I'm going to first suggest that you read:
http://Elsmar.com/ubb/Forum12/HTML/000022.html - it has some discussion on containment.
Containment is the many things a company does when a problem is identified - internally or externally. Containment is just that - to contain. To keep from getting out.
Including:
* Identify part number and determine what has been shipped, when, what is in the hands of customers (in their stock - requires close contact with a customer 'point of contact'), what is enroute, etc.
* Determine when the 'event' occurred (like lot number).
* Isolate in-house and warehoused suspect product.
* Determine earliest and latest. Eg. Did test fail? When? Is this a 'standard' that went bad? In the case of a test instrument calibration, when was the last calibration? Can you test interim product to see if stuff in between is OK or where the calibration went far enough out as to allow nonconforming product to be 'passed' and shipped. What shipped product lots are 'suspect' and 'what's in the pipeline'?
* Close communication with customers to ensure 'timely' update on possible (suspect) lots or known 'contaminated' lots.
* The customer has to have knowledge of lots - what may they have shipped that was assembled with your suspect component.
What you do to ensure that product shipped from the point at which you know of the problem is 'Conforming' [HOLD EVERYTHING!!!! - What have we got here?] (often 100% inspection/test for the specific defect and then you 'certify' [as per the SS requirement, but typical outside the sector as well - can you say SORT?] each shipment for an agreed upon time period - which depends upon the problem identified, etc., as I'm sure you know). This is what you ship to the customer within X hours (not always considering reality). The intent here with the stated hours is to avoid interrupting a customer's manufacturing schedule.
You might also want to wade through:
http://Elsmar.com/ubb/Forum4/HTML/000109.html
Run-at-Rate: if you change your process, this *may* be a part of the associated rePPAP.
[This message has been edited by Marc Smith (edited 15 February 2001).]