Why and How to Use Big Data in Manufacturing

With COVID, labor shortages, and compromised supply chains in full swing, the average manufacturer finds themselves grappling with one of two paths to follow. They can either A: pare back and attempt to save money – adopting a more lean manufacturing approach that’s optimized to only deliver what they know will bring in profit – or B: they can invest in new technology such as industrial grade PCs and big data in manufacturing practices that can help draw more profit in the form of increased productivity.

Both of these approaches, in truth, are valid. Manufacturers find themselves in unprecedented times. Both playing safe and taking risks can sound enticing to different factories based on a number of variables. And we’ve already covered several avenues of approaching option A when we discussed the top 10 ways to reduce overhead costs in manufacturing and lean manufacturing best practices.

Now, the time has come for us to reach out to those interested in transforming rather than hibernating. Now, the time comes to make a case for and explain how to use big data in manufacturing. 

Why the Aversion?

Big data in manufacturing essentially boils down to what many have already discussed in regards to industry 4.0. It’s taking data drawn from machines and sensors to dictate decision-making on the factory floor. The main philosophy behind the practice is that, with more data on how and why their machines run the way they do, decision-makers can make more educated choices in order to improve efficiency, cut waste, prevent breakdowns, yada yada yada – you’ve heard it all before. 

This sounds good on paper, but aversion still remains due to the myriad of issues we mentioned earlier being felt across the manufacturing sector.

 Another problem also contributes to the lacking adoption of big data in manufacturing: a lack of awareness as to just how much proper implementation can save, and even make, manufacturers. 

Furthermore, many believe that starting with big data in manufacturing means starting from scratch when in reality, they likely already have an abundance of data they can use to improve their decision making. According to a senior executive from ABB, a notable Swiss engineering firm, at least 80% of data already being generated across the average factory floor isn’t being utilized. Even more frustratingly, he explains that, depending on the sector, 30-40% productivity gains can be achieved if that data was actually used to inform company policy/practices. 

Why Use Big Data in Manufacturing?

To tackle the issue of not understanding how much proper use of big data in manufacturing can enhance a business, below are a few key takeaways to consider. Staying away from the obvious benefits you’ve likely heard several times before, we’ve tried to cover a few ways big data can help your business get ahead of more recent trends we see the industry as a whole moving towards. 

Improved Product Design

With supply chains completely uprooted, many manufacturers are beginning to design their own products and sell directly to consumers. Unfortunately, product design is a lengthy process. The amount of iterations that need to be created before you can even build a prototype can often be discouragingly expensive as well.  Every failed iteration, though necessary to create a product that is top-tier, requires time, resources, and manpower to create. Fortunately, many have started to blend big data into the product design process. How is this done? By unlocking the potential of that 80% unused data mentioned by ABB’s executive.

Taking data from past production processes, manufacturers have been able to run simulations for products across every stage of their development. Before, simulations were only saved for later stages such as finalized prototypes since running them was often costly and time-consuming. Using machine learning algorithms trained on data from these past productions, however, these simulations are streamlined and performed much quicker. This means every iteration of a device can be run in simulations which means a more meticulously tested, quality product. 

 Industry Week dives a little deeper into specific examples of how this is done and is a great read for those looking to convince higher ups of implementing big data in manufacturing.

Its Usefulness Compounds the Longer You Use It

The wonderful thing about big data in manufacturing is that it’s purely focused on taking past data and experiences and using them to enhance current practices. Taking the example of data being used for product design, the only reason that use case is so effective is because manufacturers and ML programs have data from past processes they can use to streamline simulations. 

That said, the longer you use big data infrastructures to gather data, the more data you have, the more you can train ML and other AI applications to perform key tasks quicker and more effectively. Thus, data gathering becomes more and more potently effective the longer you use it, meaning you can stand to gain in the short term and then gain even more in the long term. 

You’ll be Reskilling a Workforce that Will Need to be Reskilled Anyway

The skilled labor shortage in manufacturing is only increasing as more and more become unable to work. Manufacturers also continue to compound the issue by adding more tech into their operations that raise the entry level requirements of new employees. This has resulted in a renaissance for reskilling the workforce as business owners start to train employees they already have to meet new needs.

 By slowly implementing big data in manufacturing practices, you can expose your employees to smarter tech and industry 4.0 hardware in an approachable way. Odds are you’ll be needing to reskill your workforce regardless of whether you do it now or after this transformative period in manufacturing’s history. Starting now may be what puts your business ahead of the curve when the smoke clears. 

How to Use Big Data in Manufacturing

Of course, now that you’re gung-ho on big data in manufacturing, you don’t want to blow your budget on an entire program you aren’t equipped to use. So what does the process of slowly working it in look like? Here’s how to use big data in manufacturing in an approachable way that will provide all of the benefits we mentioned earlier. 

Educate Your Team on its Benefits

First and foremost, there’s no big data in manufacturing if your team isn’t on board. Remember that several manufacturers still belong to the camp of “big data and industry 4.0 is superfluous and we don’t have the money right now.”  

Convincing these people, with their very valid reservations, requires educating them on the benefits, the profitability, and, most importantly, the future-proofing of your business caused by big data implementation..

Start Small

Don’t scare yourself into thinking Big Data needs to be a plant-wide transformation. Start small by adding sensors to a few devices and preparing a data storing center. 

Plan ahead and deploy workstations such as rack mounted pcs or HMI interfaces that can draw in this data and support software/hardware needed to analyze it. Start small and prove the concept of big data in manufacturing has merit before committing 100%.

Target a Specific Issue

So, if you’re starting small, where should you deploy the few sensors you plan on deploying? That depends on you! What’s an issue your floor has been dealing with as of late? Does a certain machine commonly yield product that doesn’t meet your standards of quality? Does production often slow down during a specific stage?

Consider implementing big data in manufacturing in trouble areas where you can test if these new programs can actually solve problems and not become another application of IoT wallpaper. 

Learning How to Use Big Data in Manufacturing Enhances What You Already Have

In Peter Terwiesch’s interview that we mentioned earlier, the engineering firm president perfectly encompassed why big data in manufacturing has become so enticing in recent months. 

“When you can’t get what you want, you have to get better at using what you have. So if you can’t get people to site… you have to get better at using what you have, which is your data, which is remote connectivity.” 

Data is something you’re constantly producing whether you know about it or not. Every success, every failure, every process is a point of data that can be referenced either to recreate preferred outcomes, avoid negative outcomes, or even create all new positive outcomes. Whether you decide to use this data is up to you. However, if we were a betting group, we’d say the average manufacturer is deciding to use their data to improve their production. If you don’t want to fall behind in adopting your own untapped data, contact an expert from Cybernet today for recommendations on the hardware required to get you started.