In Luxembourg, on 10 September 2015, two French trade federations stood before the Court of Justice of the European Union and argued for making a right very small. One spoke for the country's retailers and distributors; the other for its do-it-yourself and home-improvement stores. Both had the same trouble with the same rule. The law said that anyone who supplies an object must, when a customer asks, tell them whether it contains a chemical on an official list of hazardous substances. It set the trigger at a tenth of one percent by weight. The federations' whole case turned on a single question: a tenth of one percent of what?1
Of the whole thing, they argued. Take a bicycle. A chemical in the rubber of a handlebar grip is a real quantity in the grip and almost nothing when you weigh it against the machine it is bolted to. Measure the grip against the bicycle and the chemical drops below the threshold, the question never has to be answered, and the duty all but disappears from the shop floor. The same logic would empty it out of a car, a sofa, a laptop — anything built from many parts.
The judges held the other way. The threshold, they ruled, applies to each separate article an object is assembled from (the grip, the cable, the zip, the circuit board) and not to the finished product they combine into. So the question the law actually asks is not about the bicycle. It is about the grip. That is the rule a customer lives with now. You can point at one part of a thing you bought and require the shop to tell you which listed chemical is in that part.
1. A court fight over the chemical in a bicycle grip
The Court did one more thing in that judgment that matters here. Asked whether all this amounted to too much to place on a business, it described what the rule requires of a supplier as minimal in nature. That is a duty to pass on information the supplier is already meant to hold, not an excessive burden.1 That single line will come back near the end, because it turns out to describe the cheapest part of the whole arrangement. But first, the plain shape of the thing.
2. What a shop has to tell you, free and within forty-five days
Here is the duty in one passage. Any supplier of an article (and "supplier" includes the distributor, which means the shop that sold it to you) must, on request from a consumer, provide enough information to allow the article's safe use. As a minimum, that means the name of any substance on the official Candidate List of substances of very high concern that is present above a tenth of one percent by weight of that article. Free of charge. Within forty-five days of receiving the request. And where no such substance is present above that threshold, no duty to reply arises at all.2
That last clause is not a loophole; it is the whole design in miniature. The right is not a promise that a product is clean. It is a promise. If a listed chemical is in it above the line, and you ask, the shop has to say so. Which raises the only question that matters once the law is on the page: if someone asks, does an answer come back?
3. Twenty thousand people asked, and most asked once
The best record of what happens when people actually ask sits in the logs of a German environmental charity. Between October 2016 and October 2018 the charity ran a mailbox: a person scanned a barcode on their phone, and an app called Scan4Chem drafted the right-to-know letter and sent it on their behalf. Every request left a trace.
Four researchers went and read the traces. Julian Schenten and Martin Führ, at Darmstadt, with Mattheus Brenig and Kilian Bizer at Göttingen, obtained the logs of the mail processor the charity used and counted, request by request, what the users had done.3 Of the 20,634 people who sent a request through that app in those two years, about six in ten sent exactly one. They never came back. Roughly nineteen in a hundred sent two. Five hundred and forty-two people sent ten or more. Seven in ten did everything they ever did on a single day. First-day requests account for almost two-thirds of every request in the file.3
So this is not the record of a right that failed under heavy use. It is the record of a right that people tried once, in a single sitting, and then set down. Which is a strange way to treat something free. Why would so many people reach for a right and let go of it that fast?
4. Most people assumed the question was already settled
Susana Fonseca went and asked them. A sociologist who coordinated the project in Portugal, she ran three consumer surveys and forty-four interviews, and a good deal of what she wanted to know was why people who could use the right did not.4 The answers came back with a pattern in them.
A woman in Luxembourg, between thirty-one and forty-five, on whether she had ever asked: "No, and to be honest I thought that what is sold is not harmful by definition." A respondent in Latvia, the same age band: "I have always been quite rights oriented, I know consumer law, so I thought I would have those rights, but I had not thought about it at all because I thought that EU regulations only allow safe products on our market." A respondent in France, again thirty-one to forty-five: "I did not know this right, and I thought that these substances were not present, being dangerous, they do not have to be in the articles I consume."4
The same belief turns up on the other side of the Channel, in numbers. When the UK government surveyed 3,667 people about this right, 87% had never asked a business for this kind of information. Only just over a quarter (27%) were aware the right existed at all. Asked what they expected of the system, 89% thought chemicals regulation should ensure the products they use would not cause them chemical-related harm. The government read that as a strong degree of consumer trust in chemical regulation.5
Read those quotations again and the thing they share is not ignorance exactly. It is a settled assumption that the question had already been asked and answered by somebody upstream: that dangerous things are not on sale, so there is nothing to check. That is not a foolish thing to believe. It is the thing a regulated market is supposed to earn. But it raises a sharper question than simple unawareness would: if people did know the right existed, would they use it?
5. Knowing more chemistry did not make people ask
Two researchers, Hartmann and Klaschka, went looking for exactly that relationship, and expected to find it. They surveyed 1,321 people who had signed up as interested in harmful substances in everyday products, and set out to test an obvious hypothesis. It held that the people who knew the most chemistry, who worked with chemicals, or with REACH itself, would use the right more than everyone else.6
When they looked, it was not there. In their own words, "there was no correlation between chemical knowledge or experience with chemicals and usage of the 'SVHC right to know'." And they published what they made of it, which is the interesting part: "people with good knowledge in chemistry have other means to inform themselves, so that this information instrument does not play a big role for their personal risk assessment of everyday articles." The people best equipped to ask did not ask more, because they had other ways to find out — a finding about who this particular instrument serves, not a claim that knowing anything makes no difference.
The same survey caught the gap between wanting and doing. Every second participant said they were interested in finding out more about a substance in a specific article; and, as the authors put it, "not more than 4% of all participants inquired for SVHCs with various methods." Their conclusion was blunt: consumers "would like to know more, but tend to avoid the burden of making investigations." The AskREACH project reached the same place from a different road, reporting that many consumers do not, and here they put the word in brackets, do not (want to) think about these substances when buying, because they firmly assume that what is on the market is safe.7
This is the hinge of the whole thing, and it is worth being plain about why. If hardly anyone asks, and the reason is not that they never heard of the right, and not that the people who understand the chemistry best are somehow put off — if the reason is that the task itself is effortful, delayed and usually turns up nothing — then "hardly anyone asks" stops being a verdict on the people and becomes a fact about the design. Keep that in hand. It changes what every number after it means. Because now the question narrows to the few who did push through: when someone found the will to ask, where did it actually get hard?
6. The hard part was finding who to write to
Not the asking. The addressing. The people who tried and struggled were remarkably consistent about where the wall was, and it was not the letter — it was working out who the letter should go to.
A respondent in Poland, between thirty-one and forty-five: "I spent more time looking for contact details than using the app, so the level of usefulness is not great." Another, the same country and age band, on the same wall: "…if I have to spend 5 minutes standing there next to the product on the shelf just to find a producer's or supplier's email address, well, I just don't have time for that." A third, in Poland, younger: "When I had time and I really cared about checking the product, I tried to find the contact information myself and fill them in, but more often than not I dropped it."4
The project's own plumbing tells the same story as arithmetic. Of the contact addresses the app relied on, 363 suppliers had registered their own details; the other 14,419 had to be found and keyed in by hand by the project's partners.7 The address was the scarce thing, not the willingness. Fonseca's summary of what her interviewees told her lands exactly there: "the process of sending requests on the database is easy. The difficulty increases if the contact information is not available."4
Hold onto that, because it has a consequence the app's designers spent years discovering and a reader already carries in their pocket. But first, the question this plants: the people who did find an address and did write — did an answer come back?
7. Four studies measured whether firms answer, and got four answers
Four separate attempts have been made to measure that, and they returned four different results.
The UK government commissioned a survey from Kantar Public in January 2022, in which 483 people who said they had made such requests recalled 1,164 occasions between them. On 73% of those occasions, they reported, the information they had asked for arrived. That figure the government's own methodology note calls "susceptible to bias," providing "an indication of national knowledge and experience, rather than allowing for firm generalisations." The report twice describes it as evidence that suppliers complied with their duties under Article 33(1), the business-to-business duty, even though the survey asked consumers about requests they themselves had made, which is the separate consumer duty, Article 33(2).5
Schenten and his colleagues read the German app's logs and found something dimmer. Letters drew a reply about two-thirds of the time. But once the charity BUND assessed each reply against what the duty actually requires, information on the substance reached the database for, in the authors' own words, only one-fifth of requests. Their abstract ends: suppliers "are struggling to reply to right-to-know requests."3
Fonseca asked the app's users directly, three times over, and heard something dimmer still. Among those who had sent requests, she reported, "the most common situation is that they never received an answer from producers or retailers (average of 50%)." That figure improved across the three surveys, with 32% by the third saying they got answers always or at least half the time, against 28% in the first.4 That is what app users reported receiving, which is not the same as what suppliers, as a class, actually do.
And an accredited laboratory ran the one test that paired a consumer's request with a chemical test of the thing itself. It bought articles on the open market, sent a consumer right-to-know request first, and then tested what it had bought. Where the supplier had already published the article to the project's database, testing found one discrepancy in 73. For articles no supplier had published, the assumed non-compliance ran from 10% to 50%. And of the articles that turned out to contain a listed substance above the threshold, not one supplier had given that information to the consumer who had asked.7
Set them side by side and they will not sit still: seventy-three percent, one-fifth, half never answered, and a lab arm at zero disclosure. Four honest studies of one right. How can they disagree this much?
8. The answers disagree by who was written to
They disagree because they are not measuring the same thing. Each one wrote to a different address, and the answer tracks the address.
The government survey asked UK consumers about "products." The government chose that word deliberately, explaining in its own methods that it used "product" instead of the legal term "article" because consumers might struggle with the technical definition. The definition, it noted, is narrower than the everyday sense.5 A person who emailed a shampoo brand about an ingredient, or a food company about a preservative, would answer that survey yes. The right does not reach any of those things.
The German logs and Fonseca's surveys measured the app, and the app wrote to whoever owned the barcode. As Schenten's team found, 57% of the requests in the logs were sent about products the right does not cover at all, and many of the barcodes belonged to firms outside the area where the rule can be enforced — firms with, in the authors' reading, no particular reason to answer a letter from a jurisdiction that cannot compel them.3 The laboratory, for its part, measured goods pulled off the open market, sampled to look for trouble, not drawn at random.
Four instruments, four routes, four answers. Which is not a muddle to be averaged out. It is a finding, and it points somewhere specific: nobody in that table wrote to the one recipient with the clearest legal standing of all — the shop that sold the object to the person asking. So who did write to the shop, and what did they find?
9. No one measured a consumer writing to the shop
Nobody did. And the closest anyone came to trying is worth telling exactly, because it is easy to get wrong.
When the UK government reviewed how this right works, it identified three groups: consumers, who have the rights, and suppliers and retailers, who have the duties. It designed three separate surveys, one for each. The retailer survey (built, in the department's words, "to collect evidence from those at the end of the supply chain selling directly to consumers") "received too low a response to allow meaningful analysis and inclusion in the report," and the review records the retailer's perspective as one that was, in its own phrase, missing from the review.5 The supplier survey that did run went out through a regulator's chemicals eBulletin, to subscribers who had signed up because they follow chemicals regulation for a living. Nearly two-thirds of the suppliers who answered said they were both suppliers and retailers.
This investigation continues below.
Want the next one in your inbox?
So this is not a case of a government asking the shops and being ignored. The shops were asked through a channel that reaches chemicals specialists, not the general run of retailers, and not enough came back to analyse. That is a survey response rate, not a measure of whether shops answer their customers. What is true, and only this, is that none of these studies isolates the one route with the clearest legal standing — a consumer writing to the shop that sold the article — as a measured condition of its own. There is an empty cell in the table, and it happens to be the cell a reader stands in.
The researchers who ran the app had already worked out that this was the cell that mattered. The app's weakness, Schenten's team argued, was that it wrote to the barcode's owner, often a manufacturer abroad, when it could have written to the retailer in the country where the rule bites. Retailers are suppliers too, they pointed out, and are obliged under the same right.3 The best-placed recipient and the unmeasured one turn out to be the same shop.
10. Telling people was necessary but not enough
On 25 January 2023 the five-year project that had built the app, the database and the campaigns held its closing conference in Brussels, and gave it a title that was itself a conclusion: "Art. 33 REACH – A right to know, not a right to ask!" In the room were people from industry, from public authorities, from environmental and consumer organisations and from research bodies.10
Susana Fonseca, who had run the consumer research, stood up in front of them and said the quiet part out loud. "AskREACH is a stunning example," she told the room, of "how mere information provision to the consumers does not automatically lead to behavioural change," and "politicians need to understand that consumers can never be the sole engine of the transition to SVHC free products."10 Heidrun Fammler, deputy manager of the Baltic Environmental Forum in Germany, put the same finding in the language of what it had cost: "Making Art. 33 work is a tremendous communication task… Having spent more than 5 Mio. euro for staff, we can say: It is people that make Art. 33 work."10
That deserves to be read for what it is. This was the largest Article 33 awareness effort anyone has documented, and the people who ran it reached a hard-won conclusion and said it to a room, against their own interest: awareness was necessary but not enough. Telling people the right existed did not, on its own, make an ask-and-wait design easy enough to use. What they were naming is not a failure of the public. It is a fault in the machine's blueprint: a right you have to ask for puts the whole burden of asking on the person least equipped to carry it, and leaves untouched the people who chose that arrangement, the categories it exempts, and whoever put the substance in the object to begin with.
If asking is not the engine, though, that leaves a reader who did ask holding a practical problem. Nothing came back. What does that mean, and can they do better?
11. Three reasons nothing comes back, and two you can fix
Silence has three causes, and a letter cannot tell you which of the three you got. The request may have reached someone who does not owe the duty at all. In the German logs, 57% of the requests were about products the right does not cover, and many went to firms outside the area where the rule can be enforced, who have no reason to reply. The question may have been asked about the whole object instead of a part, which is not how the threshold is measured. Or there may be nothing on the list above the threshold in that article, in which case no duty arises and nothing is owed.7
Two of those three are yours to remove, and they are removed by asking better rather than by asking again. Ask the shop that sold you the object. A distributor is a supplier, the supplier owes the duty, and the shop sits in the country where the duty can be enforced. That is exactly what the barcode app so often could not manage. Name the article and name the component: the grip, the zip, the cable, the board. The part is how the tenth of a percent is measured. And check the thing is something the right actually reaches. It must be an article, not a cosmetic, a food, a detergent, a medicine or a liquid, which are all outside it. The third cause is not a problem to solve. If nothing on the list is above the threshold in that article, no duty arises, nothing is owed, and nothing is exactly what should arrive.
That reframes what the letter costs the shop, too. The duty to know sits upstream of any letter. It is a standing obligation on the business, and the Court has already priced the answering part of it as minimal. When the UK government surveyed suppliers, 129 answered (a number it states twice is not a representative sample, and whose answers, it cautioned, "was reliant on suppliers' accurate reporting of their experiences… which when talking about compliance could produce some social desirability bias"). Among those who confirmed they supply articles above the threshold, 62% said they already provide more than the legal minimum. Even so, 64% called providing it a burden and 42% called it too expensive.5 Schenten's team, reading the logs, saw the same readiness from another angle: many firms answered requests even where the article contained nothing above the threshold and no reply was owed at all.3 Whether the information those firms hand over is always correct is, as Schenten's team themselves note, at least questionable — which is the right thing to wonder next, and it turns on what a clean answer actually rules out.
12. A clear answer covers one list, and the list keeps changing
Say the letter works and the answer is a clear no: this article contains no Candidate List substance above the threshold. That is a complete and lawful answer, and it is worth knowing — but it is an answer about one official list, measured part by part, and everything not on that list sits outside the question. A no is not a claim that nothing else is present, or that anything is harmful, or that anyone has misled anyone. The two laboratory arms make the point between them and refuse to resolve it: where firms had published, the testing largely agreed with them; where none had, the testing turned up substances no supplier had disclosed. Neither population answers for the other, and so the honest place for a reader's doubt is not the honesty of the shop that replied — it is the boundary of the list itself.7
And that boundary moves, which is the reader's own leverage, not a reason to shrug. Fifteen substance or group entries joined the Great Britain Candidate List on 15 June 2026, after a consultation that opened in March.8 Among them: the photoinitiator used in UV-cured gel nail products; a fragrance ingredient; a group of boric acid sodium salts; the standard flame retardant in printed circuit boards; a stabiliser used in PVC. Several have documented uses in everyday things: nail products, fragrances, electronics, and PVC. The list is a public file: anyone can open it and read what is on it and when each thing arrived. Which is how a reader knows that a no given last year is an answer about a shorter list than the one that exists this month.
Two more things are worth a reader's knowing, and neither is a letter. Where a supplier who does owe the duty gives you nothing, the route is not a second letter but the enforcing authority. The government's guidance sends concerns about a retail sale to the local Trading Standards office, and says that if you are unsure which authority to report to, they work together and will pass it on. Failing to comply with the duty is an offence, and the regulator publishes a register of the notices and prosecutions that follow.8 All of which is real, and all of which still leaves the reader doing the work. So is asking even worth it, if the answer is narrow and the list keeps moving?
13. The answer that arrives without a letter
The most useful thing to know is that the answer does not have to depend on the reader's diligence at all. The clearest evidence for that is what Europe did once it had watched the ask-and-wait model up close. From 5 January 2021, EU producers, importers and many distributors must file a SCIP notification with the European Chemicals Agency for any article above the threshold. It is structured information about the article and the substance, fed into a central database, and owed to nobody in particular. That is a notification duty laid on top of the old business-to-business obligation, requiring nobody to ask. AskREACH, which had spent five years building a voluntary version of exactly that, wrote that its own database now seemed redundant beside the mandatory one.7,9 The right to ask did not go anywhere. What changed is that some of the information started arriving somewhere without anyone having to send a letter to get it.
That is the property that tends to separate the transparency rules that work from the ones that mostly do not. The hygiene grade in the restaurant window is read by everyone precisely because nobody has to request it. It is simply there, in the frame, at the moment it is useful. A right that has to be exercised, one letter at a time, by the person with the least information and the least time, was never going to do that work — which is not a scandal, but a design, and one its own researchers named to a room.
Fonseca put it most plainly, and it is worth giving her the last word, because she had run the largest test of the alternative and earned it. The key, she concluded, "is to prevent the use of hazardous substances in products, because Europeans will never be experts on everything we want them to be," and the responsibility of those who place articles on the market, and those who regulate them, is to live up to the trust consumers place in them to provide non-toxic articles in the first place. The right to ask is real, and it is worth using. But the weight was never supposed to rest on the person holding the object.