How Consumer Packaged Goods companies can improve Overall Equipment Effectiveness (OEE) with AI
This blog is co-authored with Ajeet Singh
The U.S. consumer packaged goods industry is a powerhouse responsible for 1 in 10 U.S. jobs and contributing more than $2 trillion to U.S. GDP every year.
The industry experienced intense growth during the pandemic despite supply chain disruptions and shifting consumer demand.
But many hidden villains still compromise efficiency, create waste and squeeze margins in the CPG industry and impact overall equipment effectiveness, or OEE.
OEE measures the rate of fully productive time relative to planned production time, while also making first-quality products. If an OEE score is 100%, a plant produces only the good products ordered by planners as fast as possible during scheduled production time. Per published and generally accepted industry specific benchmarks, the CPG industry has much room for improvement with respect to OEE performance. CPG manufacturers, on average, have an OEE of 66.4%, indicates a study of 100 manufacturing operations worldwide.
This means that more than 30% of potential is lost, adding up to billions of dollars of losses in terms of efficiency, excess energy consumption, lost capacity and suboptimal capital deployment.
And it is avoidable. The “best of the best” in the study exhibited an OEE of 96.9% while “best-in-class” scored 82.5%—showing what is possible.
Factors eroding OEE
Improvements in OEE in all industries, including CPG, are now more possible than ever. Digital transformation continues to bring increasingly advanced technologies to the factory floor, and OEE is a prime target for improvement.
That is especially true for CPG companies where margins are everything in an industry that sells affordable products that almost all consumers repurchase often, such as laundry detergent, toothpaste, and toilet paper. Every efficiency added in the factory ripples throughout to many other products being made and then sold.
Factories will always have scheduled losses that stem from such things as necessary plant shutdowns, lack of demand, disruptive weather or safety meetings.
Exhibit 1: OEE factors with overlay of solutions which address problems from Noodle.ai products (Asset Flow and Quality Flow)
But a material amount of existing losses in the CPG manufacturing sector can be diminished. The three OEE factors driving those losses are availability, performance and quality. None is inherently more important than the others. The point is to focus on what’s causing the underlying loss in each of the three areas:
- Availability Loss takes into account such things as unplanned downtime due to equipment breaks and material shortages.
- Performance Loss relates to idling, minor stops, reduced speed due to such things as material jams, blocked flow and misfeeds.
- Quality Loss addresses process defects, yield loss, out-of-spec product and other factors.
Why artificial intelligence is key to improving OEE
The most successful CPG leaders are using advanced technologies such as AI and machine learning to drastically reduce losses in each of those areas.
For example, take a company that makes bread. Some of the bread ends up underweight and so it cannot be sold. Once that factory is equipped with sensors that feed real-time data into an AI engine, the source of the problem can be more quickly identified.
Then, AI models pinpoint the handful of factors, out of hundreds of sensor values, that cause bread to end up underweight. The out-of-sync factors might include mixer speed, temperature, conveyor speed, baking time, dough consistency and on and on.
The AI engine calculates, in real time, how one factor impacts another factor, which impacts the weight of the bread and the final product. It typically isn’t one factor that’s the problem. It is a combination of factors – factors that are constantly changing as products change, operators change, and conditions change. Over time, the AI engine learns from the data and can even predict what will happen if the mixer slows slightly and how to adjust baking time or other factors to hit the sweet spot so that the bread stays perfect.
Ability to analyze data and uncover anomalies
Not even an experienced plant operator can consume and analyze as much data as an AI engine can. As such, preventable problems get missed, which hurts OEE.
Such knowledge was exactly what separated the best performers from the worst in the study of CPG manufacturers.
The best performers knew more often than the laggards why their equipment was failing. Only 0.5% of equipment downtime reasons were unknown for the best-in-class performers versus 15.7% of downtime reasons for the laggards, the study found.
If plant operators know why equipment falters, they can take preemptive steps to prevent that, which will reduce availability losses and drive up OEE.
An average CPG production line stops 20,000 times per year. In the study, the poorest OEE performers had six times more minor stops than the best performers. Frequent stoppages hurt OEE. If plant operators know the causes behind minor stops, the more likely they can be prevented.
Unlocking higher levels of OEE with Flow Operations (FlowOps)
FlowOps is a new category of software that empowers companies to identify valuable complex patterns in operations, to predict waste that likely will slow flow and to recommend actions to restore perfect flow.
FlowOps helps manufacturing leaders drive higher OEE in plants, via purpose-built and native AI applications intended to address asset performance and quality management.
FlowOps is made possible by combining fundamental breakthroughs in Explainable AI, (XAI) high performance computing, data storage, and machine intelligence. XAI continuously takes in vast amounts of real-time data about every action in and around a manufacturer’s operations from factory to supply chain to distribution.
It identifies patterns in the complexity that people could never see. Those patterns enable XAI to make predictions about what’s likely to happen, offer recommendations on how to avoid it and arm factory operators with the knowledge to make the right decisions.
Asset Flow is a next-gen Asset Performance Management application that prevents unplanned downtime and reduces planned down time through time-to-failure prediction and curated anomaly detection by analyzing tens of thousands of signals across hundreds of components in complex assets.
Hidden anomalous patterns across all types of live sensor data enables Asset Flow to predict the time to failure and suggest actions to avoid plant operation issues. The world’s manufacturing plants are riddled with unplanned downtime, resulting in an estimated loss of almost $700 billion a year. With Asset Flow, Noodle.ai customers have achieved 25% to 50% reduction in equipment downtime, a 2% to 5% drop in maintenance cost, and a 3% to 5% improvement in equipment life.
Quality Flow provides diagnostic and predictive capability to prevent quality issues and yield losses by analyzing hundred-thousands of parameter combinations across multiple production stages.
Quality Flow detects anomalies to identify leading defect drivers, predict probable defects and product specification variability, prioritize issues based on cost to the business and recommend changes to improve quality and yield.
With Quality Flow, Noodle.ai customers typically achieve more than a 25% drop in defect rates, up to a 35% decrease in quality variability, an overall 25% to 35% reduction in total cost of quality and a 15% to 20% increase in (first pass) yield.
A combination of Asset Flow and Quality Flow addresses short line stops. If the line stop issue is equipment related, such as machine faults, then Asset Flow effectively tackles it. If the issue is related to the quality of product or materials flowing through the equipment, then Quality Flow handles it. If it is a combination, then hybrid modules from both products tackle the issue.
Start small. Start big. But just start!
While many CPG manufacturers are investing heavily in advanced technologies, every plant, every production line will benefit from a focus on OEE.
The good news is any plant can begin to tackle this problem. Quality Flow and Asset Flow work seamlessly with existing systems. OEE improvements can be made plant by plant, or line by line.
Look for incremental improvement in OEE, with each step being a stretch target that is achievable, preferably within a few months so the time is short enough to keep people engaged and long enough to achieve significant improvement.
Also, focus on improving OEE in context and with background knowledge. It can be enticing to improve OEE by tackling the lowest performing line. But if that line is not a constraint or is not of strategic importance, it might be best to concentrate elsewhere.
Also, avoid the temptation to improve OEE by creating unneeded inventory. Very often, companies make the wrong products, or too much of the right products, very efficiently. A high OEE can simply mean that production is perfect at making what planning tells them to make. Never strive to simply drive up OEE at the expense of inventory cost and obsolesce.
In general, the benefit and motivation to improve OEE will be higher for evergreen versus line extension, or net-new, where new equipment is used and is less likely to suffer unplanned downtime.
Higher efficiency: Stronger company, healthier planet
By concentrating on asset and quality flow, CPG manufacturers will boost OEE, which will help margins and profitability. Also, CPGs will waste less energy and produce fewer goods that never make it to market because of defects. Given the heft of the CPG industry, that’s a big win for the environment, too.
To schedule a demo to see the Noodle.ai Manufacturing suite of products in action, visit https://try.noodle.ai/demo/
- How to Get Actionable Answers to Yield and Defective Production Issues Finally — and for Real
- Process over Perfection: Augmenting Intelligence in Business – By Garry Kasparov
- Announcing our SOC 2 Type II Certification
- Make Profit Not Waste – Supply Chains and The Legend of John Henry
- 7 Tips to Unlock Human + Machine Intelligence in Factories.