Stage 1: Embedded Devices
Starting at stage one: embedded devices, we have any device with circuitry that fulfills a specific purpose. Think about cassette players, Walkmans, or even iPods. These stand-alone devices play music, and that’s about it. This stage of the model should be familiar and fairly self-explanatory, so we won’t dwell on it for too long.
Stage 2: Cloud Computing
Going up to stage two: cloud computing, we have devices that interface with a centralized point over the internet, usually a server. For our music example, some popular services include Pandora and Spotify, especially in their early 2000’s form before the Machine Learning (ML) explosion and the rise of recommendation engines.
The key mark of a stage two product is that one device talks to one other device. In the case of music streaming services, our computers or phones are talking to the service provider’s cloud-connected machines.
Stage 3: IoT Connectivity
Next, we climb up to stage three: IoT connectivity. This is the stage where a device truly becomes an IoT product. In the music world, the Sonos is a great example. Not only do these smart speakers connect to each other in order to optimize acoustics and create a richer listening experience, but they also integrate with a variety of other systems.
The Sonos app lets us control our speakers with a smartphone. Compatibility with Amazon Alexa and Google Assistant let us use them as voice assistants for a variety of other use cases. Apple users can use Airplay to play their music on Sonos, while Spotify subscribers can hook up via Spotify Connect. These speakers can also connect to various other devices throughout the home, including televisions and even vinyl record players.
The takeaway at stage three is that we now have a comprehensive IoT ecosystem. By linking together these interconnected devices, we begin to see the true promise of IoT—but we’re still far from the end of our journey.
Stage 4: Predictive Analytics
Moving on to stage four: predictive analytics, we can now start thinking less about the device’s baseline technical aspects and start thinking about data and what we can learn from it. This is where Machine Learning hits the scene; our devices collect the data, and then our data scientists get to work aggregating it, cleaning it, and feeding it to ML models. Essentially, predictive analytics uses “statistics and modeling to determine future performance, based on current or historical data. Predictive analytics look at patterns in data to determine if those patterns are likely to emerge again.”
A common domain where we see predictive analytics is in the Industrial Internet of Things (IIoT) with predictive maintenance. For instance, we’ll feed sensor data from a factory machine into an ML algorithm, and the output is how likely that machine is to break or when we can expect a malfunction, thus letting us repair it before it breaks to minimize downtime.
But we’re talking about music. There are a couple ways we can see this play out. One is the familiar recommendation engine; if we’re big fans of one genre or artist, a streaming service can look at the listening habits of other users who share our tastes and then recommend new music. Another possibility is a smart speaker that can tell if we’ve rearranged the furniture in the room and can recommend a new placement for the speaker to optimize the room’s acoustics. A third option is a speaker that recognizes when its sound quality degrades and can recommend picking up a new one.
Stage 5: Prescriptive Analytics
This is where things start to really get interesting. Stage five, prescriptive analytics, is today’s most cutting-edge business model. Not only are stage five products the most technologically advanced, but they also generate consistently high revenue streams. Though this stage shares some similarities with the predictive models of stage four, in that it uses current and historical data for ML modeling, the main advantage of prescriptive analytics is that it recommends which future course to take.
This is the stage where we answer the question, “How do we make the user experience (UX) better with the data we’re collecting?” For instance, I like to listen to music every night while I cook dinner, and I tell my smart speaker to put on some music when I enter the kitchen. If we introduced prescriptive analytics into this equation, then an algorithm will look at both me and my wife’s work calendars, figure out when our last meeting is, and then ask if we wanted to listen to music.
We could take this one step further by figuring out what kind of music I like when I’m in different moods. For example, it could take biometrics from a Fitbit, notice that my blood pressure is up because of a looming deadline, and play some chill beats to relax to.
The ability of businesses to use the data that they’re gathering on users from IoT devices creates valuable insights that reveal new revenue stream opportunities. A stage five IoT product can recommend new products and services to persuade users to spend in ways that are more natural and intuitive for them. I want to emphasize, however, that the goal here isn’t just to pry people's wallets open. Rather, our mission is to provide a tangible benefit that makes their lives easier or more convenient.
Stage 6: Ubiquitous Computing
Now we approach the sun, the pinnacle of the IoT maturity model, the summit that we all aspire towards. An IoT product is truly mature when it reaches stage six: ubiquitous computing. In this paradigm, devices are completely interconnected, constantly available, and downright pervasive. Muskian moonshot technologies certainly fit the bill: self-driving Teslas, worldwide internet beamed down from the Starlink, and, of course, a literal computer chip embedded into one’s skull, the Neuralink.
While there’s no company or society that’s turned this science fiction into a reality, we’re getting closer. This is especially true for the Chinese, who are investing heavily in smart cities, advanced artificial intelligence systems, and a unified data landscape on all-in-one platforms like WeChat. That doesn’t mean, however, that American companies like Amazon, Google, and the aforementioned Musk-eteers aren’t making great progress towards bringing the benefits of ubiquitous computing to the west.
A music service in a world of ubiquitous computing would quite literally be a soundtrack to your life. Putting on your jogging shorts? Cue the motivational jams. Getting ready for bed? Turn down the volume and play something relaxing. Besides these automatic transitions, we’d also be able to seamlessly control the music with voice controls, or perhaps even a good old -fashioned remote control on occasion.
Now that’s a powerful IoT product.
Long-Term IoT Strategy: IoT-as-a-Service
The bottom line is that a mature IoT product uses an IoT-as-a-service business model. When we get to the point where our device continues to add value to a user’s life because it’s so effective at giving them exactly what they want and need, we find ourselves with a pipeline of recurring value. This is precisely how a product retains relevance for years to come.
Now here’s the big question — how do you actually get your product to reach maturity? What technical challenges crop up along the way, and what do you need to do to surpass them? In our next post in this series, we’ll discuss the technological hurdles that you need to clear to advance from one stage to the next.
Ready to dive deeper? Download the full PDF of the IoT Maturity Model here.