3 Best Practices for Harnessing Big Data Analytics in Manufacturing

| June 21, 2016 | 

Quick, define big data and give a few examples for how it can be used to improve manufacturing operations. If you’re struggling with the answers, you’re far from alone. For example, the detailed Wikipedia entry on big data says that “big data is a term for data sets that are so large or complex that traditional data processing applications are inadequate,” but then goes on to note the term is often used interchangeably to describe predictive analytics that may involve smaller data sets. But big data analytics in manufacturing can be a little complex in how to make sense of the loads of data located in different systems across the organization.How it works - big data in manufacturing

Since the media hype on big data is usually focused on consumer applications, our goal for this blog post is to:

  • Demystify what data big data means to manufacturing
  • Explain how you can use big data analtyics in manufacturing to improve operations
  • And give you 3 keys to a successful big data strategy.

Big data in manufacturing environment

Before delving into how manufacturers can tap into the value of big data, let’s take a step back and consider the various types of systems that are typically employed across manufacturing operations. After all, your existing systems are the foundation for what you will be able to harness and analyze as you pursue a big data strategy. In our client engagements, we usually see a mix of a material requirements planning (MRP) system; SAP, Oracle or even BPIX ERP systems; computerized maintenance management systems (CMMS), Quality Management Systems and a lesser extent we see manufacturing execution systems (MES). We are also increasingly seeing more manufacturing tools and equipment with features and capabilities for enhanced connectivity to business systems.

Going back to the Wikipedia definition above, these environmental considerations have important implications on what “big data” could actually mean to your business in the near term. For example, if you are employing an MES and a lot of smart tools and equipment, then you may be closer to the challenges of leveraging especially large and complex data sets based on real time information coming in from equipment and processes. Otherwise, you mostly need to be thinking about using analytics or business intelligence (BI) tools to uncover hidden insights and better capture and extract maximum value from all key data that is related to a processes you’re trying to improve—while keeping a close eye on your desired future state.

No matter what, it’s important to be working on an operations big data strategy because providing the right data to the right people at the right time really can drive the types of efficiency gains and process improvements that are critical for continually improving (or at least maintaining) margins and carving out a competitive advantage.

Show me the efficiency gains

How does a 3 – 5% improvement in gross margin sound? It’s definitely nothing to sneeze at, and we’ve seen these levels of improvements in clients that have adopted appropriate, comprehensive strategies to gathering and analyzing data in their operations alone.

When you consider the complexity of processes across most manufacturing operations, there are countless opportunities for finding efficiencies. For example, one of our clients kept experiencing one or two day outages in a key process that would grind production to a halt for a day or two. When the client finally analyzed maintenance work order data against outage dates, it turned out it was due to a key machine component breaking roughly every four months. And once the client knew the issue, they had the information they needed to either figure out the cause of the break or at least replace the component prior to its regular breakage time frame.

In other words, manufacturers often have the data they need nearly at their fingertips, the problem is that they don’t have a way to analyze the data to pinpoint what may be causing a trend. That’s where the big data analytics strategy and software, comes in.

3 keys to a successful big data analytics strategy

Since there are so many possibilities for using big data in manufacturing (typically a complex environment), simply figuring out where to start can be daunting. Here are three best practices that are key to an effective approach.

  1. Put people first

    When it comes to gathering and analyzing data, the technology is obviously important. But ultimately the value of technology has a direct correlation with the skillsets of the people using the technology and analyzing and taking action on the data. That means you may either need to bring in people with the right knowledge and mindsets for tackling big data challenges or make sure the solution you choose will be able to deliver meaningful data to your different levels of users.

  2. Set clear goals

    The most important question to continually ask as you develop a strategy is, “what are we really trying to accomplish?” This question leads you down a path of which of your systems may contain useful information and what types of operations or supply chain data you need to be capturing and analyzing to realize goals.

  3. Start with lost throughput or production

    Since there are so many potential opportunities there’s no need to overcomplicate matters. Start with areas where you see know there are issues, such as lost throughput or production, but you don’t clearly understand the trends and causes.

Gaining perspective

Big data isn’t an easy fix to manufacturing challenges by any means. Success requires careful planning and strategy execution along with ongoing fine tuning. Solution choice, like Dploy Solutions, is a critical part of the equation because your strategy is only going to work if your team gets the information they need at the right time to make proactive and confident decisions. When approached correctly, the benefits of your solution and the insights that big data can uncover in manufacturing operations will pay for themselves many times over.