I was fortunate enough to exchange a few emails with Dr. Stefan Droste, Managing Director of Hohenstein Laboratories (HK) Limited.
His laboratory is part of the Hohenstein Institute, which employs about 500 employees in Germany and in over 40 contact offices globally. This group is among the most important independent research and testing institutions in the textile sector.
Dr. Droste gives us some very useful and interesting information in the Q&A below.
Q: Let’s say an importer is looking for a testing laboratory. There are lots of options in China and in Hong Kong. What attributes should they concentrate on, besides the price?
Visit the lab if possible and get an idea of the sample preparation area. Is that clean and well-organized? Visits can be problematic in larger labs (e.g. safety reasons, protection of customer information).
- Are the methods I need accredited, e.g. by DIN 17025?
- Are they experienced in the market I want to deliver (US vs. Europe)? References?
- What are the turnaround times?
- Can they report in a format I need? (e.g. directly into your database?)
Do a blind sampling with failed samples from the past. Do they find the failures? Repeat this after the testing routine has kicked in.
Q: Some importers want their suppliers to pay for the tests (in a lab they nominate) that show their products are safe. Are there any risks if the supplier is in charge of this process, and how to mitigate them?
First, there is no way to identify how often a product was tested and what the changes in production were between different “failed” tests (if any changes were made at all). Since your supplier pays, the test reports are his property and you have no chance of interfering – or even knowing that there were tests.
The 2nd risk is the pressure from suppliers towards the labs to be “flexible” with the results. The effect strongly depends on the integrity of the involved persons. Several big players I know do in-production tests locally, while the final testing is performed in Europe. In general, I would be sceptical if the products only fail by a landslide and never with a result only a little above the limit.
To really get around both problems, I recommend that the importer pays for the first test of a sample collected by a 3rd party, and if it fails, the supplier has to pay for every retest and is charged for the failed test. That way, you can rest assured that you do not miss problems and can ask for a corrective action taken by your supplier.
Q: Many importers place a PO for 10,000 USD, and don’t pay for any lab testing because all the “strongly advised” tests cost more than 1,000 USD. Are these buyers too small to source safely from China? What is your view on this?
The short answer: Yes, the order is too small. A little longer: search for a supplier who offers labels with the product, e.g. Oeko-Tex, but it will be a hard and time consuming task…
In the eyes of the authorities, you as an importer are responsible to offer safe products. That does not mean you have to test everything all the time, see for example, the Component Part Testing Rule in the US which generally allows reliance on testing/certification of raw material or physical components and finished products if “due care” and record-keeping requirements are met.
But that means, that you have good control over the supply chain and you are sure that nothing is changed, subcontracted, (is that REALLY possible in Asia?), etc.
At least you have to test high risk parameters. Define together with a lab what is really needed and what fits into your budget. For example, a test on AZO-colorants is not very expensive, just as a test for formaldehyde on wrinkle free garments, and so on. The money you have to spend in testing has to be budgeted in your P&L calculations. You just cannot expect that you are 100 % safe – that would be like paying for some QC and expecting a QA-program.
Q: In some cases, different labs find different results for samples that were drawn from the same lot. Where do you think these differences come from?
Every 17025 accredited lab has to prove once a year that the discrepancies to other labs are within a limited range. Still, there is always the possibility of mistakes in the lab, e.g. the samples were switched, wrong part taken, undetected method bias, complex method with labile components (thinking about some AZO here), etc.
Beside that, I would say that the samples or the chemistry on a fabric are not homogeneous – even if they look the same. Just compare it with some knitting machine. You expect to have some defects in the fabric, but why? The process is always the same.
So ask yourself: how many “assembly lines” were used in the production? How often were some components replenished, e.g. a new sewing thread, new barrel of ink, etc.? Often chemical reactions are involved and/or washing procedures. So, how homogenous is an 8m wide piece of fabric? The sample might be a lot less homogeneous than you think and this will show between labs. Never mind the fact that we are often measuring in mg/kg range, while sampling in the gram range – so we have to find µg/g in the sample, and the measurement uncertainty just increases with that.
Unfortunately, suppliers often “shop” around to test the garment until it has passed in some lab and will show this report. The easy answer though is not always the truth.