In part 1 of this blog, I gave an overview of business intelligence (BI), and the benefits it can bring to manufacturing. Part 2 will discuss the intersection of business intelligence and Lean manufacturing at the junction of visual management, a Lean discipline used to communicate important information to all levels and functions of an organization. business intelligence takes visual management to another level, enabling the display of process and machine data in real time to shop floor personnel and manufacturing engineers, so they can identify off-target conditions as they happen and take immediate corrective action. The result is superior process management and control that flows directly to the bottom line.
Manufacturing collects and stores more data than any other industry. Are there opportunities on the shop floor to use that data to decrease costs and increase revenue? Is there potentially valuable data that’s available, but not being collected or analyzed? When data does indicate that action needs to be taken, is it provided in real time to the shop floor, and in a format that can be easily understood? Or is the data provided too late or in too confusing a form for effective action to be taken? Does the data provide its full potential value to manufacturing, or is much of its value wasted?
SInce the early 1980s many US manufacturing companies have adopted Lean manufacturing principles to improve their bottom line. One of these key principles is to eliminate waste, or non value-added activities, in manufacturing processes.
When business intelligence provides real-time actionable data to manufacturing, rework and scrap, machine downtime, schedule slippage, excessive inventory expense, and other forms of waste can be reduced, and even eliminated.
Business intelligence – A survey by Dresner Advisory Services published in May 2017 reported that the top two business intelligence objectives of manufacturing companies were better decision making and improved operational efficiency. To achieve these goals, the most important BI initiative for manufacturers was creating reports and dashboards for internal users.
Lean manufacturing – A 2015 survey of 120 Lean manufacturing companies by the Aberdeen Group looked at the key differences between companies that were more successful at Lean (Lean leaders) from those than were less successful (Lean followers) based on five measurable criteria. Two key differences of Lean leaders were the degree to which they measured the direct ROI of Lean (72% reported improvement in revenue, profitability, market share, productivity, and quality), and the degree to which dashboards were used to communicate performance on key metrics throughout the organization.
A business dashboard is an information management and communication tool. Dashboards are used to monitor key performance indicators (KPIs), metrics, and other business-critical data at the enterprise, department, process, or machine level. Dashboards can display single or multiple data sets, using charts, graphs, digitally-represented gauges, and other visualizations, simplifying complex data sets and providing users with at an at-a-glance notification of current performance. Dashboards are most useful when they display performance data in real time, and clearly specify a critical or alarm level that mandates that immediate action be taken to correct or improve substandard performance.
Working in manufacturing management over the past 30 years, I’ve utilized almost every tool in the Lean and Six Sigma toolboxes to improve bottom-line results. Thanks to Norman Bodek*, my long-time mentor and friend, I was privileged to write the introduction to “Fundamental Principles of Lean Manufacturing” by Shigeo Shingo. Shingo (1909 -1990), and Taiichi Ohno (1912-1990), were responsible for creating the Toyota Production System in Japan, which provided the foundation, principles, and many of the tools of Lean manufacturing in the US.
The goal of Lean manufacturing is to create competitive advantage by maximizing organizational effectiveness and efficiency:
- Effectiveness – Enabling companies to sell products that deliver the maximum value to customers at a fair market price, and to produce those products at the lowest total cost.
- Efficiency – Accelerating cash flow by reducing the elapsed time between the expenditures required to make products and the corresponding income received when payment is received for those products.
* Norman Bodek, known as the “godfather of Lean”, is a Shingo Prize winner, and is one of only 51 members inducted into the Industry Week Hall of Fame (along with Steve Jobs, Michael Dell, and Jack Welch). In the early 1980s Norman introduced Lean manufacturing to the US by bringing Shingo and other Japanese manufacturing experts from Japan to consult with US manufacturing companies, by translating and publishing the books of Shingo and others, and by creating Productivity Inc, the first Lean training and consulting firm in the US.
This goal is accomplished through the practice of the five principles of Lean:
- 1. Identify the value stream – The manufacturer identifies the sequence of operations—the value stream—within a manufacturing process that will incrementally add value to raw materials and work in process (WIP) to produce a product that meets the customer’s requirements.
- 2. Pull from the value stream – The manufacturer does not incur any costs, including the purchase of raw material, until the customer places a purchase order. In effect, the customer “pulls” value through the value stream.
- 3. Continuous flow of value – The operations in the value chain are balanced so that raw materials and WIP can flow through the manufacturing process without interruption.
- 4. Elimination of waste – To achieve continuous flow, all waste (non-value-added activity) must be removed from the value stream (See “The Eight Wastes of Lean Manufacturing” below.)
- 5. Never-ending improvement – To maintain a competitive advantage, companies must continue to improve.
Identification and elimination of waste is an ongoing process that reduces cost and improves the flow of materials, products, and information. Below are the seven sources of waste in manufacturing identified by Taiichi Ohno, with an eighth critical waste added by Norman Bodek:
- 1. Transportation" Movement of materials and products that does not add value.
- 2. Inventory Purchasing and storing more raw material inventory than is needed to fill current demand.
- 3. Motion Any human movement (hands, eyes, walking, keystrokes, mouse clicks, etc.) that does not add value.
- 4. Waiting Time lost when needed materiel, information, personnel, or raw goods are not available.
- 5. Overproduction Producing more product than is immediately needed by customers.
- 6. Over-Processing Performing any step in a manufacturing process that does not directly add value.
- 7. Defects and Errors Producing products or data that doesn’t meet customer, internal, or regulatory requirements.
- 8. Underutilization of Human Potential The missed opportunity to engage the talent, creativity, and ability of employees to improve their work processes.
One of the sources of waste that Lean manufacturing is just beginning to include in its focus is information or data waste. Entering data that’s never used, searching for or waiting for data, missing or incorrect data, data that’s not available when needed, data that’s in a form or format that’s confusing or not actionable, are all sources of waste.
One of the most powerful tools of Lean manufacturing, visual management, is used to display actionable data and information on large LCD displays, whiteboards, computer monitors, flip charts, digital signs, labelling, signal lights, and other visualizations. The value of visual management is in providing understandable and actionable data to all levels and functions within a manufacturing organization.
On the shop floor, visual management dashboards provide data that production teams need to do their jobs, in a highly visible and understandable format, in a location where it’s needed most, and, most importantly, in real time. To be truly effective, visual management dashboards must signal to a production team that immediate action is needed to prevent or correct a product, process, or machine issue. To accomplish this, they must display:
- 1. The Metric - For example, the number of “good” units produced by a machine since the start of the shift. This tells the production team “this is important.”
- 2. The Target - For example, based on standard times, 100 “good” units an hour. This tells the team, “this is the expectation.”
- 3. The Variance from Target in Real Time - For example, “good” units produced in the last hour has fallen to 98 units. Knowing this, the team can answer the question, “Are we performing to expectations?”
- 4. The Alarm Level - For example, 98 good units were produced in the last hour. The dashboard tells the team that no action is required until good units produced falls to 95 units in the previous hour. This lets the team answer the question, “When should we take action?”
In my career, I have been very fortunate to lead several business-critical Lean projects, working side-by-side with some extremely talented shop floor operators, manufacturing engineers, PLC programmers, and software engineers. However, It wasn’t until recently that I realized that a handful of these projects integrated Lean with shop-floor business intelligence as well as the Internet of Things (IoT). Among those projects were:
- Implementation of automated, real-time statistical process control (SPC) on the shop floor to reduce the waste of defective product. The last operation in the manufacturing process was also one that produced an unacceptable amount of defective product. Any defective product from this operation was painful because it carried the added value from all previous operations. Since rework of this product was not an option, any defective product had to be scrapped. The solution was to implement SPC on the shop floor. A type of process dashboard, SPC has long been used to detect an unusual or unexpected change in a manufacturing process. When such a change occurs, the process is considered to be out-of-control and its output unpredictable. Detection of an out-of-control process is accomplished by applying a set of statistical rules to the data after each point is plotted. When an out-of-control condition is detected, production is halted, and immediate action is taken to identify and eliminate the root cause. If done in real time, the process can be stopped before any defective product is produced, or at minimum, the number of defective products can be minimized and contained.
Figure 1 - Statistical Process Control X-Bar & R Chart
The integration of Lean, IoT and BI – Although paper and pencil SPC charts are valuable at preventing defects, they also can be the source of a bottleneck in a continuous flow process due to the required amount of time to populate them. The manual operations of calculating, plotting and interpreting the points on the chart are also subject to human error, resulting in incorrectly plotted points on the control chart. This can lead to unnecessary process interventions when the process is still in control (and could potentially itself cause an out of control condition), or a failure to intervene when the process is, in fact, out of control.
Working with a group of software engineers, I led a project to implement automated, real-time SPC on each of the 40 machines in this operation. Test data from a random sample of 5 units pulled from each machine run was automatically uploaded by the test device to an SQL database and plotted as points on a digital SPC X-Bar and R chart. All 40 charts were viewable by anyone on our network, and any out-of-control condition on a chart automatically triggered a flashing red light (an “Andon light” in Lean terms) at each machine, signaling operators and process engineers to take immediate corrective action (the actionable insight).
Creating a real-time OEE chart (Overall Equipment Effectiveness) on an automated production and test line to improve production flow. OEE is a Lean metric that’s a combination of machine’s Availability** as a percent of planned availability, Performance (rate of production) as a percent of maximum, and Quality as percent of output that meets requirements. The product of these percentages yields the OEE, or overall effectiveness, of the machine. For example, if all three percentages were at 90%, the machine would only be 73% effective (see Figure 2 below). A machine that’s 100% effective is one that operating at peak effectiveness.
Figure 2 - Overall Equipment Effectiveness Chart
**Availability is equal to scheduled machine uptime minus unscheduled machine downtime plus changeover time.
The integration of Lean, IoT and BI – Automated testing and PLCs uploaded the required OEE data to our home-grown production database. A dashboard showing all three components of OEE, as well as the overall OEE, was displayed in real-time on a large screen LCD monitor on the shop floor. This provided “actionable insight” to our operators and engineers, allowing them to take immediate action to address machine issues that affected continuous flow, and potentially prevent machine downtime.
Developing continuous flow for each product SKU in a mixed-model production environment. Continuous flow production is often difficult to implement in a mixed-model production environment, especially when each product SKU flows through the same set of operations, and cycle times through each operation vary by product. This was the situation in one company I worked for and had been a barrier to Lean Implementation.
The integration of Lean, IoT and BI – I had the opportunity to collaborate with our software engineers to create a continuous flow system for each of our products, despite the variability in operation cycle times. The first step in this project was to set up barcode readers at each operation. As each barcoded unit was “received” into an operation, it was scanned and the clock time automatically recorded in our ERP System. This gave us the ability to track cycle-time from station to station (the actionable insight). With that information, we could then identify and address bottlenecks to continuous flow for each of our SKUs. Once bottlenecks were addressed, we could then determine time standards for each operation across all of our SKUs. Our Software Engineers then created a real-time dashboard that displayed the actual and standard cycle times for each operation in real time on computer monitors at each operation. When cycle times were greater than standard (which included an allowable variance), engineering was immediately called to the floor to take action to identify and correct the root cause.
As we reduced and controlled the variance to our time standards, an additional benefit was that we now had confidence about production lead time for each of our product SKUs, and much tighter control over the direct labor portion of cost of goods sold. Integrating those lead times into our production scheduling system enabled much tighter scheduling, more accurate revenue projections, and much more accurate lead-time commitments to our customers.
The competitive advantage that can be achieved by the integration of these Lean and BI is a strategic opportunity that every manufacturing organization should consider exploring. If your PLCs, test equipment, and machines can output data, that data can be transformed into real-time business intelligence dashboards for a reasonable capital investment. Integrated with the principles of Lean manufacturing, the real-time actionable insights provided by these dashboards can produce immediate results that flow directly to the bottom-line.
A future blog will look at Industry 4.0, so-called because it represents the fourth revolution in industrial production. Going beyond the integration of Lean and BI, Industry 4.0 envisions embedded IoT in automated:
- production equipment
- product inspection and testing
- monitoring of equipment conditions,
all linked to machine learning algorithms that monitor, adjust and control all these integrated shop floor systems. This “smart factory” will also be digitally connected to both suppliers and customers, automating communications throughout the entire value chain.
Jeremy Green, PhD
Lean Six Sigma Master Black Belt