescence, availability and lead times, shipping logistics and the other things mentioned in this article. But the further the startup is to the right side of the graphic, the more they are affected by things like design for manufacturability, raw component selection, second sourcing options, and obsolescence risk.
Summary
If you made it to the end of this article, it tells me you don’t scare off easily. And, as I mentioned at the start, I’m truly not trying to scare you away from pursuing your hardware startup idea. I just want you to enter into the venture with eyes wide open and with an understanding of some key issues and risks that you’ll want to think about before most first-timers do.
Good luck building a great hardware company!!!