From the outside, it seems like building an IoT product should be easy. But that probably hasn’t been your experience. You may say to yourself, “why is my team struggling to deliver a valuable and usable product, and are we going to miss our big opportunity?”
Let’s be honest — you suck at IoT. You suck at it because you underestimated the complexity of the problems you need to solve, because the “IoT experts” you hired have actually never done this before, because you outsourced all of your engineering capabilities in the ‘90s when all you cared about was cost reduction, and because you’re unwilling to pursue business models that are dramatically different from your current business.
But don’t worry; these are solvable problems. Before we get into the solutions, let’s talk about why it feels like IoT should be easy.
The web is easy today. But it wasn’t twenty years ago.
In 1996, it could cost millions of dollars to build a website. In 2017, it costs thousands. This reduction in costs has come from a number of evolutions that have happened in parallel. Browsers have gotten better. Servers have moved to the cloud. Winning standards and protocols have emerged and stabilized. Programming languages and frameworks have evolved to make web development easier. And many of our brightest minds have spent decades learning to do web development and have developed deep expertise that makes solving hard problems much easier.
One would expect that many of these advances would transfer easily to the Internet of Things. One would, unfortunately, be wrong.
Browsers (mostly) don’t matter in IoT, so advances in browsers don’t help. Servers moving to the cloud is a good thing, but if you thought that meant you wouldn’t have to touch another circuit board, think again — now you’ve got to deal with the hardware in your Internet of Things thing. The standards and protocols that rule the web don’t necessarily work for IoT; we need new ones that are designed for resource-constrained, battery-constrained, bandwidth-constrained IoT devices that don’t have a screen or keyboard. Same goes for programming languages; the great languages and frameworks that made Web 2.0 possible are too heavy for IoT.
Same goes for programming languages; the great languages and frameworks that made Web 2.0 possible are too heavy for IoT. To top it off, web developers come into IoT with strong biases and preconceived notions based on their web experience that may not translate to IoT, and may lead them to fail.
We’re basically starting over again. Building an IoT product in 2017 is like building a website in 1996. It’s expensive, it takes forever, and you’re going to be banging your head against the wall solving problems that you think should have been solved for you.
But let’s look at the bright side — the companies that solved the problems of the early web went on to create extraordinary value. You could be the next Amazon, eBay, or Google. So let’s talk about how to get there.
IoT is harder than you think it is
Gartner says that 80 percent of IoT projects will be over-budget and deliver past their deadline through 2018. When you put in place a team to run the development of your IoT project, don’t be upset when they spend more money and take longer than expected. It’s easy to blame the team for over-promising and under-delivering, but they probably didn’t realize what they were up against.
Instead, be conservative in your planning. If you’re budgeting one year and $1MM to your initiative, you should be comfortable with it taking two years and $3MM. Your only alternative is to cancel the project when they don’t deliver, which means that you’ve wasted time and money and are no further along than you were when you started.
You should also consider working with a partner — someone who provides an IoT platform that fits with your needs. Evaluating IoT platforms can be challenging when you don’t yet have deep expertise in the industry, but the biggest thing to look for is a partner whose platform fills the gaps in your organization. If you’ve got existing IoT devices out in the field that are instrumented and connected but you just need to collect the data, then buy a IoT data analytics platform. If you are building new IoT devices from scratch or retrofitting devices that aren’t currently connected, buy an IoT device connectivity platform (like ours). Find a partner who knows the answers to the questions that you don’t even know to ask.
Be skeptical of “IoT experts”
I get a lot of LinkedIn requests from people who are self-proclaimed “IoT experts”. I see them get hired by our customers and potential customers. Often these people are a disappointment. This is not necessarily their fault; they do have relevant knowledge, even if they’re overselling it on their resumes. Nonetheless assuming that hiring a single “expert” will solve all of your problems is setting yourself up for failure.
In many cases, an “IoT expert” is someone with IT experience who once built something with an Arduino or Raspberry Pi. In other cases, it’s someone who was project manager for connected jet engines at GE with skills that might not apply to your connected door lock. These IoT experts know enough to be dangerous; they can speak the lingo, but are likely to encounter the same challenges and frustrations as everybody else. They are, however, much more likely to pretend that they know what they’re doing and barrel through instead of admitting their challenges. After all, they’re supposed to be experts.
Recognize that, when an industry is so young, there are far fewer true experts than there are people claiming expertise. Be skeptical. And perhaps, instead of hiring for expertise, you should hire for potential: find people that have relevant experience that can be applied to IoT and have the creativity and hunger to learn IoT.
Hire great engineers with broad skills
If you’re a hardware OEM, you probably lost a lot of your engineering mojo in the ‘90s and early 2000s. At that time, Chinese and Taiwanese ODMs were getting better and better, and at some point, their hardware engineering and design skills surpassed ours, at least dollar-for-dollar. As a result, many U.S. and European OEMs divested their engineering organizations in favor of “vendor managers” — people who know how to work with and oversee their ODM partners, but don’t know how to build stuff from scratch.
All of a sudden, it’s once again important to have your own world-class engineering organization. IoT products that are outsourced to a third party suck because you lose the ability to closely oversee and iterate in-house. You don’t want your product to suck. So start hiring.
There are a lot of spectacular engineers out in the world. The best ones are extremely curious and love working through novel challenges. Recognize that, if you’re building something brand spanking new, that makes it a lot easier to hire great engineers.
Search for areas of overlap. For example, you might consider hiring web developers who worked on embedded systems earlier in their career and embedded engineers with a background in web development. This makes it possible for the team that’s building a cloud back-end to empathize with the challenges of the firmware team and vice versa. Your product will be better if your software team and your hardware team know how to talk with one another.
Pursue game-changing business models
I see a lot of companies who approach IoT by dipping their toe in the water. “Let’s just take one of our existing products and connect it to the internet. We’ll figure out the business model later.”
On one hand, I appreciate the fact that they’re not waiting for some grand vision to get started. Yes, you should get started building prototypes and proof of concepts right away so you don’t fall behind. But on the other hand, I think the best opportunities require dramatically re-thinking your business.
Let’s look at the early web and Amazon.com. In the beginning of the web, the way that product companies and retailers “dipped their toe” in the web was by putting up a website with pictures of their product. “Now we’re a technology company”, they said.
Then Amazon launched an online bookstore. In order to do so, they had to do a lot more than just build a website. Yes, they had to solve hard technical problems, like online payments and shopping carts. But then they also had to build a distribution network. They had to do a bunch of dirty, non-techy things to become one of the world’s largest tech companies.
Amazon ate everybody else’s lunch because they dared to reinvent an industry. Don’t let that happen to you. Invest deeply in IoT; be comfortable with business models that are disruptive and cannibalize your current business. Allow yourself to truly become a tech company.
IoT is hard, and its future is uncertain. Don’t let that uncertainty keep you from investing deeply in your future as an IoT company. Right now, you suck at IoT, and if you don’t stop sucking, you’re going to find yourself getting Amazon’ed.
Don’t be Barnes and Noble.
The author is the CEO and co-founder of Particle.