Industrial Data Summit 2018: How do you gain insight at scale?

Data, advanced analytics and machine learning hold a ‘Pandora’s Box’ of both opportunities and challenges, our chair for the day – Oxford Engineering’s Karim Sekkat – began.

The primary hurdle facing almost every business which is yet to start their ‘digital journey’ is knowing where to start – a deceptively ‘simple’ objective, Sekkat continued, which leads to follow-up questions such as who, what, why, when, and how.

For those who are already leveraging the value of digital technology and data, the prizes on offer are compelling: improved productivity, increased asset utilisation, preventative (even predictive) maintenance, closer customer relationships, faster innovation, heightened supply chain visibility, the creation of new revenue streams, and more agile, flexible businesses overall.

To help manufacturers see digital and data as an enabler of business growth, rather than simply an ongoing business cost, the Industrial Data Summit’s innovative format focused around a series of intimate roundtable sessions – allowing decision-makers to sit next to the experts and have their questions and concerns addressed first-hand.

Roundtables discussions took in the Internet of Things, IT/OT convergence, machine learning, data-driven sales & operations, real-time energy monitoring, supply chain insight and artificial intelligence, among other topics, and delegates were given the opportunity to sit at five tables throughout the day.

Before the first roundtable session got underway, Chris Cameron – worldwide sales leader for IBM, took to the stage to deliver the first keynote: The New Supply Chain: Collaboration + Blockchain + Artificial Intelligence.

You can read The Manufacturer’s overview of Cameron’s keynote here.

Panel Discussion #1

Overseen by The Manufacturer’s editorial director, Nick Peters, the first panel looked to answer, ‘Whose data is it anyway’.

It’s a very important topic, noted Nick Frank – a specialist advisor for Hennik Edge. As data ownership has ‘legal implications and complications’, before sharing data you have to first identify what value you’re going to get from doing so, ‘is it worth it?’, he cautioned.

“If there is value in sharing data, then be sure to put appropriate mechanisms – such as processes and contracts – in place ahead of time,” Frank added.

It’s also important to ask yourself whether you’re sharing quality data or nonsense, said Phillip Woodall – senior research scientist at Cambridge University. “Inaccurate, unformatted or poor-quality data could see you waste time, money and effort getting something into shape which doesn’t actually hold any value.

In any discussion of data ownership and sharing, it’s not long before the issue of cyber security and IP protection comes to the fore.

“It’s a rapidly changing world, and businesses can’t afford to be complacent,” said Hassan Khalid – data scientist at GlaxoSmithKline.

Technology and platform providers are aware of that need, however, and are very security conscious. A global software provider is almost sure to have more expertise, resource and capability than most other companies, so manufacturers just need to be upfront about their individual business needs and concerns.

If your organisation has to share data, or could create significant additional value by doing so, then don’t let latent fears or uncertainties stop you from taking proactive steps forward.

Morning Roundtable Sessions

The first roundtable I sat in on was hosted by Dell EMC and explored the business benefits the Internet of Things (IoT) can bring to a manufacturing organisation. Listening to the conversation, it seems that many manufacturers are still struggling to correlate how IoT relates to their day-to-day processes.

To help illustrate exactly that, one example given was of an India dairy farm which uses drones and sensor data to map out the most fertile fields, and the herd was moved accordingly. As a result, yields had increased by more than 25%.

Given that another key concern is how to bridge the gap between new technology and legacy equipment, another example given was that of a decades-old Bridgeport milling machine and a similarly aged Colchester lathe that were recently on display at this year’s MACH exhibition.

Both machines had been retrofitted with relatively simple, easy to install, low-cost (around £1,500 in total) sensors to communicate operational and utilisation data in the same way as any new machinery does.

Performance data is one thing; however, you need additional information to contextualise that data. If the average speed of your lorry is 75mph, but sensors show it only averaged 65mph this week, knowing just that isn’t enough. What is causing the 10mph drop?

The challenge lies in gathering enough data to make informed decisions, without having so much data that you become ‘stuck in quicksand’.

My second roundtable was hosted by SAS and explored what machine learning means for you and your business. After going around the table and hearing a dozen different definitions and use cases, it became apparent that there is no one definition, and there’s certainly no one-size fits all solution.

However, rather starting with the technology, a much better starting point is identifying what it is your business is trying to achieve? That will help to determine what data is required, from which machines, at what time, and who needs to receive it.

My final roundtable of the morning covered the cyber security challenges associated with integration operational technology (OT) with information technology (IT), and was hosted by IBM.

The discussion emphasised the importance of having cyber security on your digital agenda from the outset, rather than added as after-thought.

Almost all (99.9%) of business problems IBM comes across are easily fixable, but they can’t be fixed if they go undetected. One manufacturer noted how they had asked a ‘known, friendly hacker’ to the business to try and crack their system in an effort to highlight potential weaknesses.

Many products currently carry a ‘CE Mark’, declaring that the product complies with essential regulatory requirements. ‘In the future, will we see an equivalent or an amended CE Mark relating to cyber security?’, one delegate posed.

Following lunch, Tim Clark – head of manufacturing at SAS discussed how to unlock the value of data produced by machines in his keynote: Can you Talk to the Machine?

You can read The Manufacturer’s overview of Clark’s keynote here.

Panel Discussion #2

Overseen by Jonny Williamson, editor of TheManufacturer.com, the second panel attempted to address one of the most frequently heard questions of the day, ‘What data should I collect and Why?’.

Terex uses a both a targeted and open-ended approach to data collection, according to service and solution director, Matthew Skipworth. ‘Targeted for obvious reasons, but we are finding some very interesting and unexpected trends with the open data we critique,” he explained.

The business has taken an innovative approach to breaking down internal siloes – Skipworth, himself. From the outset, his focus has been to engage all functions and segments, ensuring diversity is represented across all project teams, that communication and information is both structured and frequent, and ensuring that every person fully understand how critical each function is in the business’ overall success.

Terex is also seeing significant benefits from ‘open innovation workshops’, which empowers the entire workforce – particularly those seen as lower in status but certainly closer to the front line – to share their unique knowledge, experiences and suggestions.

Real-time data insights can also strengthen and boost a manufacturer’s existing lean initiatives, noted Marco Del Seta – head of digital at industrial gas multinational, BOC.

“Continuous improvement requires straightforward measures to track and guide it; that correlates with the way metrics can offer a simple method of determining how well the business is performing,” he explained.

“For me, CI is sustainable when the feedback from any changes made is immediate, which requires data to be delivered swiftly, accurately and reliably.”

The big question for many organisations is whether to adopt a targeted or more open approach to what data to collect – with several manufacturers I’ve spoken to extolling the benefits of both methods.

Peter Bozsoki, a data scientist at GE Aviation, noted how his organisation had increasingly moved towards automating data collection – a trend he expects to only increase in the months ahead.

“The more automated the collection process, the better the data quality is,” he explained. “The timelier the gathering, i.e. live satellite feeds versus after-flight data collection, is also beneficial in providing earlier alerts to potential issues.”

Afternoon Roundtable Sessions

My first of the two afternoon roundtables was hosted by IBM and explored why gaining greater supply chain insight has become a primary focus for manufacturers – especially those seeking to exploit new business models, such as servitization.

Many organisations currently operate fragmented, disconnected supply chains which need to become more standardised and unified. However, simply investing considerable resources into gaining visibility will be wasted if you first don’t identify what additional benefit or value such visibility would bring.

One major challenge has been created by those which have outsourced logistics and distribution to third-parties. As a result, the expertise and insight is taken out of the business and yet another supplier has to be interacted with.

My last roundtable of the day discussed how to collaborate with customers, and was hosted by Marco Del Seta of BOC and Philip Woodall of Cambridge University.

‘How do you sell putting a data-capture device into customers’ equipment or facilities?’ asked one delegate. The consensus was by explaining how the information gathered could benefit both parties, i.e. by delivering ‘value’ – probably the most commonly heard word of the day.

‘Could companies afford to not have 100% accurate, quality data?’ another delegate asked. Again, the table agreed that for some applications, 100% accuracy wasn’t a prerequisite and trying to achieve optimum conditions shouldn’t hold you back from pilot projects.

Ahead of the final panel discussion of the day, Paul Brook – data analytics director for EMEA at Dell EMC, outlined The Digital Future, Fuelled by Data and Valuable Returns for Business. 

You can read The Manufacturer’s overview of Brook’s keynote here.

Panel Discussion #3

Industrial Data Summit 2018 drew to a close with a discussion presided over by Oxford Engineering’s Karim Sekkat, with a theme of ‘Raw (data) – what is it good for?’.

Data may be the ‘new oil’, but you wouldn’t put raw oil in your engine and likewise your business shouldn’t attempt to use raw data. Like oil, data first needed to be refined, advised Unilever’s Biswaranjan Sen.

A key takeaway for many delegates was that to gain the greatest value of data, you first must ask the right questions. Though, ‘do manufacturers currently have the necessary skills and knowledge to ask the right questions?’, posed Cobham’s Dr Robert Pearson.

Drawing proceedings to close, the chairman’s closing comment perfectly summed up an insightful, lively day of conversation.

“If you always do what you’ve always done, you’re always going to get what you’ve always got. Digitalisation has opened manufacturers’ eyes to what’s possible and today has been a tremendous catalyst for executives to see what their operation could – and should – look like.

“Some businesses will use digitalisation to squeeze more juice from the lemon, others may choose to switch that lemon for an orange. The most important thing is to get involved. Engaging with and using technology isn’t scary, we’ve been successfully doing it for centuries.”