Michael Hickins | Content Strategist | Nov 3, 2023
Manufacturers of every stripe—as varied as aluminum and steel producers and makers of electronic components, aircraft engines, and chemicals—use data analytics to help their factories run more smoothly, track supplier performance, increase the rate of perfect orders, identify supply chain bottlenecks, improve employee productivity, reduce product recalls, and ultimately cut costs and boost profits.
Manufacturers use data analytics to reduce unscheduled downtime, track key performance indicators, and improve factory efficiency and customer satisfaction. The broader trend is called Industry 4.0 or smart manufacturing. This involves aggregating data collected from conventional IT systems as well as industrial equipment and running analytics applications to make more informed decisions. Analytics also helps manufacturers identify the root causes of production errors and predict bottlenecks across manufacturing and supply chain processes that could disrupt order fulfillment.
Most manufacturers use sensors to collect data from their plant and equipment, known as operational data, and from the IT systems that run applications to manage their manufacturing, financial, supply chain, and HR processes. Manufacturing analytics helps business leaders make decisions based on that amalgamated data.
For example, analytics systems let business leaders track key performance indicators (KPIs) to identify which suppliers consistently deliver on schedule, identify supply chain bottlenecks, and limit the scope of product recalls. Analytics systems also interpret inventory and work order data from the ERP system and data generated by machines on the factory floor, and alert managers of the potential to miss a key delivery window because of insufficient output or machine downtime. This type of analytics helps manufacturers improve their perfect order rate—a KPI that reflects a company’s ability to deliver the right number of goods, without loss or damage, in the correct packaging, and with invoices that accurately reflect stipulated pricing and the number of goods delivered.
At most manufacturers, sensors connected to key pieces of equipment send constant streams of data, typically stored in a data warehouse, about every imaginable type of parameter—examples include the temperature at which the motor is running and the level of vibrations emitted by ball bearings—all of which can indicate a potential problem that must be addressed before the equipment fails and takes down a production line.
More sophisticated factories combine operational data with related IT to alert production units about a possible disruption and business leaders that a particular work order or production associated with that equipment is threatened. This type of analytics can also include inventory. Managers use applications to visualize where to find inventory—in different warehouses or in transit from a supplier—and apply analytics to make better, faster decisions about dealing with a potential inventory shortfall that could stop a production run if it isn’t addressed quickly.
Manufacturing analytics provides substantial benefits, the most important of which are outlined below.
Successful analytics projects share several key characteristics, outlined in the best practices below.
Involve business stakeholders, all the way up to the C-suite, in developing analytics projects. Ensure that the projects yield early, meaningful results (see KPIs section) so they’re not seen as just another bunch of IT projects. For example, demonstrate that combining IT and operational data can help analyze connected metrics, such as the impact of on-time delivery on customer satisfaction or the impact of machine downtime on the perfect order rate.
To prove the value of analytics, start with data collected from a small number of machines, ones that are bottlenecks or are particularly crucial to a production line, rather than trying to create an enterprise-scale project. This approach is less expensive than a big bang one, is more likely to show immediate results, and often leads to greater demand for more wide-scale analytics projects.
Engage in a full-scale discovery of the different data types available from different systems used by various departments. This assessment should include applications used by acquired companies; accounts payable, payroll, and other back-office applications added over time; and even that one-off application that a developer created for someone a decade ago and is still running on a server under someone’s desk.
Include data collected from factory equipment or other operations alongside data collected in applications that manage manufacturing processes to get the most accurate analysis. For example, analyzing work order data from an ERP application with operational data about a production line’s cycle time can indicate whether a given order will be filled on time, a finding that directly affects customer satisfaction and revenues.
Aggregate data from different data warehouses into a single, cloud-based data warehouse or data lake. This is especially crucial after an acquisition because different companies often use different data management systems that don’t integrate well with one another.
Scope analytics projects so that the appropriate types of data are collected and analyzed. If one goal of the project is to reduce downtime, make sure sensor data is being collected for the equipment that needs to be maintained in working order. If a goal is to improve throughput, ensure you can record volume and collect time series data so you can measure how much is being produced in a given timeframe.
By leveraging no-code ML within analytics, anyone in your manufacturing organization can uncover hidden patterns based on historical data, such as identifying backlog trends in inventory, predicting machine downtime, analyzing resource underutilization, and correlating the impact of production shortfalls to key business metrics, such as revenue and margins.
Identify key areas where data isn’t being collected and add sensors or other capabilities to let that occur. Expand both the scope and complexity of analytics projects accordingly. For example, manufacturers can start by measuring the quantity of units produced and the percentage of time that equipment is operating at full capacity, subsequently adding quality measures, such as the number of units accepted as a percentage of total units produced.
Manufacturers can use analytics-driven insights from data amalgamated from integrated inventory as well as fulfilment, customer experience, sales, production, and third-party sources to make quick decisions and adjust production plans as needed.
Manufacturers use data analytics to improve the overall efficiency of their floor operations and supply chains and to gain better insights into KPIs, such as overall equipment efficiency, equipment uptime, and yield throughput. Consider the following examples.
Most manufacturing companies use data analytics, but in many cases they have yet to implement a comprehensive strategy. That includes aggregating and cleansing data consistently, running analytical queries against that data, and systematizing responses to alerts or other information revealed by the data. Manufacturers should consider the following 10 implementation best practices.
While most manufacturers already use information technology and, to some degree, telematics or other instrumentation on their equipment, their use of IT and analytics in particular tends to be uneven. That’s because data resides in different silos, making it difficult to access and analyze.
Standardizing on cloud-based IT systems will help manufacturers consolidate all this data—both structured and unstructured data—letting them use analytics in a concerted, consistent manner to gain accurate and trusted insights to improve decision-making.
Finally, the introduction of low-code and no-code ML embedded within analytics will let business users create reports on their own, without needing to fill out a request ticket or otherwise get help from IT. This will lead to more frequent use of data and all the resulting benefits.
Oracle Cloud Supply Chain & Manufacturing, part of Oracle Fusion Cloud ERP, helps manufacturers respond quickly to changing demand, supply, and market conditions. Manufacturers using this application suite can continuously monitor inventory patterns to mitigate the risks of work-order backlogs, determine if supplier performance could affect production goals, and much more.
Oracle Fusion Supply Chain & Manufacturing Analytics enables manufacturers to increase productivity with prebuilt insights, improve shop floor efficiency by quickly detecting anomalies, and optimize plan-to-produce processes with an integrated view of supply chain and manufacturing data.
How does analytics help manufacturers?
Manufacturers use analytics for a variety of purposes, including to reduce unplanned downtime, track and improve supplier performance, prioritize work orders, boost employee productivity, and reduce product defects.
Which kinds of physical events can sensors detect?
Sensors can detect the presence of flames, gas leaks, and oil levels, and they can sense physical properties such as temperature, pressure, and radiation. They can also detect motion and proximity of objects to one another.
Where do manufacturers get the data they analyze?
Manufacturers correlate data from a variety of sources, including factory floor machines, back-office IT applications, suppliers, and third-party providers of data focused on markets; demographics; weather; regulations; patents; environmental, social, and governance practices; and other information categories.
Learn how to improve the quality and speed of your supply chain decision-making and get ahead of tomorrow’s challenges in our ebook.