How data products can drive efficiency in manufacturing

assets/files/images/09_02_23/keepler-pablorios.jpg

By Pablo Ríos, Business Manager for Manufacturing and Energy industries, Keepler Data Tech.

For years, manufacturers have been under pressure to find greater efficiencies. The formula has remained fairly consistent: targets typically centre around reducing costs and increasing quality to defend profit margins and persist in challenging markets.

While this approach has been the hallmark of many successful manufacturers, such a strategy has squeezed margins tighter and tighter, while traditional methods have long been exhausted. As limits have been reached, enterprises have had to become more innovative – thankfully, they now have the tools to do so.

Today, data is powering everything we do – so much so that it’s estimated there will be a massive 175 zettabytes of data in the global datasphere by 2025.

For manufacturers, this presents opportunities. Indeed, data has the potential to be one of the industry’s greatest assets, enabling successful enterprises to thrive in today’s fast-paced and competitive manufacturing arena. 

However, realising that potential is highly dependent on manufacturing firms dealing with data in the right way.

Data projects versus data products

At present, firms of all shapes, sizes and industries – not just manufacturers – approach data with a project mindset. Every time a business function has a problem that it wants to solve using data, the organisation starts from scratch – acquiring the data, cleanse and prepare it, then analyse it for that specific use case.

This is a flawed approach that doesn’t enable firms to make the most efficient and effective use of their data investments. It is often slow, leads to duplicated work, and the outputs from each project typically can’t be repurposed to solve other use cases.

Instead, organisations should be looking to manage data like a product, shifting focus off individual challenges and towards developing frameworks that can be used and repurposed to enable the use of data in solving key challenges on a repeat basis. In other words, they should embrace a product (not project) centric approach to data.

Indeed, data products have the potential to revolutionise manufacturing, offering several ways of driving efficiencies in innovative manners.

With data products, ready-to-use data frameworks can be harnessed at speed to deliver real-time to, for example, identify bottlenecks in production processes, which can help manufacturers quickly identify and address issues, reduce downtime and increase productivity.

As an example, we’ve seen instances where data products have been used to deliver production optimisation for a bottle manufacture, leading to reduction rates in bottle rejection of between 5% and 20%.

Here, machine learning models were created to determine the key criteria for quality in the bottle manufacturing process among hundreds of variables. A decision tree with the ranges of values for bottleneck temperature, blowing pressure and others key criteria were created. Resultantly, by applying combinations of these adjustments, the reduction in rejected bottles was dramatically reduced while quality was sustained.

Further, by analysing data from equipment and monitoring systems, data products can also predict when a machine is likely to fail, allowing manufacturers to schedule maintenance before a breakdown occurs. This helps to prevent unplanned downtime and reduces the need for expensive repairs.

Equally, the real-time element of data products can also help manufacturers optimise their supply chain by providing visibility into inventory levels and delivery times. This allows them to make informed decisions about when to order materials and components, reducing the risk of stockouts and overstocking.

Valuable insights into customer behaviour and preferences are also key. By analysing data from sales, marketing and customer service, manufacturers can identify trends and make informed decisions about product development and marketing strategies.

Identifying an ever-improving opportunity

Across these various applications, data products can provide manufacturers with significant benefits, from improved decision-making and boosted operational efficiency to reduced costs and mitigated machine downtime.

With that said, data products remain relatively novel in the manufacturing space. Why? Because old habits die hard: where manufacturers have traditionally sought and/or developed solutions that address specific use cases (taking a data project approach), this continues to be the avenue that many take. It’s a prime example of the saying, “if it ain’t broke, don’t fix it”.

Critically, however, the customisation capacity data projects reduce the benefits that manufacturers can obtain compared to personalised data solutions (data products). For this reason, it is vital that manufacturing companies change their mindset and embrace solutions which can be implemented through data products that provide a clearer process and enhanced ROI.

Moving forward, it is likely that many manufacturers will begin to head in this direction as the cost of data storage and processing continues to decrease.

As the economy of scale model offered by hyperscalers continues to improve, manufacturers will have a prime opportunity to wholeheartedly embrace data products more readily and cost effectively.

This, combined with the ability of firms to work with partners that have a high degree of specialisation in the use of native cloud services, makes it possible to drastically reduce the operating expenses associated with data products, making them even more attractive.

Culture is critical

Of course, these aspects are only one part of the puzzle. While improved ROI and reduced OPEX will help to get key decision makers on board, a wider cultural shift will be required to ensure data products are implemented and utilised readily in a manufacturing setting.

To instil this shift in mindset, it is important for firms to keep their data practices are up to scratch. That means implementing and/or enhancing key processes for improving data quality and eliminating errors to ensure more robust and reliable models are developed.

To achieve this, manufacturers should first focus on securing and leveraging the right skillsets, technology strategies and partnerships capable of propelling them forward in a relatively novel or unfamiliar space. Equally, they should work on enhancing internal understanding and skillsets, driven by both a willingness from individuals to learn and embrace new skills as well as investment into training from the enterprises themselves.

By moving these critical building blocks into place, manufacturers will be well set to begin developing and deploying data products capable of delivering a variety of transformative benefits. Indeed, those that are proactive in doing so will lead the charge in the sector and unlock vital first mover advantages as a result.

Add a Comment

No messages on this article yet

Editorial: +44 (0)1892 536363
Publisher: +44 (0)208 440 0372
Subscribe FREE to the weekly E-newsletter