Can hard guys go soft?
A meditation on hardware vs. software product management.
Software product management is sexy. When people think tech, they think of FANG and other “hot” companies like Uber, Instagram, and DoorDash. But interestingly, lot of software companies have either moved into hardware (like Google buying Fitbit); own the hardware like data centers to run their software platforms (like Facebook); or design their own software and hardware like Apple and Microsoft.
For the purposes of this article, I am focusing on hardware tech products that have a heavy software component such as IoT, wearables, and control systems. Without software, these products would be essentially expensive paper weights. As you can imagine, this is double the effort: you not only have to ensure that the hardware works as designed, but also the software that controls the hardware.
Creating hardware tech products is tough. You need to get both the hardware and software right. The product needs to work 100% out of the box. Imagine if you bought golf sensor for your club and it does not connect to the app on your phone. The app is essentially useless not matter how slick it looks. This actually happened to me recently with Garmin’s CT-10 sensors I bought to track my golf swing.
Long Lead Times
Hardware products can take up to two years to ship from initial concept to launch. There are multiple phases from concept to production and each phase could involve a procurement cycle, which is a fancy way of saying “waiting for parts”. In two years, you could be ready for a Series A in the software world, especially if you hit the product/market fit right out of the gate.
The biggest issue tends to be long procurement cycles. China is the manufacturing hub for the world. Prior to COVID-19, ocean freight took 5 weeks. Currently, it sits at 12–14 weeks. Forgot to order a part? Tough luck, cause you will have to wait for 14 weeks before it arrives, or you to have pay for expensive air freight that chews into your budget. In 14 weeks, you can easily code an MVP and maybe even an alpha depending on how many developers you have.
Discipline
It takes a lot of discipline to ship hardware products. I hear stories of software developers fixing up patches or features over the weekend. No such luck with hardware. If a part fails, it likely involves 1–2 weeks’ time to perform a root cause failure analysis. You may have to order a substitute part locally which could take another 1–2 weeks. So right there, take a month off your schedule.
Staffing
The biggest human capital need in software companies tends to be developers. The other non-technical functions like sales, UX, and customer success are essentially support functions that wouldn’t exist without the software.
On the hardware side, depending on product complexity, you may need mechanical, electrical, thermal AND software engineers to build your products. The electrical design could impact the mechanical and thermal requirements.
Additionally, you need procurement, quality, manufacturing, and logistics personnel to take the product through to the finish line.
Commercial and business model differences
Software companies tend to have relatively low overhead. You may not even need physical offices if you are small enough. All you need is a laptop and a high speed internet connection to stay in touch with your colleagues. With the advent of online collaborative tools like Miro, you can even brainstorm remotely. Hence, margins can be as high as 80–90%.
First mover advantage makes a huge difference in any industry, but I can’t help thinking that it makes a bigger different in hardware. If you buy into the narrative that hardware products take two years to develop, you have a two year head start of competitors. But if you are a software company, that head start could only be weeks, if not months before competitors replicate that feature.
Competition also tends to be pretty fierce in the software world. It is tough to build a “competitive moat” simply based on a feature list, which can be pretty easy to replicate. Clubhouse, TikTok, and Kickstarter are great examples of software companies that have built up a moat by creating a network of like-minded individuals rather than cool software features.
Hardware maybe more difficult to replicate as it involves creating a product, but also mastering the supply chain, procurement, logistics, and quality pieces. Replicating the product during the mass production phase involves manufacturing expertise, which is quite different from product design and prototyping expertise. In essence, the core advantage is the expertise in manufacturing at scale, rather than product innovation.
Bringing it all together
Software can be a lot more exciting and fast paced. You can prototype concepts really quickly by mocking up wireframes and testing the concept with stakeholders. You can release a new feature over the weekend and launch it in a month’s time. Product design also gets a lot more focus in software as UX can be a differentiating factor.
In my opinion, it is easier to transition into hardware if you have a background in energy, construction, transportation, or really anything that’s hardware oriented. Software companies are reluctant to hire product managers who don’t have a software background, although this could depend on the company and industry you are targeting. Larger companies have more resources to train you and nascent industries may not have a choice but to hire external to the industry.
I would highly encourage you to do your own research into what is the right fit for your skills, interests, and passions. Ultimately, it comes down to what your interested in and where your skills lie. In my experience, it’s tough to go back and forth. Folks usually pick either the hardware or software world and spend a bulk of their careers in there.
Hope that gives you a better picture of the tech world at a high level. I have made some sweeping generalizations, but on the whole, the observations hold. There are many more subtle differences, so I have created a handy table to summarize it all and included a few additional attributes not covered above.
