Statistical quality control tool: get a valid sampling plan

Quality Control TipsWhat is the difference between statistical quality control and unprofessional inspections? A professional inspector will follow a valid sampling plan.

A sampling plan should take care of three fundamental concerns:

  • There should be a clear limit between acceptability and refusal,
  • This limit should be set before inspection starts (actually, even before production starts) and be transparent for the supplier,
  • The inspection’s conclusions should be applicable (with a certain degree of confidence) to the whole order quantity–not just to the few samples that were checked.

The problem: no user-friendly tool

The quality control industry in Asia uses the same statistical tools. They were set as standards by the US army during World War II (to check the army’s procurements), and they were more recently taken over by “commercial” institutes such as ISO, ASQ, etc.

Everybody calls them the AQL tables (one of many jargon terms of the QC industry). The AQL tables are very functional for everyday users, but they are a real put off for buyers and suppliers.

I read that Google did a survey and found that only 8% of pedestrians in Manhattan knew exactly what a “browser” is. In the same way, I bet that less than 5% of purchasers know how their inspectors choose a sampling plan. And even less have a clear understanding of the impact of a difference in inspection level of in AQL.

An easier way to get a sampling plan

I came upon one or two online tools to calculate the number of samples and the AQL limits, but they are not really adapted to quality control in the context of international trade. And all they give is numbers, not explanations.

So we developed a very simple tool to get a rigorous plan and to understand how to apply it:

[Update: we have "killed" this feature. Sorry!]

I hope many readers will use it and will spread the word. Any suggestions to make it better or more convenient?

Print Friendly

Comments

  1. says

    Great blog and great tool as well, you did something tedious easy to be applied, and as you say , even explained… Awesome!
    Thanks for sharing your experience…
    My sister company does QC in Indonesia but always piece by piece; because it is the only way to have a proper product… I like reading about things like this but in Indonesia you take too many risksm and later they (manufacturers) don’t want to be fully responsible…
    If you ever need help in Indonesia field, remember me… I’ll do for China…
    Who shares is brave, and finally will get the best…
    Thanks!

  2. says

    @ Carlos: thanks for your feedback! 100% check is also done in China, mostly by Japanese QC firms. Of course it is better in terms of reliability, but I see 2 downsides:
    -1- It is very long and expensive, compared to statistical quality control.
    -2- It tends to take place at the end of production… But quality issues should also be checked early in the manufacturing cycle.

  3. MIKE MANDA says

    i want to learn more about QC as am currently working as assistant QC in a mining industry at the Laboratory

  4. Renaud Anjoran says

    Mike,
    There are many books about statistical QC and statistical process control… And you can read all the archives of this website!

Leave a Reply

Your email address will not be published. Required fields are marked *