Data lessons: a review of Industrial Data Summit 2023

Posted on 27 Jun 2023 by Joe Bush

The sixth annual Industrial Data Summit took place at Villa Park, Birmingham, on 26 April where the UK’s most senior manufacturing data professionals came together to talk data and analytics in their businesses.

In the opening keynote of Industrial Data Summit, Rob Clifford, Head of Digital and Data, BAE Systems, spoke about applying lessons from digital transformation in the public sector to the manufacturing industry. The manufacturing sector is the latest domain to embrace the digital and industrial data transformation ethos. While there is ground to make up, there’s also huge opportunity to accelerate change by learning from successes and failures in other sectors.

The Manufacturer spoke to Rob around the topic of cyber security within manufacturing. The full interview can be seen in the Jan/Feb issue of The Manufacturer magazine. We often talk around the importance of people here at The Manufacturer and the second keynote of the day saw Kat Dixon, Analytics Business Partner at Babcock International, discuss the realities of creating an industrial data-driven culture, warning that businesses often have a tendency to dive into the technicalities of digital transformations.

However, no amount of new technology is going to work without first having the right foundations and organisational culture. A full interview with Kat can be found in our Mar/Apr issue. In this article, however, we look at the findings from the summit’s eight discussion tables, where the table hosts share the popular discussion topics of the day and their key takeaways.

AI and Machine Learning

Tom VaughanTom Vaughan, Solutions Engineer, Peak AI

Four key points came out of our discussions. Number one was to identify the process and see where it can be improved. In order to get to an AI and machine learning base, you first need to understand the process you’re trying to improve; establish what the challenge is across the business and the process improvement that can take place.

Number two was industrial data strategy and systems. What is the quality of data and systems you currently use? Are they interconnected or are they siloed? And do you have the foundation layer to be able to bring in an AI or machine learning model?

Quite a large part of the discussion, spoken about passionately, was the change management side of data. When bringing in AI and machine learning, it’s important to remember that it’s not there to replace people, it’s there to enhance people.

We heard some great examples of where machine learning could be deployed within a business. The best results have been where machine learning perhaps performs the algorithm and builds the model, but then allows people to make the final strategic decision.

Finally, topic four was around the AI and machine learning models themselves and how to get an effective AI and machine learning model within your business. That involves the three previous elements – process improvement, having a great data strategy and change management.

Data Strategy

Jennifer-BelissentJennifer Belissent, Principal Data Strategist, Snowflake

One of the big issues that we discussed was what is a data strategy? We started with the idea that perhaps an industrial data strategy doesn’t exist on its own – it’s an element or an enabler of a larger business strategy.

So, when thinking of wider business objectives, manufacturers need to ask what they are trying to do as an organisation; whether it’s to reduce cost, increase revenues, improve customer experience or reduce risk. What needs to be done with data then flows from that, enabling the whole data value chain from data use through to investments in the pillars of people, process and technology.

It’s not just about having a data strategy that encompasses technology, it’s about the people and the processes that we need to set up. Therefore, we also spoke about where you start; a number of people highlighted the fact that they had a data strategy but were unsure of how to implement it.

We discussed how to identify data champions; whether it’s from parts of the organisation that are more data mature, or people within the organisation that have an affinity to data and really want to experiment and innovate. From there broader communities can be developed, creating a snowball effect. These champions should be given the opportunity to share what they’ve done and to showcase their work as a means of professional development.

We also spoke about what else is required to scale i.e., the communications element of a data strategy, and how you spread the word on what data is in the organisation, so everyone understands their role; whether it’s capturing, protecting or using data. We spoke about getting more people involved in this data-driven culture.

The ‘D’ in chief data officer doesn’t just stand for ‘data’, it stands for diplomacy as well, because it’s about bridging gaps within different organisations and rationalising the requirements across the business.

Data Governance

Niall BuswelNiall Buswel, Procurement Data Lead, Jaguar Land Rover

It was very valuable to hear what everyone collectively has been going through and it was reassuring to know that we’re generally facing the same challenges.

Even though the main topic was data governance, we also spent a lot of time talking about data quality and how organisations are approaching that. As well as discussions around ownership and stewardship of data, we also spoke about the standards and policies around data governance and how we get business buy in, throughout the organisation, from the top floor to shop floor.

Another theme focused on delivering value. You can have all the best policies and data quality metrics in place, but how do you showcase that to the senior leadership team and the wider business? How is it impacting day-to-day operations and how are you delivering value for others throughout the organisation?

IoT and Sensors

Matt MouldMatt Mould, Partner, Storm Reply

We learned a lot about where people are in their journey with IoT and the good news is there are similarities in terms of where everyone’s at. One universal complexity seems to be a wide, disparate collection of machines that have to be understood to satisfy leaders that a plant is operationally efficient, and that energy is being tracked. The challenge, however, is that a lot of separate vendors own the machines that need to be changed.

We also tackled the issue of talent attraction in this space. Manufacturing was referred to as the backbone of the country, yet it is operating in the background. Therefore, attracting the talent to compete with other, more recognisable industries, is also a challenge.

We also debated why we extract data from machines and why we integrate data from other systems? But these questions need to be answered in the context of the business case to enable investment in this space, particularly with IoT and sensors. Because the impact on the bottom line isn’t always immediately obvious.

Data Collaboration

Sean Robinson, Manager, Software Solutions, Novotek UK & Ireland

Our discussions kept returning to working backwards from goals that matter to leadership, and then translating that to things that are very specific. It’s a lot easier to get people to collaborate around something that’s understood (such as launching a new product, changing a product definition or even building a new factory), than it is to collaborate around industrial data for a more nebulous goal that may be seen as detached from day-to-day business operations.

Some of the cultural challenges around that collaboration relate to people feeling a sense of ownership and protectiveness for their system. It was interesting to note how few people were able to say that they had clear data owners.

That links back to keeping data specific; know what you’re doing in support of the organisation’s goals and strategy, but then understand how collaboration has to happen. From there you can create a culture that’s more capable to think about specific improvement initiatives or more detached exercises, because there will be a higher level of trust.

Security was supposed to be a big part of our discussion table. I would argue that what we ended up concluding was that more often than not, the technical security aspects of enabling collaboration, even around sensitive data, are less of a barrier than culture. You need people who are domain experts in what really matters.

Overly rigid rules can be enacted if no one truly knows what is safe and what should be available for both internal and external collaboration. Therefore, the technical aspect of data can play a secondary role to an overzealous culture.

Predictive and Advanced Analytics

Richard JeffersRichard Jeffers, Founder and Managing Director, RS Industria

There were four recurring themes on our table. The first was around the importance of starting with the problem at hand rather than falling in love with the technology. It’s not about deploying data science for the sake of it.

Linked to that is the fact that you can do a lot with some relatively unsophisticated tools – statistics can take you a long way before you need to move into AI and machine learning. To solve any of these problems, you need to bring together the people who understand data and the maths with the people who understand the domain; you’ve got to have people who are embedded into the problem, as well as people who are embedded into the technology.

The most important discussion point was around culture change because people have to respond to the alerts and insight that come from the data and use that to drive improvement activity in the business. The maths won’t actually deliver any business value without cultural change.

Data Collection and Integration

Mike KierseyMike Kiersey, Head of EMEA Technology Organisation, Boomi

We touched on just about every aspect of data as part of our conversations, and there was some great sharing of best practices and ideas, particularly around keeping things simple, starting small and growing from there. Businesses can go quite broad with data, very fast, but that won’t necessarily yield the desired results. Much like the second keynote address, we also touched on culture and data literacy.

Do people really understand why they’re making data-driven changes and who is leading the way on that journey? Do we really know what and how much data we need and how long we need to keep it? And how do we make it easy for people to access, organise and use?

Data Leadership

Phil WarePhil Ware, Digitalisation Manager, Edwards Vacuum

We looked at building a strong data-driven culture, implementing robust data management practices and developing a strategic approach to leverage data to achieve business objectives. I think the last one answers all the other questions.

If you know where you’re going, you know what you’re looking for and you can communicate that. That drives your data strategy all the way through the business. Everyone understands what you’re trying to do and why you need the data, and they want to find out what data you need.

Cross collaboration was a common theme; getting people who work in different areas to understand data from different parts of the business, creating a data flow. Knowing what to do with a piece of data and how it impacts the customer is critical. Leadership was also highlighted; the senior team need to drive down. If not, the shop floor can still drive back up. Get good information and feed that upwards to drive culture change at the top.

Rob CliffordRob Clifford, Head of Digital and Data, BAE Systems

Being a data leader doesn’t mean you do it all yourself. It takes bravery, intellectual curiosity, but doesn’t require you to be a data ninja. It’s more important that you create the conditions for other people to succeed and give them the cover that they need from senior teams if it doesn’t go the way it should. That’s what the real leadership part is. And it may be cliché but those are the leaders we all enjoy working with.

For more stories on Industrial Data and AI click here.