That Day, When the CPK Report Came Out, the Whole Room Was Silent for Three Seconds
I still remember several years ago, our production line shipped a batch of products, and we received a flood of complaints from the customer. The reason? A critical dimension of one of our components had consistently drifting tolerances. The process report at the time showed a CPK of only 1.08, barely scraping by, yet the yield rate wouldn't improve. The boss, fuming, slammed the table in a meeting and demanded, "What exactly is your IQC doing?" For three seconds, no one in the room dared to speak. Our IQC strategy was the most easily overlooked link in the chain.
Where Did the Problem Lie?
To put it bluntly, IQC (Incoming Quality Control) is essentially a health check for supplier materials before they enter your warehouse. What's the purpose of this inspection? It's simple: to ensure these materials are usable and won't cause you trouble during production. The core point of contention here is, "Should we conduct 100% inspection, or is sampling inspection sufficient?"
100% inspection sounds ideal; every part is checked, seemingly foolproof. But consider this: if you receive a million screws every day, can you really measure the dimensions and inspect the appearance of each one individually? The costs of manpower, time, and inspection equipment alone would be enough to skyrocket your department's budget. In other words, 100% inspection trades a high price for theoretical "zero defects," but in practical operation, it is often unrealistic.
As for sampling inspection, it involves randomly selecting a subset from the incoming batch for inspection. If these sampled items are all acceptable, we assume the entire batch is acceptable. This might sound a bit like gambling, right? Indeed, there is a risk. The key lies in how you control this risk within an acceptable range.
How Is It Done in Practice?
To decide between 100% inspection and sampling inspection, the simplest and most straightforward approach is to assess "supplier quality stability" and "material risk level."
- Supplier Quality Stability: If your supplier's historical data indicates that the quality of their delivered materials has been consistently stable, with a defect rate (DPMO - defects per million opportunities) below 500, and they provide clear quality reports, then sampling inspection is definitely the preferred choice. You can establish a more lenient sampling plan, for example, based on MIL-STD-105E AQL levels, conducting sampling under "General Inspection" and "Normal Inspection." Assuming a batch of 1000 units with an AQL of 1.0, you might only need to sample 80 units, and if the number of non-conforming units exceeds 2, the batch is rejected.
- Material Risk Level: What constitutes a high-risk material? Consider if a failure of this material within your product would lead to the entire product being scrapped or even impact personal safety; then it is high-risk. Examples include critical chips in medical devices or sensors in automotive braking systems. For such high-risk materials, even if the supplier's quality is excellent, you might still consider enhanced inspection, or even partial 100% inspection. Conversely, if it's merely a decorative button with little impact, then sampling inspection can be comfortably applied.
Therefore, the key point is that this is not a simple either/or multiple-choice question, but rather requires dynamic adjustment based on actual circumstances.
The Most Common Pitfalls
The most common mistake I've seen is the "one-size-fits-all" approach. Many companies, especially new ones, to save effort, use only one set of sampling standards for all incoming materials. What's the result? For suppliers with poor quality, a multitude of non-conforming products might be missed due to insufficient sample sizes; conversely, for suppliers with good quality, too many unnecessary inspections are conducted, wasting manpower and resources.
There's another pitfall: "only looking at the surface." Some engineers simply review supplier reports, and if the data looks good, they approve the materials directly. But honestly, reports are static; processes and people are dynamic. You need to trace how this data was generated: Were rigorous tests conducted? Is there a possibility of falsification? I've encountered suppliers who, to pass acceptance, would only send their best materials for inspection, but the actual shipments were entirely different. Therefore, regular supplier audits and building trust-based partnerships are far more important than merely reviewing reports.
One Thing You Can Do Today
Re-evaluate your department's IQC strategy: Is it too rigid?