Manufacturers are facing a quality crisis. With product recalls reaching a seven-year high and high-profile failures like the Boeing jet door plug disaster, it’s clear many manufacturers have let their once-stringent quality standards slip.
Some assume it’s cheaper to tolerate poor quality, recall products, or replace them under warranty than to address the root causes. But is it really? I would say no. They’re likely massively underestimating the cost of reputation damage, brand impact, and lost customer and market confidence.
These impacts can be extremely hard to quantify, especially when they’re not directly connected to the manufacturing or design process. In the auto industry, for example, recalls are handled at the dealership, which is far removed from the factory. Not to mention it’s nearly impossible to know how many customers opted for another brand on their next purchase because of past experience.
AI is Not the Answer
In an effort to claw their way back to superiority and rebuild trust, many manufacturers are leaning heavily on high-tech solutions like AI and automation to achieve aggressive zero-defect goals.
But the truth is, technology alone can’t solve the root cause of most manufacturing quality problems, and collecting terabytes of data won’t address the underlying issues around process, supply chain, labour practises and standard work.
Get Back to Basics by Focusing on Good Manufacturing Fundamentals
Rather than pouring millions of dollars into complex, high-tech solutions that require teams of experts to install, manage, monitor and analyse, manufacturers should get back to basics by focusing on good manufacturing fundamentals. Here’s how:
- Proper equipment maintenance and calibration. A surprising number of companies neglect routine maintenance and just wait for something to break. Maintenance is not the same as repair. Manufacturers need to prioritise proactive cleaning, inspection, changing filters, lubricating, calibrating, and servicing equipment at the right intervals to catch potential problems before the machine goes out of spec or breaks down. Inspection and testing equipment should be calibrated routinely to make sure you’re getting gauge repeatability less than 10% of the tolerance. Don’t give up all your tolerance to gauge deviation.
- Measuring process and input variables. KPIs and performance systems often focus too much on outcome metrics—making sure the finished piece meets requirements. While that’s important, the secret to driving better outcome quality is to measure and mitigate process and input variables. Let’s say you’re making disposable plastic drinking cups, and suddenly you’re getting missed punches on the die that forms the cups from the dual-layer resin. It could be the punch that’s at fault, but it could also be that the material temperature, layer adhesion, extrusion pressure or speed of the rolling process is off before it even gets to the punch. Using AI to analyse the cup at the end of the process can help you spot the defects, but it isn’t going to help you troubleshoot the process up the chain.
- Re-examine the engineering. Automating a process won’t solve a quality problem if the issue is in the design of the process itself—robots will do exactly as they’re programmed to, even if it’s wrong. As blueprints, models and design schematics go through multiple iterations over time, each update or change introduces the possibility of errors or conflicts. Examine blueprints and specifications for measurement errors, transposed numbers, a number eight that can be mistaken for a three—you’d be surprised how the smallest mistake can propagate. And it might be so small that it’s not a problem 99% of the time, but that 1% is costing you.
- Implement good standard work. Creating templates and colour-coding components can help eliminate some errors, but beyond that, you need standards that spell out the details on the tools, process settings and the definition of quality for each component and/or step in the manufacturing or assembly process. Make sure every operator is trained properly in what zero-defect work looks like and the steps required to achieve it. Identify potential failure points and develop mistake-proofing interventions. Trust but verify both the people and the process.
- Audit suppliers. Because supply chain issues have become such a huge pain point for manufacturers, many are happy they can even get the parts they need. They’ve prioritised availability over quality and rely on first article inspection to verify the entire lot. When you’re scrambling to find supplies while also controlling costs, companies don’t have the time or resources to invest in doing a deep dive into their suppliers’ process and underlying standard work. Especially if you opt for low-cost, offshore suppliers, you might get some labour arbitrage, but you pay for it in poor quality down the road. Instead, conduct a process capabilities analysis and regularly audit every supplier to validate their internal operations, rather than relying on the luck of the first article inspection.
- Cultivate company culture. AI can’t help an operator who is burned out, bored with their job, or tyred of telling their boss about the same issue multiple times with no solution implemented. Disengagement, boredom and feeling undervalued is a recipe for complacency, mediocrity and indifference that results in poor quality. Start with the core belief that everyone comes to work every day intending to do their best and nurture that behaviour. Implement cross-training and job rotation so no one is stuck doing the same repetitive task 40 hours a week for 20 years. Create an employee task force that brings hands-on experience to the table to identify and help rectify issues. Engage staff in creating a better work environment in which they feel invested and proud.
- Walk the factory floor and talk to people. Computers and robots are only as good as the people who write the programms, and high-speed cameras can’t catch every mistake. Some problems can only be seen and solved by human beings. There’s a tremendous amount of knowledge on the shop floor that is so often overlooked. Make it a priority for management to walk the factory floor regularly and talk to operators. Ask what they’re doing, how they do it, and if there’s anything they need or if they have any ideas for how to do it better. All the tech in the world is no substitute for boots on the ground.
There’s no doubt AI and other technology can play a huge role in improving manufacturing quality. But it’s no silver bullet, and in fact can be a double-edged sword. AI solutions require massive amounts of data which takes resources, expertise, and intention to cleanse the data, analyse it properly and act on it appropriately in order to be useful. If you don’t take any better care of your data than you do your equipment, it’s a waste of time and money.
The best way to achieve zero-defect manufacturing is to get back to basics and address manufacturing fundamentals before you begin introducing new technologies. And be honest and realistic with yourself; you didn’t get into this situation over the course of a single quarter, and it’ll take longer than that to right the ship. Improving quality requires a continuous investment in people and tools working together with a long-term view of performance and risk mitigation.
