With all of the hype in the marketplace about analytics, many manufacturers understandably have the impression that all you have to do to get actionable business insights from an analytics tool is “plug in” the software and let it work its magic. The reality, however, is more nuanced. Analytics are a wide set of capabilities and there are some key steps that need to happen behind the scenes before end users can start maximizing value from dashboards. This blog explores three of the top considerations for getting the most value from analytics tools in manufacturing operations.
Measuring the right data
Recently CBS’s 60 Minutes ran an interesting piece on a U.S. satellite company that uses a large fleet of bread-box size satellites to capture the entire land mass of the earth on a daily basis. One of the more eye-popping facts in the story was that the company would need to hire more than six million analysts to make effective use of all of the data it’s gathering.[i] In other words, even with all of the sophisticated analytics tools the company is probably using, the quantity of data is enormous and analysts still have a very hard time keeping up.
It’s a dramatic example, but it brings up a great question for manufacturing operations that also have a lot of potential data points: should you measure everything and figure out what’s important later or is it better to build dashboards around the data that you know is important? Generally speaking, the measure-everything approach rarely makes sense and would require a much deeper bench of analysts and a significantly larger budget to work. For the vast majority of manufacturers, the focused approach is much more practical and effective when getting started with analytics. Some key questions to ask when narrowing down what data to gather and analyze with a new analytics solution include:
- Where do you need focused visibility in operations?
- What equipment, lines or shifts seem to be causing grief during production?
- Are there any areas of the business that seem to be under-performing on daily, weekly or monthly basis?
- Where do you intuitively believe opportunities lie?
Put another way, the best way to get started with an analytics solution is to keep a narrow focus on the things that keep you and your team up at night. So, for example, you might want to build a dashboard that provides you as much visibility as possible into a particular line or piece of equipment that’s very sensitive and hard to control, or one that has a history of high maintenance or poor quality
Cleaning and prioritizing data for review
Collecting data for analysis involves more than simply pulling data from various sources. Your team must also do work behind the scenes to ensure that data being analyzed is clean and free of issues that could muddle results. Moreover, you also need to consider whether current or historical data is going to provide you the best insights into a question or issue. For instance, to look into specific issues related to a piece of equipment or line, you would typically want to do two things:
- Set up a dashboard that provides a real time or near real-time view of KPIs that you are trying to keep on track
- Analyze historical data to see if there is anything in your area of interest that seems worth further investigating.
Without analytics,it is often difficult or even impossible to determine the cause of anomalies that pop up and then disappear in operations. With analytics, however, it’s not only possible to spot and analyze anomalies in real time, but you can also determine root causes and contributing factors to prevent future recurrences. And sometimes the historical analysis can produce more immediate value than the ongoing real time monitoring. For example, maybe a plant that relies on several preferred materials suppliers and a couple of backup providers had a particularly quality problem that was never fully understood. By looking at the historical data, you might be able to tell if there was any correlation between who supplied the raw materials and production quality.
At the same time, when choosing data to track and analyze it’s also important to focus on things you can ultimately control. A company tracking the quality of products coming off a line against weather conditions might discover a relationship, but they wouldn’t be able to do much about it. While the impact of external events may be interesting, only things where you have a button to push or a dial to change to make corrections to a process really matter. And that relates closely to the next critical consideration around who should be getting different types of information and how.
Developing role-based dashboards that create value
When it comes to getting value from dashboards, customization by audience is essential. After all, dashboards with irrelevant data for a given job role are simply a confusing distraction that reduces productivity instead of improving it. That’s why the level of information in a dashboard should be aligned with the role and the viewers’ ability to do something with what they are seeing. Read about how you can create an analytical culture at all levels in this recent blog post. An executive looking at line performance, for example, probably won’t understand the context and why it’s running the way it is. She needs much higher-level information related to revenue, profitability and how the company is tracking against strategic goals. Conversely, line supervisors need insights about machine or the production line that they can act on to solve issues and keep operations humming. That’s not to say lines or other lower-level supervisors and managers never need to see bigger picture information. It often makes sense to share data about a line’s contribution to the bottom line, for example, but the data needs to be put into context that makes sense for the people that are consuming it.
Putting it all together
Getting dashboards right may sound complicated, but it’s really just a matter of smart planning. The good news is that you don’t need to worry about putting sensors on everything and getting more and more data. A strategy that helps you focus on what’s important from existing data sources and filters out the noise will provide better returns in the near term and set you up for success with more involved data collection and analysis down the road.
Dploy Solutions business process management software is an ideal platform for getting started with analytics. It includes powerful embedded analytics capabilities from Sisense, a leading analytics solution provider. Moreover, you can rely on our manufacturing and operations management expertise for help understanding what data will be most valuable to different levels of your operations based on business goals and customizing your dashboards accordingly. Ready to get started? Contact us today.
[i] Private Company Launches “Largest Fleet of Satellites in Human History” to Photograph Earth, 60 Minutes, January 2019.
About the Author:
Brian Tilley, Managing Director of Technology Solutions at TBM Consulting Group, leads and develops TBM’s technology and service offerings which includes Dploy Solutions. Brian’s background is in leading teams that develop and implement Industry 4.0, IIoT, analytics and digitization projects. Over the years he has worked for IBM, Accenture and within the manufacturing industry with companies such as Vulcan Chemicals.