Azure IoT Suite gets down to business

Forget two-year IoT pilots. Prebuilt solutions and fast prototypes mean you can ascertain the business value of IoT quickly enough to stay competitive.

Comments

Whether your business is making jet engines, delivering towels or emptying rat traps, you need to be looking at how you can improve margins and customer satisfaction with the Internet of Things – because your competitors is.

Services like Microsoft’s Azure IoT Suite mean you don’t have to become experts in distributed systems to make IoT work for your business. That’s handy, because it’s more often business leaders than IT teams buying IoT. It’s not about gadgets or wearables, but the fact that what you do with sensors and big data can change your business model. In fact, McKinsey & Company estimates that nearly 70 percent of the potential value in IoT – which it estimates as high as $1.1 trillion a year by 2025 – will come from B2B use.

When the team that created Microsoft’s Azure IoT Suite looked at their customers in 2015, they found that only 2–5 percent of IoT projects were being run by the IT group. “It’s not the same people we usually talk to,” Microsoft’s Kevin Miller says. “Instead, they’re being run out of the business groups. They’re line of business decision makers. We typically see somebody whose primary goal is to get return on investment, to control costs, to drive revenues into a business. Some of these people have never really run software groups before, and now they're saying ‘I’ve got a thing, I can run the thing better if it’s connected to the internet, but my business unit doesn't have the skills for this’.”

This adoption directly by the business side follows the same pattern as software-as-a-service. So rather than letting different business units adopt their own approaches to IoT, CIOs should have a solution to offer them. That way, they can get things moving faster than the typical pilot project, which can take a year or two, says Steve Hoberecht, who works on IoT Suite.

“Talking to businesses who’d done pilots, we found the first few phases – where you envision and do your planning and ROI calculation and moving on to vendor investigation and selection process, and looking at a proof of concept – an enterprise customer will take anywhere from three to 18 months,” he says. (Usually it’s somewhere between nine and 16 months just for the pilot, but actually rolling out the final production system takes much less time.)

The idea behind the IoT Suite, and the Microsoft Azure certification for IoT – which now covers 35 partners including Dell and HPE – was to simplify the choice of the dozens of different technologies and speed up prototyping, Miller says. “Customers were telling us, I don’t even know where to begin … and I really don’t want to start by hiring a system integrator and paying a million dollars and taking 12 months to create a prototype. I want someone in the business team to be able to spend a couple of days and stand something up, customise it to my business and look at it and say ‘OK, I see what this IoT idea really is’ is and then have that built out of real, at-scale components. I want to start with 10 connected devices, but I'm going to put a million in and I don't want to have to re-architect.”

That’s why the IoT Suite includes not just services like the IoT Hub (which reached general availability this February) for dealing with the flood of data from connected devices, and analytics tools for getting value out of that data, but preconfigured solutions for remote monitoring and connected maintenance that businesses can customize (the code for the solutions is on GitHub under an open source license).

Remote monitoring may not be new, but now it’s affordable regardless of company size, claims Miller. “The difference is hardware got cheap, sensors got cheap, computing got cheap, cloud availability became global and, just as important, the ability to analyze and so something with that data got mature enough that you could actually do something.” That expands both the range and the reach of what you can do with IoT.

“Ten years ago you could do it, but you could only afford to do it if it was something like a jet engine and failure was catastrophic and an unthinkable problem. Now, connected things are not super expensive and we find people putting tags on everything from hospital and hotel linens, to pets and livestock. They want to tag the oil when it comes out of the ground and track it to the customer; you can afford to do that when connectivity is pervasive and cheap.”

Location isn’t the only useful sensor: Audio and vibration data is starting to transform maintenance, Miller points out. “You can listen to the microphone patterns and there’s something squeaking that shouldn’t be, something vibrating that shouldn’t be. You can diagnose almost any mechanical device in world with a microphone.”

From efficiency to new business models

IoT for business often suffers from hype, but Miller says there’s a pattern to successful deployments: Start with the data and look for inefficiency before you start thinking about big changes. “There's been some press about IoT which suggests IoT is the magic unicorn that will solve all your problems and transform your business. You can get to transformative business outcomes, but most of the successful ones I've seen have not tried to start with that. They’ve started with ‘I'm going to run my existing business better, I'm going to find efficiencies better’ – because almost anything with humans involved has inefficiencies.”

One of Microsoft’s most-publicized IoT customers, ThyssenKrupp, is a good example of what he calls the maturity model of IoT in business. The first step was putting sensors on elevators to discover whether the model they were using for regular maintenance actually matched how often they needed servicing. “It turned out their elevators were more reliable than they thought,” Miller explains. “They realized they could adjust their maintenance window and not see a perceptible increase in the number of customer calls. They saved maybe 15 percent – on a billion dollar business.”

After efficiency and reducing costs, the typical second stage is innovation that increases revenue, using insights from the data you’re already using to improve efficiency. In the case of ThyssenKrupp, “they realized they could predict what was going to fail with high reliability and they’re somewhere in the mid-‘90s accuracy at getting the right part predictively on the truck. Now they save more money still by having the technician show up ready to do the repair, but they also have this tremendous customer experience win.”

Interestingly, ThyssenKrupp has been careful not to suggest that the prediction system replaces the expertise of their engineers. “Instead, they position the IoT system as a coach for the repair people: ‘this is going to make you better at your job, because you’re going to go out armed with the parts you need to do your job’.”

Another customer got both savings and better customer satisfaction by putting sensors in their industrial rat traps. “It doesn’t take a dead rat long to stink, so to avoid the call that says ‘it stinks here’ they’re checking their traps very frequently, and most of the time they’re checking empty rat traps,” explains Miller. “They’re able to save themselves a lot of money driving around checking empty traps, plus when a trap catches a rat, they can now get there really fast to pick it up.”

A linen service that delivers towels and sheets to hotels and hospitals found it was losing so much inventory that it was costing them millions of dollars a year. They already had trucks and laundry equipment that could read sensor tags, but they’d never put tags on the linens. “Now they can say ‘I delivered a thousand towels and I only got two hundred back’. That was pure inefficiency.’ The next step was to use the data about what linens were being laundered to run the equipment more efficiently as well. “Then the CEO said, I know what's in all the trucks, so I can run my system leaner; if I need towels for a customer, I can route a truck with towels on.”

The third stage is the business transformation that can give you a whole new business model. “If an elevator in a factory that moves something expensive being down brings the line down,” says Miller, “ThyssenKrupp can go back to those customers and say ‘we'll sell you an uptime guarantee.’ They now know their business well enough to know what sort of claims they can make for uptime and they know how to think about the potential costs of the guarantee and what offer they can afford to make to customers – and they've made a new business out of this.”

That’s a pattern Miller says is typical. “When they saved a lot of money, the CEO said ‘is there more free money here?’ They saved enough money that they could start thinking about investing and they started following their nose on the data and going across the maturity model.”

Even the more basic implementations made a significant difference to the business, and Miller points that that if you’re not thinking about IoT for your own business, you may find it hard to compete.

“When someone like ThyssenKrupp gets really good at predicting failure, that puts pressure on Otis. When somebody who is tracking materials is able to do so more efficiently, that puts pressure on their competitors. We're starting to see this. Customers who adopt IoT are able to run businesses more efficiently. It's an ‘Ah-hah’ moment: I can lower my prices and keep my margins where they are and that's going to drive my competitors nuts.”