Last year when we published our first Advanced manufacturing report, we soon discovered a worrying trend. An astonishing 98% of manufacturers reported struggling with data-related challenges. In this blog, we examine the top data challenges in manufacturing, assess the common factors and explore how we can address them with a focused strategy.
Bridging the gap calls for a change in perspective from merely recognising its value to actively transforming our approach to the way we collect, manage and put data to work.
Here are the top 7 data challenges our research uncovered.
- Difficulty effectively sharing data and insights between teams to achieve common goals
- Inability to identify actionable insights to help employees
- Difficulty ingesting/integrating external data
- Difficulty accessing the data we have internally
- Inaccurate data
- Out-of-date data
- Incomplete data
It’s tempting to go through these and treat them individually, asking and answering the who, what, how of each. A more sophisticated approach is to ask what these 7 factors have in common. What’s the bigger picture here?
With a bit of thought, we can reduce those seven points to just three recurring factors: Data quality, accessibility and utilisation.
Quality vs quantity: problems upon problems upon problems…
The scale of the data revolution is beyond comprehension. Just about everything we do creates data. Every time you pick up your smartphone, the movements are measured by internal sensors. Every keystroke and mouse movement on your computer is measured. Back in 2016 Intel calculated that the average person generates about 1.5 gigabytes of data per day – a 200% increase since 2012. If you own an electric vehicle you can add an extra 3 – 4 terabytes of data every day.
Electric vehicles generate vast amounts of data every time you start the engine
With figures like that, and a 200% increase in 4 years, we need to stop thinking of data as numbers on a screen. It’s fluid, it’s dynamic, evolving and growing at an exponential rate. If you’ve got a problem with data quality, every second that goes by your problem is compounding. You’re piling problems upon problems.Timeliness is key. We’re seeing real-time data collection, but manufacturers need real-time processing systems too, otherwise you’ve got a bottleneck of monumental proportions.
For many of us, the data we collect represents a snapshot in time. It’s like a photograph of an object, a product, a process, whatever it might be. Once we’ve captured that point in time it then aways and inevitably recedes away from us.
So has the big change in data been akin to the progression of still-frame photography to film and moving images? In a word: no.
We’re already well past that point. The development has been far greater and that’s where our problems stem from. We are underestimating the scale and magnitude of the change in recent years.
A better analogy would be the change from still frame images to the fully immersive 3D worlds of the latest video games, realistic, detailed 3D worlds that withstand the highest scrutiny.
Navigating the changing data landscape
We can repeat ourselves over and over again, talking about centralised data systems, cloud adoption, integration tools, AI, all those things. That’s great. It’s part of the solution. But to really navigate the changing data landscape, we need a change of approach. This means rethinking how we perceive and use data across the organisation.
Recognising data as a strategic asset is a given. We must take that idea a step further and remember that business assets are not all static. Some are dynamic. Data falls into the second category. It changes and evolves in an almost organic way. Its sources, volume, velocity, and variety are constantly shifting.
The fundamental change we’re advocating emphasises a proactive approach. Instead of being a passive resource, data becomes an active part of our strategy. To tackle the seven data challenges, we must first see them as symptoms of a bigger problem. That’s the exponential growth in data collection and quantity.
As the volume and variety of data we collect grows, organsations face challenges in managing this vast amount of information.
All this data, however, is crucial for developing learned models. These are models which use historical data to recognise patterns and then extrapolate them into the future with a high degree of confidence. That’s what we mean by moving from reactive to predictive capabilities.
For example, predictive models can forecast when a machine is likely to break down, allowing for pre-emptive maintenance rather than reacting to a breakdown after it occurs.
The ultimate goal, perhaps the future of manufacturing, is to achieve fully autonomous capabilities. This is where systems go beyond predicting events to suggesting the best actions to take. They can make decisions and act independently, without human intervention. A fully autonomous, prescriptive system would not only predict a machine breakdown but also automatically schedule maintenance to prevent it.
This transformation turns data into a dynamic asset that drives future success, allowing organisations to overcome the seven data challenges and harness the vast quantities of data as a strategic force for proactive decision-making.
