Hardware Is SEXY!

I was recently asked by The Things Conference to give a keynote to help inspire hardware related companies. Here is my vision of how to make hardware valuable and the future of embedded technology. A video of the keynote is included at the end!

For the much of the last decade hardware has often been looked at as a commodity, and hardware startups as less valuable than software or SaaS. Three trends are now changing this equation, making hardware the new SEXY:

• User experience (UX)
• Hardware as a service
• Personalization

One visionary of hardware is Elon Musk. He brought EVs into the mainstream by rethinking the UX and introducing hardware as a service to an old-fashioned industry. The Tesla Models S completely re-thought the automobile from a user centric point of view, from the purchasing process to controls and the product feature lifecycle.

Every hardware product needs an obsessive focus on the user experience; from mechanical design to user interfaces and the entire lifecycle.

Back in 2015 Wired reported on a massive remote control vulnerability in the Jeep Cherokee. This serious security hole resulted in a massive recall of 1.4 million Jeeps. The fix required each vehicle to be updated with a USB stick (usually at a dealership), resulting in serious losses and a PR nightmare for Chrysler. A similar attack through physical access to the Tesla Model S was reported around the same time. Tesla quickly patched the reported vulnerability with an automatic over-the-air update resulting in PR acclaims for treating the vehicle as a service, not just a piece of hardware.

All of us in the hardware industry need to start looking at our product as a service, and take advantage of the value that it brings, not just the manufacturing BOM.

In 2016 I had the once in a lifetime opportunity to help the BBC Micro:bit partnership turn an amazing educational hardware project into a global non-profit product. The Micro:bit was the ultimate UX challenge - aimed at one of the most challenging user groups possible: children starting at 8 years old, teachers and parents. This is where I really came to understand the power of user focus, and keeping the entire organization laser-focused on their success.

If engineers were to have designed the Micro:bit, it would have been yet another difficult to use embedded devboard!

Instead, by including teachers, psychologists, media experts and the children themselves in the process from the beginning (with a focus on gender-neutrality), something magical was born. The Micro:bit was designed to be attractive to young children, and the entire experience was created for an educational environment. Technology choices were made in support of the experience, not the other way around. The result was a very affordable device, that was easy to manage in schools and at home, and did not require any special tools besides a web browser. Since the original national deployment in the UK in 2016, the Micro:bit has reached millions of children in countries around the world and has had a real impact in education.

The next frontier for hardware is going to be personalization. Recent improvements in embedded processing and wide bandwidth sensors are rapidly making Machine Learning accessible for real-time embedded sensor applications at low power (TinyML).

TinyML is huge, as it not only opens up the 99% of sensor data that is discarded today, but it allows us to train hardware to work better for the specific environment it is deployed in, or even the individual user.

This personalization of hardware will enable a new level of value for the services we provide around it. I foresee this combination of TinyML and personalization as the driving force of the embedded industry for the next decade. We started Edge Impulse to help developers take advantage of this new technology from getting started to production scale!

Investors are taking notice of hardware companies that are driving value through user experience and services. I recently spoke with Butterfly Ventures, an early-stage Nordic VC that focuses on hardware related investments. Their data shows that hardware startups are able to generate better returns and require less customer acquisition costs compared to SaaS startups.

The underlying market dynamics are being driven by the continued growth of connected devices vs. human population. On top of that research shows that customer acquisition costs for software have grown to double that of hardware. This may explain why hardware startups are seeing an 81% uptick in valuation between Series A and B, compared to just 42% for software.

Then why are software startups more valuable than hardware startups? This is just a myth - take a look at the return on invested capital across unicorns! Hardware startups perform at least in well across return multiples, in particular for the most valuable returns over 20x.

There is a story behind the title of this talk, and it comes back to Mr. Musk. He had a master plan in mind for Tesla products, and stuck to it regardless of Ford blocking use of the Model E name. Have to love the spirit!