I like the answers of both Fred and Cliff, but am going to offer some counterpoints
While .6% may seem like a small number, it is certainly well above "six sigma" levels, and I am not sure Iwould be willing to live with that, though I agree with Fred it really depends on the value of the items being received. Also, wonder if the errors are random over/underages, or whether it is usually shortages.
The challenge with Cliff's answer might be that you cannot get accurate product weights from the suppliers, or they use random carton sizes, either of which would potentially thwart the "check weight" approach.
You may be able to work with suppliers on both issues, however, and/or weigh the products yourself if that's the only recourse. Not sure how many total SKUs you are dealing with, but if you are opening the boxes now to get a weight as part of the receiving process, adding a weigh step for a short while would be fairly painless. If you can capture that and have vendors always use the same size cartons, this certainly could work.
I don't think a sampling plan will really help you much. Auditing processes like that are usually to ensure that the overall process/quality seems to be working - it would only really tell you whether that .6% error rate was growing or shrinking over time. It this scenario, it would not be really useful for finding the errors.
I agree with Fred that I suspect there are some patterns that may not be apparent. Could be, for example, that 250 vendors are almost always accurate. Cutting the inspection down from 400 suppliers to 150, just for example, might be a reasonable compromise. I don't know your receipt profiles, but for example do you receive some "full cases" and some "mixed cases," or cases with less than full case quantity. If so, the mixed or less than full cases are much more likely to have errors. So, you let the full cases go straight through, inspect the mixed or partial cases. There may be other variants on this theme.
The last idea I have concerns whether you can capture when a carton is "consumed" in manufacturing. If yes, can you then compare that to the amount your manufacturing systems thinks has been consumed? Flagging those exceptions will at least allow you to react the "shortages" (overages are less of an issue) from an operational perspective, though who takes the hit for the inventory I don't know.
Seems to me from all of the thoughts thus far you ought to be able to craft a solution that works for all. For example, maybe some suppliers/types of shipments are exempt from inspection, and the remaining are check weighed for errors, rather than opened and counted, except for high value items which are opened. Something like that.
Supply Chain Digest