David Dayton’s latest article illustrates the bad practices of the largest testing laboratories in China. It really shows that something is “broken” in the sourcing process.
The dynamics of the consumer products testing industry is driven by a few large actors (the regulators of importing countries, the mega-retailers in rich countries, and the 3 or 4 biggest certification companies). As often in China, it results in a system that is consistently gamed by the most savvy players, at the expense of all the small importers.
I listed some dark practices adopted by most of the large testing laboratories, and their perverse effects on the safety of Made-in-China goods.
High prices, to be borne by small & medium buyers only
At SGS, BV, and ITS (the three largest certification & inspection firms), the “consumer products” division is usually the most profitable… Thanks to laboratory tests. Margins are insane, simply because there is no real competition and no risk.
There are hundreds of labs, but only a handful of them have an international network and are “authorized” by large retailers or governments. In Dayton’s experience:
You can only use a testing company from a US govt approved list—and of course, those on the list charge more for their certification than companies who are not on the list. In addition to that, some large box stores also “strongly suggest” that if you want to place product in their stores you have to use a specific testing company.
Why do mega-retailers play this game? Because they benefit from it. Their suppliers are obliged to pay for testing in a specified lab… At very high prices (example: 60 USD for a simple drop test in their facilities). And in compensation, product inspections, which have to be paid by the importer, are artificially cheap (sometimes 50 USD).
No transparency of methods used
Most buyers do not realize that even the largest testing houses regularly subcontract certain jobs. A laboratory cannot do all tests on all products!
Add to this the poverty (or lack?) of internal control in these huge structures, and you get to inconsistent results:
When testing, the responsibility is on whomever requested the tests to make sure that everything is done correctly—and really, how in the world can you know if it was done correctly or not?! Unfortunately, the reality is that if there are problems in the methodology, you’ll never be told about it.
Can you imagine requesting to be present when your samples are being tested? I doubt it would be accepted, especially when tests are subcontracted.
Profit motives get in the way of proper conduct
Testing laboratories are for-profit companies. And some of them are more “flexible” than others when it helps in getting more orders.
Let’s say you work in a Chinese factory. You have been exporting the same type of products to the US and Europe. You have noticed that different testing laboratories (required by different customers) give you different test results for the exact same products. When given the choice, will you go for the toughest one? Of course not!
Dayton gives a first-hand account of this phenomenon:
We pulled and sent all the SAME samples for ALL three clients and mailed them ourselves to the two different testing companies so I can personally testify that NOTHING changed in the samples between tests. But the results were different—radically different.
If tests were really performed in controlled environments, under the relevant standards and with the right tools, this type of surprises should not be that frequent. I heard similar stories from other sources. There is something wrong, for sure.
I remember Paul Midler also denounces the ease of “circumventing third-party testing” in Poorly Made in China, a great read if you want to see how experienced factories can cheat naive importers.
No advice to small & medium clients
Some of my clients have no idea what tests they should perform on some products they buy from China. So I contacted several testing labs (who gave me different answers, of course). They gave me lists of supposedly compulsory tests. Sometimes it reached 2,000 USD in testing fees, when the whole order was only 10,000 USD…
In this case, any importer prefers to close his eyes and ask for a certification from his supplier (even though everybody knows it’s only a piece of paper with no value in 90% of cases). It is just not manageable for small orders.
So I asked for a distinction between “the tests that most often fail for this type of product” and “other tests that usually pass”. No way. QC firms run a huge liability if they advise what tests should be done, they forget one, and if it causes some safety issues after delivery. So they have an double incentive to stuff the list up.
Don’t ask a barber if you need a haircut, right? But if you want to source quality products that represent no danger, who should you turn to, if not to testing labs??
For advice related to lab testing:
Do not trust Chinese manufacturers’ certificates
5 tips for testing China products against regulatory standards
First of: great page, like reading your articles. For this one I finally have an opinion… . A biased one for sure, since I am the one from the dark shadows of a testing lab (insert dark laughter here)
a) Testing cost: So you want a report about all the relevant contaminants, which involves machinary in the range of 1 mil US$, and I take the risk as giving you the “all is fine” report, but since the order is 10.000 US$ you only can pay what? 300 bucks? At the end of the day it is like shopping for meat. Please I want the best meat without growth hormons, but I buy the cheapest available one where I keep my fingers crossed that no shortcuts are taken… . Right. If you want to know what is likely to fail – well, the same problem you have during inspection: I have seen so much shit, I do not know where to start… .
b) Different results. WELL, you know about statistics. Very well since you cite Deming from time to time (In God we trust – the rest has to provide data)… . This is always a cheap excuse, but yes it is definetly important.
So are there different results to be expected? For SURE. Try a fabric, take a role and make a colour fastness along the roll. What do you see? A homogeneous product with one result? Doubtful. I would be pissed not knowing about measuremnet uncertainty, but well that is another story and customers I know are NEVER interested in measuring problems. They want a pass.
c) Different results part II. Yes, it is frustrating. Do you think I like it when my result is different from another testing lab? Nope. That is the reason why I have a certification by a neutral 3rd party following e.g. 17025. That says “yes I do my work correctly”, after my lab got an in depth audit that takes a week and includes the normal operation in the lab. I can tell you about those costs… but anyway, all the testing labs are certified. Look at the results from a profiency test where you have to participate once a year and you see good fitting results for the most labs. Do we pay more attention to those samples? Or are the samples homogenous? I do not know for sure, and I have no time and no money to cross test with my competitors. But I can guarantee you, if you look in the compexity and see the statistics, it is not suprising that you see different results. And the methods or norms are not as detailed as you would hope for stating things like “take a method which is sufficient”… .
d) No time? But it takes for EVER when you test? Interesting, since in the textile world, the allowed time for me to report a test is 3 to 5 days. Ah, and if I need more material to get a meaningful sample for e.g. sewing thread, it is not available… Uh, and if it is a fail, I rather not pay… . I am so glad that I have crazy margins since there is NO competition between testing labs. Easy money. (again: insert dark laughter here).
Got a bit long, and have more to say – but anyway if you have never seen a “testing lab” from the inside – I would like to invite you 🙂
Renaud Anjoran says
Interesting to read your viewpoint! Thanks Stefan.