IoT has been prophesized as the next industrial revolution. With predictions of a multi-trillion dollar market and tens of billions of connected devices, nearly every tech company on Earth wants their piece. The problem, though, is that they're doing it all wrong.
The success of Nest and Philips Hue spawned a generation of copycats that fragmented the consumer IoT market so much that it basically destroyed itself. Think about how many smart bulb manufacturers there are. Now think about how many smart bulbs work together. We're seeing this collapse now with stories like Nest's disappointing numbers. The novelty has worn off and the value's not there.
The problem is not the products. Nest, Hue bulbs, and the million others are amazing pieces of technology. The problem is that each of these manufacturers attempted to own the market. They built siloed environments and in some extreme cases actively blocked interoperability.
IoT is not something you build your product with, it's something you build your product within.
Device manufactures should think of IoT as a landscape and design their products to appropriately fit within that landscape. It's a scary proposition. The company has to freely give up some level of control and accept the fact that products will be used in unintended ways.
It's the unintended use-cases where the real value comes. The software industry has already realized this and there are many examples of products that are infinitely more valuable because of the ecosystem built around them. Twitter is probably the best example. By itself, Twitter is pretty basic messenging platform, however there are thousands of products built on top of this basic concept. For example, you can now transfer money using Tweets.
A smart power outlet that I can turn on and off with my phone is a neat, but ultimately a pretty weak value proposition. I'm looking at you, Wemo. A smart outlet that I can turn on and off based on information from some other, unrelated, product. Now we're getting somewhere.
In its defense, the Wemo does have a local API, but this isn't LoT (LAN of things), this is IoT. The "I" is very important. If I can't access your IoT product using a remote API over the Internet, then your product is ultimately useless and irrelevant in the landscape of IoT.
What's funny is that nearly all of these products have remote APIs, they just refuse to release and document them. For example, I can control a Wemo outlet from anywhere in the world, but only if I use their mobile app. Philips Hue had a remote API, took it away, then has been promising to bring it back for years.
Small device makers are the same. I walked around this year's Launch conference talking to device startups and heard the same story each time. "You can control it from the app, we'll have a remote API at some point in the future".
API First is now a widely adopted approach to software development. Software vendors know that interoperability and customizability is a requirement. We're just waiting on the hardware makers to catch up.
An ecosystem of integrations, with your product in the middle, is what will make you successful.
IoT feels a lot like the Internet of the 80's. The groundwork was there, but it wasn't until the World Wide Web was invented in 1989 that things really started to take off. IoT doesn't have its WWW equivalent yet, but it needs one.
A free and standardized way to exchange data between disparate connected devices is vitally important to the success of IoT.
The industry, as I see it, requires two sets of standards. The first is to control how a device sends data to the cloud and how it retrieves commands from the cloud. The second is how external services can receive data and send commands - basically the remote API side.
The first would allow any device to connect to and be controlled by any cloud service. We're seeing very early versions of this through things like Apple's Homekit. Apple provides an opinionated way home appliances can connect and communicate. All appliances that use this standard can be controlled in similar ways, making it very easy to develop for.
The second facilitates interoperability between disparate systems. If I want to read information from the Wemo and my Philips Hue, I have two different API endpoints, two different authentication mechanisms, and two different data formats.
Having this set of standards is still far off. IoT developers have widely adopted REST and MQTT as transport mechanisms, but what goes in those payloads is still the wild west. Basically we have the Internet with no World Wide Web to make sense of it.
A single company's IoT product line, living in a enclosed ecosystem, has very limited value. It's not until developers and consumers can combine products from multiple vendors into personalized experiences that the real potential of IoT will be realized.
Your company won't own IoT, so don't try. It can only live harmoniously inside an IoT ecosystem. Your products will only be successful if they can be combined and integrated with other products in new and novel ways.