Industrial IoT’s biggest problem? The fear of failure in a risk adverse industry. So says Ericsson, which recommends a two-pronged approach to spur takeup among overly-cautious manufacturing companies.
The first fix is to drive starter IoT with proven low-level sensor solutions; the second is to drive advanced industrial automation with higher-end networking and analytics. One is easy, and one is hard. That is the conclusion from a conversation with Kushala Silva, head of technology at Ericsson IoT Studio, the Swedish vendor’s IoT developer lab in Silicon Valley, where it seeks to “move the needle” on industrial IoT.
Kushala joined Ericsson from VMWare in August 2019. But his recruitment was tactical for Ericsson, also, because of previous roles on the other side of the digital sales line. He is a veteran of Procter & Gamble (P&G) and Coca-Cola, where he was global head of digital innovation and group director of innovation, respectively.
His recent experience, and his candid conversation, makes his views of how industrial IoT – covering high-end sensing (5G, IoT) and sense-making (AI, edge computing) – should be bought and sold more telling, even as he pushes the familiar case for private wireless networks.
Factories are prime candidates for industrial IoT and networking, he reasons; but they are also killer clients for sales teams in the technology space, as their owners and operators want proven technologies – and don’t much like upstarts on their turf, either.
Here, he discusses the challenges and opportunities for digital technologies in industrial sites, and how to bridge the IT/OT divide, drawing on the fizz and pop of his digital experiments at P&G and Coca-Cola.
…
Why is IoT so well suited to manufacturing, and the wider industrial space?
“Simply, because any optimization in the industrial space turns into profit – directly impacting the bottom line. So organizations are naturally interested in what IoT can do for them. The other reason industrial, and manufacturing in particular, is so well suited is because these organizations are closed environments.
“P&G, Coca-Cola, Unilever – these guys can all change and create the toolsets within their own environments. That’s why conventional wisdom says IoT is such a perfect fit for these kinds of areas.”
But it is not working – or it is not working as well as it should be? Is that the point?
“Yes. I mean, in the last year, many organizations have started to adopt IoT, even if it is in a small way. But historically manufacturing is very conservative. I remember when P&G started implementing SAP across the globe, way back in the 2000s – it started off in finance, and then distribution; the last place to roll it out was manufacturing.
“Because the leaders on the manufacturing side made clear they wanted a tried and tested solution. They want proof. The message was not to brake something that is not broken. The SAP implementation happed everywhere else, first. That attitide is the same in manufacturing with IoT.
“It was the same at Coca-Cola – which has 238 manufacturing plants across the globe. When I talked about IoT with the heads of each manufacturing category, they said they were super excited, of course, but that they were also worried about reliability – because they are running mission-critical machines and porcesses.
“‘How do we know it is going to work?’ That is the message we see, over and over. At Coca-Cola, we tried to enable supply chain networks using Wi-Fi. We retrofitted IoT sensors to some of the forklifts, and packaging [processes]. Corporate Wi-Fi is strong, but it doesn’t deliver the reliability we needed.
“That was the issue. Reliability is everything in these environments. The plant managers wanted to know, in trial, [how to respond] if a rogue device was planted in the organization. I mean, there are muliple ways to skin a cat, but there is no really good answer to that.
“I wish we knew about industrial [cellular] networks when I was at Coca-Cola. I am excited at Ericsson, that we offer private 5G networks to these types of plants. And the minute you have a very reliable, very secure network, then all of a sudden the infrastructure is there to explore all of these aspects of IoT. That is a huge value proposition.”
It sounds like part of the solution to this blockage for industrial IoT is technical – that ultra-reliable connectivity will provide some assurance to the industrial set. Is that right, or are you saying they are still a conservative bunch, and these projects are still getting stuck in proof-mode?
“Yes, actually, the latter is correct. The risk-averse nature of these organizations means they will look into these things very carefully. Even when technical organizations show the technical aspect of these networks is tested and proven, there is still convincing to do for these industrial organizations to move.
“The high-tech industry – which includes companies like Ericsson, with its own manufacturing plants – will adopt industrial IoT [and industrial cellular] quickly. That is the entry point.
“We have partnerships within high-tech industries – with companies like Hitachi. As these companies start deploying these technologies, then the CPGs and FMCGs (consumer packaged goods and fast-moving consumer goods manufacturers) – with the highest number of industrial sites – will also start adopting.
“But it will happen little by little. I don’t expect these organizations to jump into it, neccessarily, just because a tech organisation or a consulting organisation tells them to.”
This seems to be key – that industrial IoT does not always fall into the IT domain. It crosses into the OT domain, which is controlled by a different group. Is there a perception the IT crowd is hustling-in, and some resistance for that reason?
“Yes, that is what you find. That is a problem. In most companies, the IT function looks after the network. But in the supply chain, there are the operations folks, who are very technical, and are even running the IT infrastructure inside the plants in some cases. And this bunch doesn’t report into the CIO.
“They report into the supply chain organization; the OT team always rolls-up into a president-level function. So that is a challenge; technology companies have historically sold to CIOs, who now, all of a sudden, have IoT capabilities somewhat in their control. They are making decisions around IoT platforms, for example – which are [used and managed by] the OT side.
“When I was at Coca-Cola, one of the key learnings when we started was to work with the OT teams – they were my partners, telling me what needed to be done, and not just the people that received the capability. That was a great way to get it done.”
Is Ericsson selling private LTE and 5G networks to the IT or OT side?
“To the CIO. But they (private networks) are not really mainstream yet. We are working to get them in front of the right people. And there needs to be some socialization of this capability within the OT community – to get them involved, too.”
How is this resolved, then – if cellular provides technological reassurance, but more reassurance is required to make industrial IoT mainstream?
“It requires a multi-step approach, much like with analytics a decade ago, when no one believed it until they saw the results. If you talk to a general manager in an industrial plant, they’ll say, ‘Yes, yes, we know the fancy words, and we know these things will come, but we’ll adopt when the time is right’.
“The industry needs a few IoT use cases to be proven, and to deliver huge value. Because if you are looking to apply IoT in an FMCG-type of company, and you wait until the supply chain function or the manufacturing chain function adopts it, then it is a tall ask.
“But the marketing side, and the sales and distribution sides, are always looking for ways to improve performance and optimize efficiencies. If you can get IoT into those realms, then the general managers and presidents will start to talk about it – and the production side might follow.
“That is the startegy we took at Coca-Cola – to go for the low-hanging fruit, first. Because the OT side will not jump ship just because of a Gartner or McKinsey report. They want proof, within the organization itself, that IoT is real, and that the data captured from it is being used to improve business.”
Can you talk about these use cases that turn into business cases? Enterprise IoT Insights has just posted an interview with Orange, which chimes very closely with a previous talk from Vodafone, which says there are only two use IoT cases – tracking and monitoring – and that everything hinges off those. Can you talk about that, and also give a couple examples of what worked at Coca-Cola?
“Yes, and I basically agree with that assessment – not everything about it, but most of it. Coca-Cola identified four potential IoT capabilities: manufacturing was one, and another was logistics. Coca-Cola has more trucks than UPS, FedEx and DHL combined, so logistics is a big area for Coca-Cola.
“The third was dispersed assets – Coca-Cola has about 22 million coolers across the globe. And the fourth was the employees, the contractors – that type of aspect, and how to manage the real estate. Those were the big focus areas, which would take two-to-three years to materialize.
“But while we were looking at them, other simpler use cases popped up. One of those was to do with thefts in the vending business, from these machines selling cans of Coca-Cola for $1 each – where some [delivery] drivers would buy cans from Asda and Walmart for 20 cents a pop, swap them into the machines, and pocket the difference.
“That was a real issue at Coca-Cola; the company was losing about $18 million per year in North America. I was speaking to one of the presidents, who indicated this was a recurring problem – not just in terms of the $18 million, but also the compliance aspect. The way we solved it was to put an IoT lock on the vending machines that recorded when they were opened and closed.
“Because drivers were switching products in their off-hours, and not on their regular timetabled routes. By putting the locks on, and marketing the fact to the drivers, the thefts stopped overnight. And it was not a complex solution – it was a cellular-enabled IoT lock. That’s all. But we got a huge amount of publicity internally, and it has been rolled out across the globe.”
It seems the vendor community has to be smarter about pitching these solutions, in order for IoT to be noticed first, and to scale from there. But how do you sell campus-wide private networks if you are being tactical? Because the opportunity and strategy for Ericsson is not to sell sensors on locks, but to sell millions of private networks.
“What Coca-Cola did was build credibility. But, yes, you need to be careful when you put these plans together. Some items need to be tactical, to prove the point, to prove it really works. But you also need to keep your eyes on the bigger picture. As you say, the private network is not a tactical sale.
“But from Ericsson’s point of view, these private networks, and the niche capabilities you can provide with them, can be proven easily – without changing the operational lines, without changing much at all within the operations environment. And the value they bring is very clear.
“Safety in a factory, for example, is of paramount importance; a lot of money goes into it, to know who is going into which zones, and so on. Yes, the latest versions of Wi-Fi do that. But with cellular, you can create sectors, and check the vital signs of people going into those sectors.
“These kinds of things will provide huge financial benefits for companies. You can drop your insurance premiums if you have capabilities like this – which is a number on your bottom line. So the strategy should be to use private networks to solve cases that are solvable today.”
