Partner Article

The Internet of things recipe for success

By Denis Dorval, VP EMEA, Apigee

Finding the best use for a new discovery or innovation often takes time. So it is with the Internet of Things (IoT). When the concept first gained currency, we expected great things from it without appreciating what its real benefits would be. For too long it was illustrated with the tired old trope of the ‘intelligent fridge’ that would automatically order milk. The good news is that the reality of connected devices is far more exciting.

At Apigee’s annual I ♥ APIs conference in September, delegates were privileged to hear some fascinating examples of companies working at the cutting edge of the IoT. SAP demonstrated a vending machine that can change inventory prices based on contextual information. For example, it might alter the cost of drinks according to factors such as weather or demand. The machine can also provide customer analytics and reward repeat purchasers with complementary items.

Another large global company showcased the Connected Fitting Room, which is transforming how the fashion industry and other retailers will serve customers. When a customer walks into the fitting room, the tagged products trigger a screen that displays related products and relevant information to the customer. It demonstrates how digital data can be used effectively in bricks-and-mortar shops.

Swisscom presented the world’s first digitally connected e-bike. Through intelligent wireless interaction enabled via the cloud, owners can remotely communicate with the bike through their smartphones, tracking location data, performance, and battery status, as well as activating anti-theft settings.

Also at the show was VGo, showcasing its robotic telepresence solution. The 100% remote-controlled robots enable a person to replicate themselves in a distant location and have the freedom to move around as if they were physically there.

Of course, the long-awaited flourishing of connected devices is attended by a host of potential complications, lessons to be learned, and techniques to be perfected. Not least of the questions surrounding the IoT is how we can manage the huge volumes of data that it will create, and how to ensure that all these connected devices can share data effectively – not just between the device and a user interface, but between the devices themselves. For example, it’s useful to have your car tyres tell you that they’re beginning to run flat; it is even better if they can tell an intelligent compressor exactly how much air they needed.

We have noticed that companies initiating successful IoT projects, like those mentioned above, tend to exhibit a common approach that incorporates two key ingredients. First, they use predictive analytics to bring unbelievable levels of usable intelligence about their customers, markets, and products. Second, they use Application Programming Interfaces (APIs) to enable the worldwide developer community to interact with these devices, and build an ecosystem of other devices with which they can communicate.

Predictive analytics is one of the most exciting fields in data management; especially machine learning, an artificial intelligence (AI) technology that permits computers to learn from big data without requiring re-programming. For example, before the advent of machine learning, to predict patterns of behaviour – such as when a shopper makes a purchase – a data analyst might hypothesize three features: total customer spend, total number of purchases and the last date of purchase. The analyst would then test which of these features carry predictive power and experiment with various algorithms. With machine learning, this process is automated, requiring much lower levels of human input and therefore placing it within the ambit of a much larger number of companies.

The second ingredient, APIs, have been around for some time by modern standards. Yet the complexity and scale of the IoT, with such huge numbers and varieties of connected devices, requires organisations to take API management extremely seriously. This involves considerations such as gateways, security, access management, version management and a variety of other tools that enable businesses to control the relationships between their devices and the wider ecosystem of people, developers and machines with which those devices are connected.

Getting APIs right in this new world is a journey. You need to look at what has been done up to now and think of APIs as a key constituent of digital enablement. Ease of consumption is also key. Developers - whether internal or external - need to consume these APIs, so for this to happen they need be easy to use and combined with Internet-scale capability and enterprise-grade security, versioning etc.

Predictive analytics and API management may be necessary ingredients for an effective IoT project, but they do not necessarily constitute a recipe for success. Ultimately, it requires enterprises to create IoT implementations that will bring real benefit to their business and to users – not simply because they have the capability to do so. The best type of project is one that displays vision, imagination and creativity, while suiting the appetites of those for whom it is designed. If enterprises follow this simple rule for the Internet of Things, they will be creating masterpieces for many years to come.

This was posted in Bdaily's Members' News section by Apigee .