Analytics Factory of the Future

Posted on 9 Mar 2022 by Lanna Deamer

As the UK looks to pave its new way in the world, external shocks and pressure resulting from climate change and tightening supply chains has pushed the productivity of UK manufacturing into sharp focus.

In order to be competitive on a global scale, manufacturers must turn to digital capabilities to improve productivity, increase innovation and find all important marginal gains and often the Factory of the Future is posed as an answer.

However, in a landscape of siloed data, unconnected systems and skillset scarcity pressures how can such a nirvana be achieved?

A recent virtual roundtable hosted by The Manufacturer and Tableau, a business intelligence and analytics software company, saw the manufacturing community come together to discuss further.

If you are interested in attending a TM Directors’ forum, register your interest here.

Key pain points for manufacturers:

“Data is the future of digital transformation and we have a wealth of data and insights within our factories, but very little of it is elevated above factory. We do very little on a regional basis or indeed get factories talking across factories.”


“I’m Head of Data Strategy at my company, we sit as part of the central IT team with a broad scope to support HR analytics across all functions, including manufacturing and supply chain. We have got legacy and new facility challenges such as making use of the plethora of data that’s already there. And then also, how ambitious should we be in setting up new things? There’s lots of options out there – should we go after all of them or are some of them a bit of a fad?”


“Before you even consider the importance of the data and leveraging it, you need to consider what data to store, why you need it and what you’re going to do with it, and therefore, figuring out the best way of storing it. It almost seems to be the most important element before you can work out how to turn it into useful information.”


“How can we get the people to trust the data, especially when deploying it on traditional areas of work. The people on the shop floor have tried and trusted experience that says they know how to solve problems. And a lot of issues arise around getting them to agree that it’s the right way. That trust and understanding is key and goes beyond being just about analytics.”


“In my role, I have a strategic initiative underway, which is data analytics. However, I’m trying to get the business, particularly supply chain, to create a roadmap for what the factory of the future looks like. The danger is that they all come up with great proof-of-concepts, but the problem is figuring out how it all fits together from an end-to-end perspective.”


“One of my responsibilities is to the lead the data strategy across the business, working with all departments. We could run the factory on an Excel spreadsheet (and many do actually try) but of course, there’s a lot of challenges with that, because everybody that has their own copy of the data, has their own interpretation of the data and will display it how they choose to tell their story. It’s not their fault but that’s when the trust issue starts to creep in. We’re relatively early on that journey – moving on from very manual gathering of data.”

Key questions:

When getting data shared across systems how do you build a valuable big picture?

Data is of course different between different sectors, and if you look at the connectivity across a cell, line, or a whole facility, it does vary and, in some places, it can be quite immature. However, if you’re speaking about value, the other secondary and tertiary datasets that you start to involve, you start seeing other patterns. For example, you might begin to see that you had dips in quality on certain lines when introducing freshly qualified technicians that are just coming off their apprenticeships, or you can start to see patterns with environmental data.

Discrete automated manufacturing organisations who have increased or solved quality issues have started to augment some of that secondary and tertiary information. While there is value in looking at the cell and line level and having it across a specific facility, there is also value when you start to bring in that higher level and bring in some of those other data points. You can start seeing patterns as well. It is not to say there aren’t differences because there are, but once you put a level of abstraction on it, you can start to see these patterns. So I think there is value there.

Manufacturer insight: “A key point to focus on is what the data is telling you, rather than what the data itself may be.”

“Currently, our company is looking into how we could integrate our supply chain, which are currently very complex with lots of companies that feed through to us, so it’ll be a huge job. But there will also be huge benefit to having the data flow through to us from suppliers and having that relationship sped up by the data connection and the analytics of that data.”

“If you abstract the data up, a big percentage of the cost is internal and external in the supply chain, and some of those key level metrics at the top do vary across different sectors. Data can be abstracted to support those important commercial KPIs where you can begin to look at the different facilities to work out where you’re perhaps underperforming or excelling.”

“In my experience, you can never have too much data – especially if you’re looking into why things are performing differently to how you expected. We’ve had plenty of examples where environmental conditions within the factory have changed the behaviour of some of our products. We hadn’t noticed this until we started correlating data with the information we were getting from the factory environment sensors. We’ve also had scenarios where the same material from two different suppliers has resulted in different properties of the products that we’ve produced.

“So providing that you have the right analytics suite, having all that richness of data can be valuable when it comes to diagnosing the end product of the production line.”

How have other organisations maximised the inherent numeracy and creativity of science and engineering teams in an analytics and insight context?

In terms of utilising skills broadly, engineering teams within many organisations are often relatively untapped. Data leaders rarely see data projects that scale across the enterprise – it tends to be a lot of proof-of-concept solutions and a lot of heavy data processing happening in those areas. Scaling data and analytics across an organisation often comes from very different areas, and most data usage will come from R&D teams, as well as finance teams.

Data is really hard. Most people in various organisations aren’t data fluent, and they don’t know how to code so the more you can make it accessible and humanise data, the better your results will be. There are various techniques from businesses trying to humanise data whether that’s been telling stories with data or elevating use cases and video content in order to increase the demand for data. Once manufacturers do get that demand, it’s important to then follow that up with the supply of data for it to be readily available.

It’s going to take a mindset shift; companies are often reluctant to trust their employees when it comes to making good use of data. It’s all about handling the demand and supply of the data at the same time – you need to be supplying governed, accurate data that people can make decisions from at the same time. It’s important to be stimulating demand for data from various areas of the organisation.

Manufacturer insight: “This is what set us up on our self-service analytics journey. We have one extreme to the other where it’s almost a ‘free for all’ on people’s desktops with Excel and then at the other end, we have very fixed reporting methods that churn out an automated PDF which is no use to most people because they want to deep dive with more detail.

“We need an agile tool that will allow people to look at disparate datasets, potentially enabling them to draw their own conclusions – there’s a lot of value in this.”

“We’ve all gone through that process of having extremely capable engineers who fill in spreadsheets and maintain spreadsheets to produce reports. A few years ago, we appointed a chief data officer, who was responsible for standardising the reports and the ways we generate the data and visualise it. Not only has this made the decision making that we can take from data far easier, but it’s also actually saved us a lot of money. One of the best pieces of advice I can give people is to do something similar.”

“We’re currently working hard to educate our end users on data terminology. This is vital because the danger is they’re unfortunately not understanding their own data. Our principle is to go down the self-service analytics route, and my goal is to source the data, load the data, make sure it’s clear and then let the end users do what they want with it.”

What approaches can I take to try and offer beneficial intelligence to processes whilst minimising emissions increases caused by adopting them?

Blockchain is exciting and it’s a hot topic that we all get enthralled by, but the reality is that until quantum comes online, the current platforms won’t scale. And the way they’re scaling currently, they will consume more and more power so there’s a catch up of the tech that has be happen. However, there are lots of datasets available that nobody is getting value from.

The right framework needs to be put in place to allow individuals to get more out of their existing datasets using technologies such as the ones Tableau offer, so that incremental gains can be made from what already exists, without sucking the data out of everything and trying to use blockchain as the solution. It will be a balance with the current trajectory, as some of the current technology stacks that exist are not going to scale, they are going to use more power than they will save.

Manufacturer insight: “We are in the same position with regards to data analytics at the moment -gathering data from multi-section sites in a common format that we can use is really critical. However, we keep using the word ‘data’ but I keep using the word ‘information’ because it’s easy to get the data, but it’s difficult to get good quality information from it.

“A situation arose recently where we had capable engineers working with capable developers across multiple sites, creating lots of valuable information using different platforms. This is an incredibly hard thing to try and reverse because there’s no standardisation across the software and platforms. Another one of the significant challenges is creating a whole new physical network that is able to manage that volume of information. It’s hard to get all the information in a common format that can be used in a way that everybody understands.”

How do we build trust to enable the sharing of data across the whole supply chain whilst maintaining competitiveness?

There are three key points:

  • Using the data to achieve the visibility required.
  • Having the right commercial models in place to incentivise that change.
  • Having the tools to allow for collaboration.

Manufacturer insight: “When we’re talking about sharing data within the supply chain and making sure its beneficial, we’ve suddenly got a position where all the data has value. There is going to be monetisation of the value of that data. So there’s going to be a whole new dimension, will it make you competitively advantaged or disadvantaged.”

“During the roundtable discussions at last year’s Digital Manufacturing Week, we had some really interesting conversations around this and one of the themes that came out from it was the fact that there’s visibility in some sectors which are more consumer driven.

When you start talking about net zero, and you start talking about the entire supply chain, and the entire environmental impact of the product end to end, you must share that data in the supply chain. And, that might drive the monetisation, the competitive advantage of a supplier who would be willing to share the information for the purposes of environmental impact and net zero. But very soon the expectation from the consumer will be that all manufacturers will be able to make very confident claims on the full end to end picture the environmental impacts.”

“Senior sponsorship of data and digital initiatives are critical to rolling out data analytics at scale across an organisation. It’s vital to change the mindset of everyone within the organisation – making people see that data is a fundamental capability is key for the future of data going forward.

“Large manufacturers often have big graduate intakes each year and the graduates often form communities within these organisations. Manufacturers should be tapping into those young graduates who already have digital skills. This will be beneficial as it will enable them to stimulate lots of ideas around the business.”

Summary

Important takeaways from the discussion clarified that everyone is on the same journey – regardless of individual sector, the challenges are very similar. The challenges that arose several times in the discussion are mostly human in nature. No matter what the data is, or what technology is being used, its clear to see that the problem lies with the adoption and upskilling, but it was positive to hear how the manufacturing sector is accelerating change more than ever before.

Another element that was emphasised across the evening was trust. Although data is the future of digitalisation, trust needs to be gained. And data comes with great benefits, but it also comes with great vulnerability.

It’s well-known within the manufacturing industry that we have an ageing workforce, but the younger generation are the solution. Manufacturers can use them and their digital skills to harness the power of the older generation.