Q.5 RFC Component v. Unit and 3rd party testing

[Eric H]

Ah, this is the big one, isn't it? I have a theory that the lack of control over their contractors is probably part of what caused the big recalls of 2007.

You have a problem first of making sure the contractor isn't substituting components. They either have to supply tested components, and you have to have some check on that, or you have to supply everything to them and make sure they don't shade on you. I suppose if you're working with foreign contractors, you're going to have them do all of the sourcing. And here your troubles begin....

So I suppose if you did the sourcing, then you need to check the component testing. If they do the sourcing, you are still responsible for the component testing as the manufacturer. So what do you trust them to do? Heck, given enough time, I'm sure I could generate a lab test from Bureau Veritas in pdf format that shows fishing weights tested lead-free. So how do you trust your contractors when they are sourcing? And how does a regulator trust you to do due diligence on this point? I suppose you could specify the testing lab, but then how do you know that they aren't using ringers for test purposes? And then how do you keep from exposing your entire supply chain from everyone else in the market?

[Vesta]

Yeah, I'm sure you're right about this being the root of the problems in 2007. Obviously, if Mattel made all of their toys at a company-owned factory, there would never be lead introduced into the process (or brought onto the factory grounds, for that matter).

So either you source and arrange for the component parts to be sent to your factory (which is just not how it's usually done, on a large scale - that's part of the joy of having a complete package provider).

Or you spot check finished units for compliance. And you pull units to check from your TOPs, not from samples provided for testing. (sigh)

Honestly, I can't believe Mattel wasn't doing this already. Lazy. Even we do this.

[Eric H]

TOP?

[Esther]

I don't know where the Mattel break down happened. I suspect it is because the approved factory did a bait and switch. A common problem in China, where factory workers are taught to lie during inspections and only a small portion of an order is produced in the approved factory.

As far as toys go, batch testing by a third party *MAY* solve some of this problem. But again, if only a portion is produced in the approved factory, it would be difficult, if not impossible, to test the rest because you may not know where it is. Mattel would be stuck with the unenviable task of testing batches (what determines a batch?) after it's all produced and before delivery.

[Vesta]

Sorry, Top Of Production. When contracting out production, one should receive the first pieces off the production run (one of each SKU), made with the final components, by the actual production process to be used for the full run. It's the last chance to catch a problem before the QC inspection on the finished products.

[Melissa McKeagney]

This is the hardest question for me, because I don't contract out, and have no experience with the manufacturing process outside my own basement.

What about randomizing your testing. Instead of pulling the first few from a production run, you pull and spot check somewhere down the line. That way if they have substituted some component for another, you ought to be able to tell from a visual check of your sketch sheet swatches whether anything has been substituted.

As for toys, or other products, this might be harder to determine. Whether a certain paint has been swapped for another, it seems to me it would be difficult to tell without doing unit testing. But, if you did do unit testing on one of the first few from the production run, and then later on a switch was made, unit testing wouldn’t have made that product any safer. It really comes down to random quality control, and using a contractor that is trustworthy.

What about having an exemption for certain countries where this has not been a problem, and random unit testing for places that have run into problems with lead, ie. China. In other words, there is no historical reason for not trusting products made in Europe, Japan, the US, and Canada, etc., but we do have reasons for not trusting products made in China and some other South East Asian Countries. So, while you can get products made cheaply in those countries, the manufacturer must pay a certain price, ie random unit testing for using contractors in those countries.

Melissa

[Alison Cummins]

(Jen, you are the manufacturer; you order goods and have them shipped directly to your contractor.)

In theory, you have fewer worries than someone who relies on their contractor to source. However, as Eric points out:

[J C Sprowls]

Actually, there is a real risk - low, though it may be.

There are times that a contractor managing a customer's inventory can make mistakes. An example might be more applicable to small components (e.g. zippers, non-branded buttons, fusible, etc.).

If the warehousing area isn't well managed, it's a possibility that someone else's buttons might be used on someone else's order. The contractor might discover this error and scurry to replace the buttons as quickly as possible. If they do not have the contact information for the original source, they'll search several other suppliers for a substitute.

[Melissa McKeagney]

So how do you manage that JC? How do you ensure that the tested component is the one that is actually used?

Melissa

[Eric H]

Ha! I knew it. I said, "I could generate a lab test from Bureau Veritas in pdf format that shows fishing weights tested lead-free".

I'm looking at an e-mail from someone who was going to get their logo printed on stainless steel water bottles. They decided to request testing.

[cinthia]

Hi Eric -
I wrote that email - let me follow up. I spoke with the testing company branch in the USA - and they suggested that I open an account with them under my company name. Since I'm paying for all this testing anyway - the factory will submit the samples to the local branch but the results will only be sent to me - not to the factory.

[J C Sprowls]

[Esther]

[Kathleen F.]