How many complaints starts a recall

Hi all!

We are conducting our scheduled review of our Recall Procedure, and we are trying to elaborate a tool capable of make easier and quicker taking the decision of having a product recall or withdrawal, something like a decision tree or a questionnaire to use on our mock recalls or in a real event.

We are trying to do this as we have our issues to answer How many complaints are tolerable? When it comes to quality defects.

When it comes to health issues or legal issues and your documentation shows a that you probably had a problem, then recall decision gets a little easier.

But when it comes to Quality Issues and reputation is the only thing that is on stake, then how many complaints do I need to receive to think on a product recall or withdrawal?

i.e. Should 3 customer complaints in 1 week be enough to detonate an alarm? 

Thanks for your help!

FA

I would start out by saying, it depends on what the complaints are for.  3 in a week for a chemical taste to your food, BIG issue, 3 for perceived shortages, much less of an issue.

Maybe start with a definition of what circumstances would you recall for quality?  Would Mishandled product, temperature abuse...be reasons to recall?  If you had incorrect ingredients, I feel that could be a safety issue, with allergens, ingredient statements, etc.But quality becomes a softer target.

Setanta

Dear ajrfrank,
 

 

We are trying to do this as we have our issues to answer How many complaints are tolerable? When it comes to quality defects.

IMO an absolute answer to yr general question is not possible.

ISO 9001 presumably relies on "fitness for purpose" as the overall criterion.

Perhaps you need to do a risk analysis analogous to that in HACCP.

Never used it myself but the guidance requirements for SQF level 3, eg CQPs (?), should give you a few ideas on the subject perhaps.

From a QA POV, loss of reputation seems rather questionable as a critical variable. :smile:

Rgds / Charles.C

PS - posted before seeing Setanta's reply which attempts to quantify the same kind of opinion (I think :smile: )

Yes, but you do it with nice cites and references...  :cheezy:

I have just uploaded a file in the Risk Assessment section of files. It was given to me by a consultant but haven't used it myself.

What Setanta and Charles C said.  :smile:

Hi FA,

To a large extent recalls based on quality are a 'commercial decision' where there is a need to gauge costs of recall vs. cost/number of complaints & loss of customer/reputation/future business.

The risk assessment posted would only work for food safety as quality defects will always and up as low risk - action unlikely

If you do trend analysis you will be aware of a normal level of complaints. I would have an arbitrary 'alert' level based on experience whereby let's say 10 times the normal number would trigger an emergency review by the management team. In terms of measuring complaints I do prefer to analyse by complaints per million units as complaints levels will naturally rise with volume.

As well as the type of defect what you do is also dependent on the shelf life of the product. If it is close to expiry then a recall may have little impact but if it has 2 years left on the shelf then a recall would be much more sensible.

Regards,

Tony

Hi Tony,

I like yr suggestion (qualitative statistical control?) and hope the resources of OP can do it.

Of course sometimes gross blunders cause a non-safety defect where one instance is a catastrophe . Numerous encounters come to mind. :biggrin:

Unfortunately the product / process / sensitivity / current reputation  is here unknown.

Rgds / Charles