A sourced, structured investigative archive — animal welfare, fast fashion, factory farming, women's rights, environmental destruction, corporate accountability, and the advertising that obscures all of it.
Every article we publish is sourced from peer-reviewed research, government databases, or named NGO reports. We never speculate. We always cite.
Each industry has its own statistics, investigations, and archive entries. Most issues belong to several at once.
Connecticut's Supreme Court found that a 1922 state law banning unfair trade practices could be used against a firearms manufacturer whose marketing targeted civilians. The $73 million settlement, reached in 2022, opened a path no other Sandy Hook lawsuit had been able to find.
87 million Facebook profiles, a quiz app, and the regulatory questions still unanswered.
A police shooting, an arrested union president, and a 56% increase that fell short of what living-wage research recommends.
Defeat devices, an EPA notice in 2015, and a Goodhart's Law problem that recurred at multiple manufacturers.
The CSDDD, Germany's Lieferkettengesetz, France's duty of vigilance law — what changed when human-rights due diligence stopped being a brand choice.
“Clean,” “natural,” “conscious,” “eco” — the words that built a marketing decade now face statutory definition.
The Digital Services Act, the Click-to-Cancel rule, the long ledger of consent — what changes when the regulators stop treating tracking as a feature.
A short weekly brief from Ayden — what the archive added that week, the investigation in progress, and the figures worth knowing. No advertising. No filler.
Select a topic and watch one of our pre-written research summaries appear. This is the format every published Feral Femme summary follows.
Structured field notes on documented issues. Each entry lists what happened, the evidence, the named sources, and what has changed since.
On 24 April 2013, an eight-storey commercial building in Savar, Bangladesh, collapsed. 1,134 garment workers, the majority of them women, were killed. More than 2,500 were injured. The building had been evacuated the previous day after cracks appeared in load-bearing pillars. Factory managers ordered workers back in the next morning to meet Western brand deadlines.
The Accord, then the International Accord, established legally binding building-safety oversight that did not previously exist. Compensation funds were paid out, though Clean Clothes Campaign documented years of delays in full settlement. Bangladesh's minimum wage for garment workers was increased — most recently in late 2023 to 12,500 taka per month, a level that union federations including IndustriALL described as still below a living wage.
Mulesing is the practice of cutting skin from the rear of a Merino lamb to prevent flystrike. It is performed on the majority of Merino sheep in Australia, which supplies most of the world's Merino wool. The procedure has been illegal in New Zealand since 2018. Australia, the world's largest exporter of Merino, has not banned it.
"Merino wool" on a garment label does not indicate whether the sheep was mulesed. Where brands disclose, they do so under voluntary schemes (Responsible Wool Standard, ZQ, New Merino) — none of which are required by law in any major market. Consumer Reports and Four Paws have both documented the gap between brand commitments and traceable, certified non-mulesed supply.
"Ag-gag" is a class of US state laws that criminalises undercover photography, filming, or employment at agricultural facilities — laws aimed not at what happens inside the facilities but at the documentation of it. Several have since been struck down on First Amendment grounds. Others remain.
Ag-gag is, on its face, an animal-agriculture issue. It is also a press-freedom issue (criminalising journalism), a labour-rights issue (criminalising whistleblowing employees), and a consumer-information issue (limiting what the public is allowed to know about how its food is produced).
No quotas. Field notes publish when the evidence is on public record and the sources are named.
Roughly 70% of the world's cobalt is mined in the Democratic Republic of Congo. A substantial fraction of that comes from artisanal mining — small-scale, hand-dug operations where workers including children dig in tunnels without safety equipment. The cobalt enters global supply chains supplying consumer electronics and electric-vehicle batteries.
The Responsible Minerals Initiative was established and refined its audit standards. Major brands published cobalt-supply-chain due diligence statements. The EU Battery Regulation (2023) introduced mandatory supply-chain due diligence for batteries placed on the EU market, with cobalt explicitly named. The EU Corporate Sustainability Due Diligence Directive (CSDDD, 2024) extends mandatory due diligence to a broader range of products. Critics — including RAID and Amnesty — argue the audit-based model has not fundamentally improved conditions in the artisanal sector.
In January 2023, a joint investigation by The Guardian, Die Zeit, and the non-profit SourceMaterial published a multi-year analysis of Verra's rainforest carbon credit programme — the largest in the voluntary carbon market. The investigation concluded that a substantial majority of the credits assessed did not represent the emissions reductions they claimed to.
Several major buyers quietly withdrew "carbon neutral" claims tied to forestry offsets. The voluntary carbon market saw a sharp drop in trading volume through 2023 and 2024. The EU Green Claims Directive and the UK CMA's Green Claims Code enforcement now treat unsubstantiated "carbon neutral" labels as actionable consumer-protection breaches. The underlying market structure that incentivised low-quality credits has not been fundamentally restructured — though Verra's revised methodology, published in late 2023, claims to address the worst of the over-issuance problems.
In the European Union, animal testing on finished cosmetic products has been banned since 2004; on ingredients since 2013 (Cosmetics Regulation 1223/2009). In the United States, there is no federal cosmetics-testing ban. "Cruelty-free" has no agreed legal definition. The brand may be telling the truth. The ingredient supply chain may be a different story.
The number of countries with cosmetics testing bans has grown steadily — Mexico (2021), Australia (2017 federal ban on testing in country; imports a separate matter), Colombia, Brazil (limited). China's 2021 reform was the most consequential single change in over a decade, given the size of that market. Major brands have committed to specific certification standards. The gap between "brand cruelty-free" and "supply chain cruelty-free" remains. Consumers depend on certification — Leaping Bunny, Beauty Without Bunnies, Choose Cruelty Free — and the certifications, in turn, depend on the brand telling the truth to its auditors.
In July 2020, The Sunday Times published an undercover investigation in Leicester garment factories supplying the fast-fashion retailer Boohoo. Workers were documented being paid under UK minimum wage, in conditions that did not meet basic safety standards. The factories had been audited. The audits had not flagged this.
The Leicester garment industry has been the subject of sustained regulatory attention since. Boohoo published a supplier list. The UK Government commissioned a separate independent review (the Apparel and General Merchandise Public & Private Protocol). The wider point — that brand audits and Modern Slavery Act statements had not surfaced what existed — became a critical example in the case for mandatory due diligence, which the EU CSDDD (2024) now requires.
In the textile mills of Tamil Nadu, a labour scheme known as Sumangali ("happily married woman") recruits young, often unmarried women — many from Dalit or rural backgrounds — on multi-year contracts. They live in mill-controlled hostels. The lump-sum payment is framed as dowry. The ILO and multiple human-rights bodies have classified the arrangement as a form of bonded labour.
Some brands withdrew from Sumangali-linked suppliers. Tamil Nadu state introduced (limited) enforcement mechanisms; the Madras High Court ruled in 2019 against certain restrictive contract clauses. The scheme has not been ended. As recently as 2023, SOMO and partner organisations have continued to document its variants in operation. The structural conditions — low rural wages, high dowry expectations, and mill demand for compliant young female workers — remain.
In January 2020, the Norwegian Consumer Council (Forbrukerrådet) published Out of Control, an investigation into the real-time bidding advertising infrastructure that underwrites most "free" online services. By tracing data sharing from ten popular apps, it documented how a single user's profile is routinely shared with hundreds of third parties within seconds.
Several apps tightened their privacy practices following the investigation. GDPR enforcement actions against real-time bidding intensified across the EU. Apple's App Tracking Transparency framework (iOS 14.5, April 2021) materially reduced cross-app tracking on iPhone. Google announced — and then repeatedly delayed — its third-party cookie phase-out on Chrome. The underlying business model of surveillance-based advertising is, however, broadly intact, and the regulatory enforcement remains uneven across jurisdictions.
In October 2017, France's "Loi 2016-41" came into force. It requires commercial photographs of models whose body shape has been digitally altered to carry the label "photographie retouchée." Norway extended a similar law to influencer content in 2022. The UK Advertising Standards Authority has, on a case-by-case basis, banned individual ads for "irresponsibly thin" portrayals. Together these regulations mark a slow regulatory acknowledgement that advertising images have measurable downstream effects.
Disclosure has not eliminated retouching — and may not have substantially reduced it — but it has changed which images can be presented as natural. Brands operating in multiple markets now manage retouching policies at portfolio level. The EU Digital Services Act extends transparency obligations to very large online platforms, including disclosure obligations on advertisers. The US Federal Trade Commission's 2024 rule on AI-generated endorsements is the first major regulatory response to deepfakes and synthetic models in advertising. The broader cultural question — whether a regulatory disclosure changes the underlying advertising harm — remains contested in the literature.
In February 2022, families of nine Sandy Hook victims reached a $73 million settlement with Remington — the manufacturer of the rifle used in the 2012 elementary-school shooting. The case did not turn on product safety. It turned on advertising. Specifically, on whether Connecticut's Unfair Trade Practices Act applied to the way Remington had marketed the weapon.
The legal mechanism — state consumer-protection law as a route around federal preemption — has implications well beyond firearms. It is the same framework EU consumer authorities have begun applying to greenwashing under the Empowering Consumers Directive. It is what gives the UK CMA's Green Claims Code its enforcement teeth. The principle is that a company is responsible for what its marketing says, regardless of whether the underlying product is legal.
A new generation of ultra-fast-fashion brands — Shein, Temu, Cider — operate at speeds that traditional supply-chain auditing cannot keep up with. New SKUs launch in the thousands per day. The data that would let consumers, regulators, or NGOs verify labour conditions is structurally not produced.
The US in 2024 began considering de minimis reform; the EU launched investigations under the DSA; France passed a 2024 fast-fashion bill imposing per-item taxes on ultra-fast-fashion imports. None of this has changed the production conditions documented in the source factories. The fundamental tension — that the speed-to-market business model is incompatible with the audit timelines required to verify supply-chain conditions — remains unresolved.
Per- and polyfluoroalkyl substances — PFAS — are a class of approximately 10,000 industrial chemicals used for stain-resistance, water-resistance, and grease-resistance. They do not biodegrade in any practical timeframe. They have been documented in human blood serum worldwide. Several have been linked to specific cancers, immune dysfunction, and endocrine disruption.
Stain-resistant carpets and upholstery. Waterproof outerwear. Grease-proof food packaging (fast-food wrappers, microwave popcorn bags, pizza-box liners). Non-stick cookware. Long-wear and waterproof cosmetics. Dental floss. The Maine law's "intentionally added" framing acknowledges the chemistry: PFAS are everywhere in the consumer goods marketplace because they work, and removing them requires substitute chemistry that for many applications does not yet exist at scale.
In April 2018, Australian broadcaster 60 Minutes aired footage from the live-export vessel Awassi Express. The footage showed sheep dying of heat stress in their thousands during voyages to the Middle East. The political response — building on years of advocacy by Animals Australia and the RSPCA — has reshaped the Australian live-export industry.
Specific, evidenced documentation — not abstract advocacy — drove legislative change in two countries. The Australian phase-out is the largest legislative restriction on animal-agriculture practice in the country's history. The mechanism mattered: video footage produced by a crew member with an animal-welfare conscience, given to an NGO with the legal and political infrastructure to translate it into reform.
In September 2015, the US Environmental Protection Agency issued a Notice of Violation to Volkswagen. Diesel vehicles sold by VW had software — a "defeat device" — that recognised when the car was being emissions-tested and adjusted the engine accordingly. On the road, the cars emitted up to 40 times the legal limit of nitrogen oxides. The catch had taken years of work by a small research lab.
A regulatory regime that tests products under standardised laboratory conditions creates a Goodhart problem — the test becomes the target, not the underlying performance. The Dieselgate response (real-driving emissions testing) shifted the regulatory model toward conditions that more closely match real use. Equivalent shifts are now happening in cosmetics testing (real-world contamination data versus controlled lab samples), in fashion lifecycle assessment (in-use environmental impact versus factory-gate), and in algorithmic accountability (the EU AI Act's requirement for ongoing post-market monitoring).
Between October and December 2023, garment workers across Bangladesh's industrial belt staged the largest sustained labour action since Rana Plaza. They were demanding a near-tripling of the minimum wage from 8,000 to 23,000 taka per month. At least four workers were killed by police. The wage offered — and ultimately set by the government — was 12,500 taka. The strikes have continued sporadically since.
Brand statements supporting "living wages in principle" mean little if the same brands continue placing orders at unit prices that, by the suppliers' own accounting, would make a living wage commercially impossible. The Worker Rights Consortium has published price-disaggregation models showing that doubling Bangladeshi garment wages would raise unit production cost by approximately 1–4% depending on garment category. Whether brand or consumer absorbs that cost — versus the worker continuing to absorb its absence — is a question with a price tag attached, and the price tag is known.
In March 2018, The Observer, The New York Times, and Channel 4 News jointly published an investigation based on documents from whistleblower Christopher Wylie. The data of approximately 87 million Facebook users had been harvested without their consent and used to build psychometric profiles for political advertising in the 2016 US presidential election and the UK Brexit referendum.
The technical capability to build psychometric profiles from social-platform data has not been removed. The major platforms now operate with tighter API restrictions, but the underlying inferential techniques continue to be developed and deployed — particularly via real-time bidding ad infrastructure (see Norwegian Consumer Council's Out of Control). The Cambridge Analytica case made the practice politically uncomfortable. It did not end it.
Between March and September 2020, COVID-19 outbreaks in US meatpacking plants resulted in at least 269 documented worker deaths and approximately 59,000 infections, according to investigations published by the Food & Environment Reporting Network and the US House Select Subcommittee on the Coronavirus Crisis. The industry was designated "critical infrastructure" via a presidential executive order. The workforce — disproportionately immigrant, refugee, and people of colour — was required to continue working.
The US meatpacking industry has consolidated dramatically since the 1980s. Four companies — Tyson, JBS, Cargill, Marfrig (National Beef) — control approximately 85% of US beef processing, per USDA reporting. This concentration was a precondition for the COVID-era policy response: with a small number of large firms, the industry could be designated critical infrastructure with one executive order. The downstream consequence was a workforce with no realistic exit option and a regulator with no meaningful enforcement.
"Parts pairing" is the practice of cryptographically tying a component to a specific device's serial number. Install a third-party replacement and the software refuses to recognise it, or restricts the device's functions. iFixit, the European Commission, the US Federal Trade Commission, and PIRG have all documented this as the central obstacle to independent repair — and therefore to extending the useful life of consumer electronics.
A consumer electronics device's lifecycle emissions are dominated by the manufacturing phase — typically 70–80% of total embodied carbon. Every device repaired is a device not manufactured. The UN Global E-waste Monitor 2024's figure of 62 million tonnes of annual e-waste is, in significant measure, the consequence of products that could have been repaired and were not.
From 2018 onward, the UK Advertising Standards Authority — working alongside the Competition and Markets Authority — has built up the most comprehensive enforcement body of work on influencer disclosure in any major jurisdiction. Named accounts have been publicly censured; an Influencer's Guide has been published, revised, and enforced. The model is not perfect but it is, demonstrably, working.
The ASA model is consequential because three elements coexist: a published Code with specific rules; public, named enforcement (a censure against a named influencer is searchable and reputationally durable); and continued monitoring so that compliance does not regress. The US FTC has the regulatory authority but issues warnings less publicly; the EU has expanding jurisdiction but variable national enforcement. The ASA — for all its limits — has built a precedent body of work that other regulators are increasingly citing.
A curated library of investigative reporting from named publishers and NGOs — The Guardian, Reuters, Associated Press, The New York Times, BBC, Humane Society International, Amnesty, Greenpeace, Fashion Revolution, the UN, and others doing the slow, rigorous work. Each entry is summarised in our editorial voice, attributed to its publisher, and links out to read the original at source.
A note: we link to the original publishers — we do not republish their work. All links open at the named outlet’s site. Some pieces are paywalled; where a link breaks, search the title at the named outlet. Entries span from foundational, long-running investigations to recent reporting.
Longer-form essays that connect multiple sourced reports into a single picture. Shorter posts that respond to whatever is currently being marketed at women. All in the same restrained, evidence-led voice the rest of the site uses.
A look at how the beauty market — McKinsey valued it at roughly $446 billion in 2023 — sustains itself by manufacturing the dissatisfaction it then sells the cure for.
Most industries solve a problem someone already has. The beauty industry, almost uniquely, makes its money by first persuading you that you have a problem at all.
This is not a new observation. Dove's Real Truth About Beauty report — first conducted in 2004 and repeated several times since — found that only a small minority of women describe themselves as "beautiful." The same campaigns that turned this finding into a marketing platform were, of course, run by Unilever, which also owns brands selling skin-lightening creams in markets where they remain commercially significant. The contradiction is the entire shape of the industry.
The structure repeats across categories. An advertisement, a billboard, an Instagram filter, a retouched cover image — each presents a face that does not exist in nature. The viewer compares her face to it, and the comparison produces a small dissatisfaction. Into that dissatisfaction, a product is sold. The product is presented as the cure, but the marketing was the cause.
The mechanism is documented. The Dove Self-Esteem Project's research links exposure to retouched media with measurable body-image consequences, particularly in adolescent girls. France's 2017 law (Loi 2016-41) now requires "photographie retouchée" disclosure on commercial images where a model's body has been digitally altered. Norway extended a similar disclosure rule to influencer content in 2022. The UK's Advertising Standards Authority has, since 2019, prohibited ads that promote harmful gender stereotypes (CAP Code Rule 4.9). Several countries — including Israel since 2012 — have set minimum BMI thresholds for models in commercial advertising.
What these regulations have in common is an acknowledgement that advertising images are not inert. They do something to the people who see them, and what they do — particularly to younger women — is documented enough that several states have decided the harm warrants legal intervention.
McKinsey's State of the Beauty Industry 2023 put the global market at approximately $446 billion. Statista, GlobalData, and others publish similar mid-hundreds-of-billions figures depending on category inclusions. WARC's Global Ad Trends reports beauty as one of the largest ad-spend categories worldwide — somewhere on the order of $220 billion in 2023.
To put that in scale: global humanitarian aid in a given year sits in the tens of billions. The beauty industry's advertising budget alone is several times the total global humanitarian-aid budget. That money is overwhelmingly spent persuading women that they are not quite acceptable as they are.
Two recent shifts matter. MoCRA (Modernization of Cosmetics Regulation Act, 2022) is the first significant federal cosmetics-safety law in the US in over 80 years. Facility registration, product listing, adverse-event reporting — basics that have existed in the EU since 2013 (Regulation 1223/2009). Implementation has been slow and uneven, but the framework is finally in place.
And on the marketing side, the UK Competition and Markets Authority has been quietly active. Its enforcement portfolio includes ASOS, Boohoo, Asda, and Unilever — all for misleading sustainability or wellness claims. The CMA's Green Claims Code (2021) is the clearest current statement of what regulators expect of marketing claims. None of it has stopped the industry. But it has marked the boundaries.
The "cruelty-free" label still has no agreed legal definition in the US. "Natural" and "clean" remain regulatory vacuums. The China import pathway for animal testing has changed in form but not entirely in substance. The advertising machine continues — now augmented by AI-generated imagery, which the US Federal Trade Commission's 2024 rule was the first regulator to substantively address.
The reader of the labels has to do a lot of the work the labels should be doing. That asymmetry is not accidental. It is the industry.
Approximately 75 million people make the clothes the world wears. Roughly 80% of them are women. Most of them are paid less than what their countries themselves define as a living wage. This is what we know about who they are.
The International Labour Organization's figures suggest the global garment workforce sits at roughly 75 million people. Of those, the consensus estimate from the ILO and Clean Clothes Campaign is that approximately 80% are women. It is one of the largest female workforces in the world.
It is also one of the lowest-paid. The Asia Floor Wage Alliance — the umbrella body that calculates a region-specific living wage by methodology agreed across unions and NGOs — has documented for over a decade that garment workers in major producer countries (Bangladesh, Cambodia, India, Vietnam) earn a fraction of the figure their own methodology defines as the minimum needed for basic dignity.
The industry calls it "speed-to-market." The Ellen MacArthur Foundation, in A New Textiles Economy (2017), called it something else: clothing production approximately doubled between 2000 and 2015. The average number of times a garment was worn before disposal fell sharply over the same period.
That production growth was not absorbed by new factories with better conditions. It was absorbed by squeezing more output from the existing workforce — longer hours, shorter deadlines, more aggressive subcontracting through tiers the brands themselves cannot reliably map. The 2013 collapse at Rana Plaza killed 1,134 workers. It was not an accident. It was a building that had been illegally extended, had visible structural cracks the day before the collapse, and was knowingly entered by workers who had been threatened with withholding of wages if they refused.
Sexual harassment in garment factories is not a series of bad apples. The Clean Clothes Campaign's Gendered Voices series, ActionAid's investigations in Bangladesh, and the work of Tamil Nadu-based unions have all documented harassment as a routine feature of how line supervision works — a way to enforce production quotas and discipline workers who can ill afford to lose their jobs.
And then there is the Sumangali scheme: documented in detail by SOMO and the India Committee of the Netherlands. Young women — often unmarried, often from rural Tamil Nadu — sign multi-year contracts at spinning mills in exchange for a lump sum at the end, framed as dowry. They live in mill-controlled dormitories. They are often not free to leave. The ILO and multiple human-rights bodies have characterised the arrangement as a form of bonded labour. It supplies, among other things, cotton yarn for global fashion brands.
Three legal frameworks have begun to close the gap between what brands claim and what is verifiable in their supply chains. The UK Modern Slavery Act 2015. The French Loi de Vigilance (2017). The EU Corporate Sustainability Due Diligence Directive (CSDDD), adopted in 2024. The 2021 US Uyghur Forced Labor Prevention Act presumes goods made in Xinjiang are produced with forced labour and requires importers to prove otherwise. Cotton supply chains were particularly affected.
The International Accord on Health and Safety in the Textile and Garment Industry — the legally binding building-safety agreement that emerged from Rana Plaza — has, by its own published reporting, completed inspections at thousands of factories. Where workers had no contracts they could refuse, they now have an enforcement body that can compel remediation.
None of this has made garment work fair work. But it has made it slightly less invisible. And invisibility, in the cases documented above, was the precondition for everything else.
If you buy fewer garments and wear them for longer, you reduce the demand pressure. If you ask the brands you buy from for their tier-2 supplier lists, you create pressure for transparency. If you support Clean Clothes Campaign, the Worker Rights Consortium, or the AFWA, you fund the people whose job is to track this when no one else will. None of it is sufficient. All of it is something.
In January 2023, a joint investigation by The Guardian, Die Zeit, and SourceMaterial published an analysis of Verra's rainforest carbon credit programme. It concluded that the majority of the credits it had assessed did not represent real emissions reductions. The aftermath is still unfolding.
For about a decade, the standard corporate response to scrutiny over carbon emissions was a carbon-offset purchase, a press release, and a logo on the packaging. The mechanism was simple: count your emissions, buy an equivalent number of "credits" that represented emissions reduced or avoided somewhere else, and call yourself carbon neutral.
This was always going to be vulnerable to verification. The credits represented things that did not happen — trees that were not cut down, methane that was not released. To check whether the credit was real, you had to ask: would this counterfactual have actually occurred?
In January 2023, The Guardian, Die Zeit, and the non-profit investigative outfit SourceMaterial published the results of a multi-year analysis of Verra's rainforest credit programme — the largest in the voluntary carbon market. The investigation cross-referenced Verra's claimed deforestation-avoidance against satellite imagery and reference areas. Its conclusion was that for a substantial majority of the credits assessed, the deforestation that the credits claimed to have prevented either did not actually occur in comparable reference areas, or occurred at a much lower rate than Verra's methodology assumed.
The investigation did not say all carbon credits are fraudulent. It said this specific category, accounting for billions of credits, was systematically over-issued.
Verra's CEO stepped down. The methodology is being revised. Several major corporate buyers — Shell, EasyJet, Gucci, Disney — quietly stepped back from offset-based "carbon neutral" claims. The EU's Empowering Consumers for the Green Transition Directive, adopted in 2024, prohibits generic environmental claims like "climate neutral" or "eco-friendly" unless backed by recognised certification. The UK's CMA Green Claims Code is moving in the same direction.
What didn't change: the underlying market structure that incentivises low-quality credits is still intact, and a substantial fraction of corporate net-zero commitments continue to depend on offsets of various qualities.
The Verra case is one instance of a broader pattern: voluntary corporate environmental disclosure has proliferated faster than the verification capacity to test it. CDP (the Carbon Disclosure Project) now collects climate, water, and forest disclosures from over 14,000 companies. The Science Based Targets initiative has approved targets from thousands more. ESG fund assets at peak exceeded $5 trillion — until the regulatory crackdown that began in 2022, which led to hundreds of funds being reclassified or removed.
Voluntary disclosure was always going to have a verification problem. The question now is whether mandatory disclosure under the EU CSRD (Corporate Sustainability Reporting Directive) and the various jurisdiction-specific climate-disclosure rules (SEC, ISSB, California SB 253) will materially close that gap.
The honest answer is: too early to tell. The frameworks are new, the auditing infrastructure is being built, the enforcement record is short. What we know is that the system before this was easy to game, and most of the players gamed it.
A short note for the next time you're standing in a supermarket aisle weighing a six-pack of eggs.
"Cage-free" in the US and EU contexts means hens are not kept in battery cages. It does not mean the hens go outside. It does not mean they have meaningful space per bird. In many cage-free systems the floor density is high enough that the hens still live in close confinement.
"Free-range" requires outdoor access, but the standards vary widely by jurisdiction. EU "free-range" eggs (Marketing Standards Regulation 1308/2013) require at least 4 m² of outdoor space per hen. US "free-range" has no equivalent federal standard — claims are largely self-regulated by producers.
"Organic" requires the animals receive organic feed and access to the outdoors. It does not by itself address welfare standards beyond that.
The most welfare-substantive UK certification is RSPCA Assured. The most welfare-substantive US certifications are Animal Welfare Approved (AGW) and Certified Humane. Neither comes close to addressing the welfare of male chicks in laying-hen systems, which are routinely killed shortly after hatching because they cannot lay eggs and are not the right breed for meat production. In-ovo sex determination is now commercial in Germany and France and is the largest welfare reform in the egg industry of the last decade.
None of which is to say cage-free eggs are not better than caged eggs. They are. The point is: the label is doing less work than the marketing implies, and the question the consumer is being asked to answer is harder than the label makes it appear.
The EU's 2024 Right to Repair Directive (Directive 2024/1799) is being framed as a consumer-rights win. It is. It is also, quietly, one of the most consequential environmental policies of the year.
The UN Global E-waste Monitor 2024 puts the world's annual e-waste at 62 million tonnes (2022 figure) — a stream that has been growing roughly five times faster than the documented recycling rate. Only about 22% is formally recycled. The rest is either landfilled, informally processed, or exported — frequently to Ghana, India, or Nigeria where workers strip devices for precious metals under hazardous conditions.
Most of that e-waste did not need to be e-waste. A phone with a cracked screen and a depleted battery has not failed; it has only had two replaceable components reach end of life. The same is true of laptops with swollen batteries, dishwashers with one broken sensor, and a substantial fraction of the consumer electronics graveyard.
Manufacturers have not made repair easy. iFixit's Repairability Scorecards have documented for years the structural reasons: glued assemblies, proprietary fasteners, parts-pairing software that bricks a replacement component if it doesn't match the device's serial number, and the simple non-availability of spare parts after a few years.
The EU's Right to Repair Directive (2024) requires manufacturers to make spare parts available for longer, restricts the worst parts-pairing practices, and creates a legal framework for independent repair. France's existing repairability index (in force since 2021) has produced measurable behavioural change in manufacturers — products with low scores have lost shelf space and brand goodwill.
The climate impact will take years to measure. But the upstream calculation is straightforward: every device repaired is a device not manufactured, and the manufacturing phase is where most of a consumer electronic device's embodied emissions sit.
This is the kind of policy that the environmental movement has been arguing for for two decades. That it has finally landed — quietly, under the radar of most consumer media — is a real victory.
The phrase that should determine whether an influencer post is an advertisement. Three words. Most posts fail them.
The US FTC's Endorsement Guides — the foundational document for influencer advertising disclosure — require that any material connection between an endorser and a brand be disclosed "clearly and conspicuously." The UK's Advertising Standards Authority requires the same under CAP Code rule 3.1 and the Communications Act.
What does "clear and conspicuous" mean in practice? The ASA's 2021 monitoring sweep and its subsequent rulings give a working definition. "#ad" at the start of a caption is acceptable. "#ad" buried at the end, after thirty other hashtags, is not. A Stories sticker that only appears in one of a six-Story sequence is not. A vague "#partner" without indication of what the partnership is, is not. An entire post about a product you are paid to promote, with no disclosure at all, is — obviously — not.
The ASA's 2021 sweep found widespread non-compliance across UK influencer accounts. The US FTC has brought multiple enforcement actions. Norway extended its retouching-disclosure law to influencer content in 2022. The EU's Digital Services Act (2024) creates a legal framework for very large online platforms to enforce disclosure standards on the creators they host.
None of this is academic. Influencer marketing is one of the highest-growth segments of global advertising spend. Consumers — particularly younger ones — increasingly source product decisions from accounts they trust. When those accounts are paid promotion that has not been disclosed, the consumer is being deceived about the nature of what they are reading.
If you follow an account whose product mentions are not labelled, ask yourself when you last saw a transparent disclosure. The answer is usually informative.
The social audit industry exists to provide credible-looking documentation that supply chains are clean. Repeatedly, it has documented things that were not the case.
In April 2013, the Rana Plaza building in Savar, Bangladesh, collapsed. Inside the building had been factories producing for major Western brands. Several of those factories had been recently audited by major social-audit firms. The audits had not flagged the structural issues.
This is not an isolated incident. MSI Integrity's 2020 report Not Fit-for-Purpose documented that the major multi-stakeholder initiatives — Fair Labor Association, Ethical Trading Initiative, Roundtable on Sustainable Palm Oil, and others — have systematically failed to deliver on their stated mandates. Clean Clothes Campaign's audit-of-the-auditors work has shown the same pattern across the commercial audit industry.
The structural problem is straightforward. Social audits are typically:
That last point is the structural flaw. The auditor's client is the brand, not the worker. When the auditor finds problems, the cost — both financial and reputational — falls on the brand that commissioned them. When the auditor finds nothing, the relationship continues.
This is why a building like Rana Plaza could be audited and certified. Why Boohoo's Leicester supply chain could be cleared by audits before the Sunday Times documented sub-minimum-wage payments. Why repeated audits of palm-oil plantations have failed to catch the deforestation, the child labour, and the worker abuses that NGOs operating in the same regions document routinely.
The alternative is not "no audits." It is independent investigation, worker-driven monitoring (the model the Worker-driven Social Responsibility Network has been building since 2011), and legally binding agreements like the International Accord. These work because the worker, not the brand, is the principal. When the auditor's client is the worker, the audit catches what audits were designed to catch.
A few numbers that don't change much from one year to the next, and shouldn't.
Approximately 80 billion land animals are slaughtered for food each year. The figure is calculated from UN FAO production statistics and varies modestly by year and accounting method, but the order of magnitude is stable. About 78 billion of them are chickens.
The figure for fish is harder. The Fishcount project, drawing on FAO capture data and Mood & Brooke's average-weight estimates, places annual wild-fish capture at between 0.79 and 2.3 trillion individuals. The wide range reflects the uncertainty of estimating from weight-based industry data. Farmed fish add another ~100 billion. The honest answer is: we do not know, to within an order of magnitude, how many fish humans kill in a year.
The UN FAO attributes approximately 14.5% of global anthropogenic greenhouse-gas emissions to livestock specifically. The full food system, per Crippa et al. (Nature Food, 2021), is roughly a third of total emissions. Poore & Nemecek's 2018 meta-analysis remains the most comprehensive comparison of food-system environmental impacts by category — beef and dairy are the highest per calorie by significant margins.
The Sentience Institute estimates that more than 90% of farmed land animals globally live in intensive systems (CAFOs, broiler sheds, concentrated dairy). Compassion in World Farming's facility documentation gives the texture: typical broiler-chicken stocking densities, lameness rates, mortality figures. The numbers are matter-of-fact and the lives they describe are not.
The human side is documented too. The US Bureau of Labor Statistics records meat-and-poultry processing injury rates at multiples of the average manufacturing rate. Human Rights Watch's Blood, Sweat, and Fear (2004) and its 2019 follow-up document the systemic nature of those injuries — line speeds engineered for output, not for the bodies producing it, falling overwhelmingly on immigrant and refugee workers.
The 2023 NYT investigation by Hannah Dreier — Pulitzer-winning — found unaccompanied migrant children working in US slaughterhouses through subcontracted cleaning agencies. That investigation produced enforcement actions. The underlying conditions it documented have not yet been substantially changed.
None of these figures are new. The reporting bodies — FAO, ILO, USDA BLS, Human Rights Watch, NYT, the welfare NGOs — have been documenting them for years or decades. The reason they do not move public sentiment more than they do is not lack of evidence. It is that the evidence is consistent, and the act of attending to it is, for most readers, a choice they make once and then have to keep making.
For two decades, corporate responsibility was a marketing function. With the EU's CSDDD landing in 2024 and the French Loi de Vigilance already producing court cases, that era is closing — slowly, partially, but unmistakably.
The voluntary corporate social responsibility era — broadly, the period from the late 1990s to the late 2010s — produced an enormous infrastructure of codes, multi-stakeholder initiatives, third-party certifications, and annual reports. It produced fewer measurable improvements in the supply chains it claimed to police. MSI Integrity's Not Fit-for-Purpose (2020) is the most comprehensive published audit of that infrastructure's failure.
The pattern was structural. A brand commissioned a third-party social audit. The auditor depended on the brand for continued contracts. The audit confirmed what the brand needed it to confirm. When investigators — journalists, NGOs, worker-driven monitoring bodies — looked at the same suppliers, they found different things. Rana Plaza, Boohoo Leicester, Shein Guangzhou, the Tamil Nadu spinning mills, the Brazilian cattle ranches: all of them had been audited. None of those audits surfaced what the investigations did.
Three legal regimes now operate in parallel and increasingly interlock.
The UK Modern Slavery Act 2015. Large companies must publish annual statements describing steps taken to address forced labour in their supply chains. Compliance has been demonstrably poor — the UK Independent Anti-Slavery Commissioner has documented routine boilerplate statements that say nothing specific. But the Act produced two important precedents: corporate transparency is now legally required (not just expected), and large investors increasingly read these statements as substantive risk indicators.
The French Loi de Vigilance (Duty of Vigilance Law, 2017). Goes further than the UK Act: requires companies above a size threshold to publish a "vigilance plan" identifying risks across operations and supply chains, with civil liability for failure. Has already produced court cases — notably against TotalEnergies, EDF, and others. The Notre Affaire à Tous and Sherpa-led litigation has shown that the law can be enforced.
The EU Corporate Sustainability Due Diligence Directive (CSDDD, 2024). The most ambitious of the three. Mandates risk-based due diligence across human rights and environmental impacts for large EU and EU-trading companies. Was watered down significantly during negotiations — the original Commission proposal covered a wider scope of companies — but as adopted it still represents the most consequential change in mandatory corporate accountability since the 1970s.
Several major gaps remain.
We are at the end of the voluntary era and the beginning of the mandatory one. The mandatory framework is imperfect, narrower than its advocates wanted, and dependent on enforcement infrastructure that has not been built. It is still — incontestably — a substantial improvement on what came before. The companies that genuinely improved in the voluntary era will find compliance straightforward. The companies that did not will face documented, legally-actionable obligations they have not previously had to meet.
That is not a complete solution. It is the beginning of one.
"Eco-friendly." "Carbon neutral." "Sustainably sourced." For a decade these phrases did a lot of marketing work and almost no regulatory work. In 2024 that changed.
Until recently, environmental marketing claims sat in a regulatory grey zone. The US FTC's Green Guides (2012) and the UK CMA's Green Claims Code (2021) existed as guidance — applied to specific complaints but rarely enforced at scale. The result: a marketplace where "eco-friendly," "sustainably sourced," and "carbon neutral" appeared on packaging with minimal substantiation requirement.
Three things changed in 2023–2024.
In January 2023, the joint investigation by The Guardian, Die Zeit, and SourceMaterial published an analysis of Verra's rainforest carbon credit programme — the largest in the voluntary carbon market. The conclusion: a substantial majority of the credits assessed did not represent the emissions reductions they claimed to. The fallout reshaped the offset industry: Verra's CEO stepped down; major buyers including Shell, EasyJet, Gucci, and Disney quietly withdrew from offset-based "carbon neutral" claims; the voluntary carbon market's trading volume dropped sharply through 2023 and 2024.
The investigation did not say all offsets are fraudulent. It said this specific category, accounting for billions of credits, was systematically over-issued. The damage to the broader category was reputational rather than analytical, but it was severe.
Adopted in 2024 (Directive 2024/825), the directive explicitly prohibits generic environmental claims such as "climate neutral," "carbon neutral," "eco-friendly," "natural," and "biodegradable" unless backed by specific substantiation under recognised certification. It also restricts claims based on offsetting unless the climate impact is genuinely real and measurable.
This is the most significant piece of consumer-protection legislation on environmental marketing in any major jurisdiction. Companies operating in the EU now face actionable consequences for unsubstantiated green claims — fines based on a percentage of annual turnover. The Directive is being transposed into national law across the EU through 2026.
The 2022 ASOS, Boohoo, and Asda investigation forced commitments to substantiate sustainability claims. The 2023 Unilever investigation extended to FMCG and continues. The CMA's enforcement track record is now a published portfolio that other regulators are citing. The legal authority — the Consumer Protection from Unfair Trading Regulations 2008 — has been available for over a decade. It is only now being used systematically.
For consumers: a green claim on packaging is now more likely to mean something. For brands: a green claim now needs a documented evidence base behind it. For the underlying environmental problem: not very much, in the short term. Greenwashing was always a symptom, not a cause. Reducing emissions, water pollution, microplastic release, and habitat destruction requires actual operational change, not better marketing language.
The optimistic reading: regulators are removing a tool that allowed companies to project environmental performance without delivering it. The realistic reading: companies still set their own operational policies, and the regulatory pressure is on what they say, not what they do. Both readings are true.
The honest reading: when "carbon neutral" stops being a useful marketing phrase, the companies that built their identity around it will either start doing the underlying work or stop saying the words. Most will do the latter. A meaningful minority will do the former. The job of the people watching is to know which is which.
An honest account of why the same digital infrastructure that ran the 2016 election manipulation now runs your skincare ad targeting — and why that's the same problem, not two different ones.
The Cambridge Analytica case in 2018 — 87 million Facebook profiles harvested without consent, used to build psychometric models for political advertising — produced widespread public alarm. What it did not produce was widespread public understanding that the same infrastructure underwrites everyday consumer advertising.
Cambridge Analytica's methodology was distinctive in one respect: the data was acquired in violation of Facebook's terms (such as they were). The underlying technique — inferring personality traits from social-media behavioural patterns to optimise advertising delivery — is, by contrast, the default state of digital advertising. It runs on every smartphone, in every app, served by hundreds of data brokers most people have never heard of.
The Norwegian Consumer Council's 2020 Out of Control investigation traced the data sharing in ten common apps. A single user opening a period-tracking app or a dating app or a fitness app would, within seconds, have aspects of their profile shared with dozens or hundreds of third-party advertising and data-broker firms. The mechanism is "real-time bidding" — when an ad slot becomes available, a bid auction runs in milliseconds, and the auction packet includes information about the user.
The information varies. Location. Device. App use history. Demographic inferences. For sensitive apps, the category of sensitive data the app handles (mental health, sexuality, pregnancy, fertility, chronic illness). The fact that a user has opened the app at all.
The IAB Europe's Transparency & Consent Framework — the consent infrastructure underlying real-time bidding in the EU — was ruled non-compliant with GDPR by the Belgian Data Protection Authority in February 2022. The framework continues to operate in modified form. Reform is ongoing.
Beauty and personal-care advertising is one of the most data-rich consumer categories. The product cycles are short, the consumer base is highly segmentable, the marketing language is emotionally targeted. The combination produces a specific kind of advertising experience: a user searches once for an acne treatment, and their digital environment fills with anti-acne advertising for months. A user mentions a wedding, and bridal-industry ad targeting follows. A user's location data suggests proximity to a fertility clinic, and fertility-industry ads escalate.
The downstream effects on body image, on financial decisions, on emotional well-being have been documented across multiple academic studies. The Dove Self-Esteem Project's research has documented links between advertising exposure and body-image outcomes in adolescents. The Reuters Institute's work has documented adverse mental-health correlations with intensive social-media advertising exposure.
The EU Digital Services Act (2022/2065), in full application from February 2024, prohibits targeted advertising to minors and based on sensitive data (sexual orientation, religion, political views, health). It requires transparency on the criteria used for ad targeting.
Apple's App Tracking Transparency framework (iOS 14.5, April 2021) requires explicit user consent for cross-app tracking. Adoption rates of opt-in have been measured at well below 50% — most users decline tracking when asked. Facebook (Meta) attributed approximately $10 billion in 2022 ad-revenue loss to the change.
GDPR enforcement in the EU has produced billion-euro fines for major platforms over consent and transparency failures, though enforcement has been slow and remains uneven.
US state privacy laws — California (CPRA), Colorado, Virginia, and a growing list of others — provide opt-out rights for sale and sharing of personal information.
The underlying surveillance-advertising business model is intact. The major platforms make most of their revenue from it. The infrastructure of data brokers — companies whose entire business is the commercialisation of personal data — operates in most jurisdictions with minimal oversight. The 2024 FTC settlements with X-Mode, Avast, and Kochava show enforcement is increasing, but the addressable universe of similar firms is large.
The honest assessment: the regulatory architecture is improving faster than the underlying practice, but slower than the technology is evolving. The infrastructure for personalised advertising is now being adapted to generative AI — synthetic personalised content, deepfake endorsements, AI-generated influencer personas. The FTC's 2024 rule on AI-generated endorsements is the first regulatory step in that direction. There will need to be many more.
A note on the most common claim on a skincare label.
"Clinically proven" or "clinically tested" appears on a substantial fraction of skincare products sold in the US, UK, and EU. The phrase has no legal definition in any of those jurisdictions. It does not mean what "clinical trial" means in a pharmaceutical context. It does not require independent verification.
In practice, the underlying evidence base for a "clinically proven" cosmetic claim can range from a small open-label study commissioned by the manufacturer (n=20, 4 weeks, no control group) to substantive, peer-reviewed efficacy research. Both legitimately produce the same words on the packaging.
What the cosmetics regulator can require is that the claim be substantiated — that is, that the manufacturer be able to produce some evidence to support the claim if asked. The UK ASA, US FTC, and EU competent authorities have brought enforcement actions where the substantiation file was inadequate. But the words themselves remain available to anyone.
The same applies to "dermatologically tested" (does not require dermatologist-led testing; can simply mean a dermatologist reviewed the study), "hypoallergenic" (no agreed standard), and "non-comedogenic" (specific industry standards exist but are not regulated).
The two best independent sources for evidence on skincare ingredient efficacy are: EWG's Skin Deep database (which rates ingredient safety against published toxicology) and peer-reviewed dermatology literature (PubMed). Both will tell you, on most ingredients, what is and is not substantiated. Neither will be on the packaging.
A short, slightly bleak fact about washing machines.
The IUCN's 2017 report Primary Microplastics in the Oceans estimated that approximately 35% of primary microplastics entering the marine environment globally come from synthetic-fibre textiles, principally shed during washing. That's the largest single source — larger than tyre wear (28%), city dust (24%), or any of the other contributing sources.
Synthetic fibres — polyester, nylon, acrylic — shed microscopic fragments during every wash cycle. A typical 6 kg load of synthetic clothing releases hundreds of thousands to millions of fibres per wash, depending on garment type, fabric construction, and wash conditions.
The mitigations are partial. Higher-quality synthetic fabrics shed less than low-quality ones. Cooler washes at lower agitation shed less than hot/intensive washes. Front-loading machines shed less than top-loaders. Microfibre-capture filters (Guppyfriend bag; Cora Ball; in-line washing-machine filters) capture a meaningful fraction but not all.
The structural fix is upstream: less synthetic-fibre clothing in the first place, and longer use of each garment to amortise the production-and-shedding cycle over more wears. France's 2025 microfibre-filter requirement on new washing machines is the first major regulatory move on this. The EU is considering parallel measures under the Ecodesign for Sustainable Products Regulation.
The simple version: every time you wash a synthetic garment, some of it goes to the ocean. The complicated version is what to do about it. The honest version is: this is one of the issues that does not solve itself by buying a different product. It solves itself partly by buying less, washing less, and wearing things longer.
A short note on fish, and why we don't think about them the way we think about chickens.
The Fishcount project, drawing on FAO capture data, places annual wild-fish capture at between 0.79 and 2.3 trillion individuals. Farmed fish add roughly another 100 billion. Compared to the ~80 billion land animals slaughtered annually, the order-of-magnitude difference is striking — and the public attention to it is correspondingly inverted.
The reason is partly historical (fish do not feature prominently in the welfare frameworks that built up around mammals and birds) and partly empirical (fish welfare science is younger than mammalian welfare science, and the policy infrastructure has not caught up).
What the welfare science currently says: fish are sentient. The UK formally recognised this in the Animal Welfare (Sentience) Act 2022, which covers cephalopod molluscs and decapod crustaceans (octopuses, lobsters, crabs) as well as vertebrates. The EU Treaty on the Functioning of the European Union, Article 13, has recognised animal sentience since 2009. Peer-reviewed work — Sneddon, Braithwaite, Brown, and others — has documented nociception, learning, social behaviour, and stress responses in fish.
What welfare looks like in practice: high-density aquaculture systems produce stress responses, disease prevalence, and behavioural restriction comparable in scale (though different in character) to intensive land-animal agriculture. Slaughter practices for fish are usually substantially less developed than for mammals — many wild-caught fish die through asphyxiation on deck rather than through any controlled stunning method.
The infrastructure for change is starting to exist. The Aquaculture Stewardship Council and the RSPCA Assured aquaculture standards are imperfect but better than no standards. Norway's 2021 amendments to its aquaculture welfare law are among the most substantive. The Fish Welfare Initiative — a relatively new NGO — is producing welfare research focused on the species (Indian carp, Vietnamese pangasius) where the population numbers are largest and the welfare attention has historically been least.
The honest assessment: the welfare conversation we have had about chickens, cows, and pigs needs to happen about fish too. The number of animals involved makes the scale much larger. The scientific and policy infrastructure to support that conversation is still being built.
A short note on what board-quota progress actually measures.
Among the genuinely useful things to come out of corporate-governance reform in the last decade is the rise of female board representation in major markets. Norway's 40% quota law (2003) was the early example. The EU's 2022 Women on Boards Directive set a 40% target for non-executive directors by 2026. California's now-overturned 2018 SB 826 led to substantial increases in board diversity before being ruled unconstitutional in 2022. The UK's Davies and Hampton-Alexander reviews drove voluntary progress to 40% on FTSE 350 boards by 2024.
The progress is real and worth recognising. It is also, on its own, a very limited measure of gender equality in corporate life.
What it does not measure: the gender pay gap below board level (UK Office for National Statistics: approximately 7-14% depending on measurement); the gender breakdown of executive committee membership (much slower to shift than board membership); the gender breakdown of people in pipelines to executive committees; harassment and conditions in the workforce as a whole; the gender composition of the supply chain (overwhelmingly female in beauty, fashion, food processing); and the lives affected by the products the board oversees.
Board-quota progress is a useful proxy because it is countable. McKinsey's annual Women in the Workplace report and LeanIn.Org's parallel work have built the data infrastructure to track it. The same reports also document that progress at the top has not produced commensurate progress at the middle or bottom of the corporate pipeline.
The honest framing: a 30% female board is not gender equality. It is a useful tracking metric for one specific kind of progress. The other kinds of progress — equal pay, equal access to executive pipelines, harassment-free workplaces, equitable supply chains — require their own metrics and their own enforcement, and have generally not been making the same kind of progress.
A short field guide for what to look for when you're about to sign up to something.
A "subscription trap" is the consumer-protection term for a service that is engineered to be easy to start and hard to leave. The US Federal Trade Commission's 2024 "Click to Cancel" rule was specifically targeted at the structural patterns documented in the field for years. Here is what those patterns look like in practice.
The 2024 FTC rule addresses several of these — particularly the "cancel through the same channel as you signed up" requirement. The UK Consumer Rights Act 2015 contains overlapping protections; the EU Digital Services Act and Unfair Commercial Practices Directive cover others.
What the consumer can do, regardless: assume any service that does not let you see and cancel the subscription on a single screen is intentional. Set a calendar reminder for trial periods and annual renewals. Read the cancellation policy before signup, not at the moment of trying to leave.
A short note on where the wild-animal-captivity debate currently sits.
The cultural conversation about wild animals in entertainment has progressed in ways the cultural conversation does not always notice. Blackfish (the 2013 documentary about SeaWorld's orca programme) marks a useful watershed: in 2016 SeaWorld ended its orca breeding programme; in 2021 the United States passed the SAFE Act extending federal protections; multiple countries have banned or severely restricted dolphinariums.
The current landscape:
The structural insight is that captivity discussions have moved from a moral binary toward a regulatory framework. The Five Domains welfare assessment (Mellor et al.) is increasingly used as the scientific basis for evaluating captive contexts. The Animal Welfare (Sentience) Act 2022 in the UK and the EU Lisbon Treaty Article 13 (2009) provide a legal basis for ongoing reform.
The honest reading: the cultural energy that pushed back against wild-animal entertainment in the 2010s produced real policy results. The category of "circus with wild animals" is now legally precarious across most of the developed world. The category of "captive cetaceans" is shrinking. The category of "zoos" is being reformed rather than abolished. Wildlife tourism is the next front, and the work is still ongoing.
A short note bringing together the available evidence on advertising exposure and body image — particularly for younger women and girls.
The evidence base linking exposure to beauty advertising and social-media imagery with documented mental-health outcomes — particularly in adolescent girls — has grown considerably since 2018. The relevant findings, briefly.
The Facebook (Meta) internal research. In September 2021, The Wall Street Journal's "Facebook Files" series published internal Meta research documents acquired from whistleblower Frances Haugen. The documents included Instagram's own commissioned research finding that for a meaningful proportion of teenage girls using the platform, Instagram exacerbated body-image concerns. The findings — Meta's own — appeared not to have driven product changes prior to public exposure.
The Dove Self-Esteem Project longitudinal research. Commissioned by Unilever, methodologically published, peer-discussed. Documents correlations between exposure to retouched/idealised imagery and measurable body-image, self-esteem, and disordered-eating outcomes. The same research has been used by legislators in France, Norway, and Israel as evidence-base for retouching-disclosure laws.
Peer-reviewed dermatology and psychology literature. Studies in Body Image, the Journal of Adolescent Health, and equivalent journals consistently find moderate effect sizes for the relationship between social-media use intensity and body-image concerns, with stronger effects for visual platforms (Instagram, TikTok) than text-based ones.
The regulatory responses have been incremental. The UK Online Safety Act 2023 places duties on Very Large Online Platforms regarding content harmful to children, including content promoting eating disorders. The EU Digital Services Act prohibits targeted advertising to minors. France's retouching-disclosure law has now extended to influencer content via Norway's 2022 amendment to its Marketing Control Act.
The honest assessment: the link between advertising exposure and mental-health outcomes is documented well enough that it is no longer empirically contested. The translation from documentation to regulatory action has been slow. The advertisers themselves — particularly in beauty — are starting to commission their own research on the question (Dove being the most public example), partly out of genuine concern and partly because anticipating future regulation is a commercial strategy.
What the consumer can do: notice. The first defence against advertising that exploits insecurity is the recognition that the insecurity was, often, manufactured by the same industry now selling its cure.
Every membership directly funds independent research and keeps our content free for those who cannot pay.
For those beginning to engage with the work.
For committed researchers and organisers.
For leaders needing full access and direct support.
No registration required · Pay via PayPal · Money goes direct to you · Cancel any time by emailing info@feral-femme.co
A one-time or recurring donation funds the same independent investigative work — with no strings attached.
“I didn’t start Feral Femme because I wanted to build a brand. I started it because I realised how easy it is to unknowingly participate in harm when systems are designed to hide it.”
I didn’t start FERAL FEMME because I wanted to build a brand. I started it because I realised how easy it is to unknowingly participate in harm when systems are designed to hide it.
Like most people, I was taught to trust labels, regulations, and distance. I was taught that cruelty was something obvious — something extreme, visible, and rare. But the deeper I looked into the beauty industry, the clearer it became that harm doesn’t survive through violence alone. It survives through language, convenience, and silence.
FERAL FEMME was created to sit in that uncomfortable space between ignorance and accountability.I didn’t want to create a platform driven by outrage, guilt, or moral performance. I wanted something quieter, more precise — a space where information is allowed to speak for itself, and where empathy isn’t softened to make it easier to ignore.
This platform is intentionally restrained. The discomfort is deliberate. The education is the point.
FERAL FEMME is not here to tell people what kind of person they are. It is here to show how systems function — and to trust that once people truly understand, they can no longer unknowingly support harm.
Two addresses cover everything — general enquiries and direct contact with the founder.
Tips, corrections, press, membership, partnerships — everything starts at one of these two inboxes.
For source tips and factual corrections, press & speaking requests, membership and account questions, partnerships, syndication, and anything else.
For founder-direct correspondence, editorial collaboration, and matters that warrant a personal reply rather than the general inbox.
How Feral Femme operates, your rights as a reader and member, and the South African legislation we follow.
This page contains all legal information governing your use of feral-femme.co (the “Site”), operated solely by Ayden Rosemary Brown (“Feral Femme”, “we”, “us”, “our”) in South Africa.
By accessing or using this Site, you agree to be bound by all sections below. If you do not agree with any part, please do not use the Site.
Operator details (ECT Act 25 of 2002, Section 43):
Unlock custom research requests, priority email response, early access to investigations, and personalised brand audits for $49/month.
Each industry below opens into a sourced summary: who pays the price, the figures worth knowing, the documented research, and how it connects to the rest of the archive. Click any heading to expand.
The global beauty industry is one of the largest consumer markets in the world — McKinsey valued it at roughly $446 billion in 2023 and projects it growing past half a trillion dollars by the end of the decade. Most of that revenue is spent by women, on products marketed to address problems the same industry teaches them to see.
Two structural problems run through the sector. First, animal testing on cosmetics is banned for finished products and ingredients in the EU since 2013 (Cosmetics Regulation 1223/2009), and a growing list of countries have followed. But "cruelty-free" claims on labels are largely unregulated in the United States, and ingredient supply chains routinely extend into markets — historically including mainland China for imported cosmetics — where testing has been required or assumed by default. The 2021 reforms in China created exemptions but did not eliminate testing pathways, and implementation has been uneven.
Second, the industry sells daily-use products containing substances of regulatory concern. The Environmental Working Group's Skin Deep database catalogues thousands of cosmetic ingredients against published toxicology and endocrine-disruption data; the European Commission's Cosing database lists substances banned or restricted under Annex II/III. The US passed the Modernization of Cosmetics Regulation Act (MoCRA) in 2022 — the first significant federal cosmetics-safety legislation in over 80 years — and full implementation continues.
The third layer is psychological. The industry's marketing is built on the production of dissatisfaction: showing a face the consumer is then sold the tools to "correct." Dove's own commissioned research (The Real Truth About Beauty, 2004 onward) documented that only a small minority of women describe themselves as beautiful. The product was the cause, then sold as the cure.
The garment industry employs roughly 75 million workers globally, of whom approximately 80% are women, according to the International Labour Organization. It is, in scale, one of the most female-dominated industries in the world — and one of the lowest-paid.
Fast fashion compressed the production cycle from four seasons a year to weekly drops. The Ellen MacArthur Foundation's A New Textiles Economy report documented that clothing production approximately doubled between 2000 and 2015, while the average number of times a garment is worn before disposal declined sharply. A garment now passes through global supply chains, retail, and landfill within months.
The 2013 Rana Plaza collapse in Bangladesh — 1,134 garment workers killed, the majority women — produced the International Accord on Fire and Building Safety in Bangladesh, the first legally binding building-safety agreement between brands and unions. The Accord has, by its own reporting, completed inspections at thousands of factories and identified more than 100,000 hazards. Wage theft, forced overtime, and union suppression continue to be documented by Clean Clothes Campaign, Worker Rights Consortium, and IndustriALL.
Animal-derived materials carry their own architecture of harm. Leather is the byproduct of a cattle industry that slaughters roughly 300 million cows annually (FAO). Wool, particularly Australian Merino, involves mulesing — banned in New Zealand since 2018, ongoing in Australia. Fur: a number of countries (Netherlands, UK, Austria, Israel) have banned fur farming; production continues at scale in Finland, Poland, and China. Silk production conventionally requires boiling silkworms inside their cocoons.
Approximately 80 billion land animals are killed for food each year — UN FAO production figures, summed across cattle, pigs, sheep, goats, chickens, ducks, turkeys and the rest. The figure for fish is much harder to estimate; the Fishcount project (Mood & Brooke) puts the global capture of finfish at between 0.79 and 2.3 trillion individuals per year, plus farmed fish. Most of these deaths are uncounted at the individual level because fish are weighed, not enumerated.
Factory farming — concentrated animal feeding operations (CAFOs), industrial poultry sheds, intensive dairy — accounts for the great majority of animals raised for food in the United States and the European Union. The Sentience Institute and Compassion in World Farming have produced concordant estimates that more than 90% of farmed animals globally live their lives inside intensive systems.
The labour conditions inside slaughterhouses are documented and severe. The US Bureau of Labor Statistics records injury rates in meat and poultry processing at multiples of the average manufacturing rate. Human Rights Watch documented, in Blood, Sweat, and Fear (2004) and subsequent reports, line speeds that systematically produce repetitive-strain injuries, lacerations, and amputations — predominantly affecting immigrant and refugee workers.
Three further dimensions matter. Antimicrobial resistance: the WHO, FAO, and OIE jointly classify livestock antibiotic use as a major driver of human AMR — roughly two-thirds of global antibiotic use is in animal agriculture. Climate: the FAO's Tackling Climate Change Through Livestock attributes approximately 14.5% of global anthropogenic greenhouse-gas emissions to livestock. Zoonosis risk: the UN Environment Programme's Preventing the Next Pandemic report names intensive livestock production as a key zoonotic-disease driver.
Animal welfare is not a sector — it is a thread running through every other one. The same animal is a unit of inventory in agriculture, a test subject in cosmetics and pharmaceuticals, a material input in fashion, a research tool in technology, an entertainment good in tourism, and a status object in retail.
The IUCN Red List of Threatened Species currently assesses more than 150,000 species; over 42,000 are listed as threatened with extinction (latest IUCN reporting). CITES — the Convention on International Trade in Endangered Species — regulates international wildlife trade across 184 parties, but enforcement and the scale of the illegal wildlife trade remain enormous: the UN Office on Drugs and Crime documents wildlife trafficking as a major transnational organised-crime sector.
Captivity industries — zoos, aquariums, dolphinariums, circuses, exotic-pet keeping — sit on contested ground. Several European countries (Greece, Austria, parts of the UK) have banned or severely restricted wild animals in circuses. The 2013 documentary Blackfish drove sustained scrutiny of SeaWorld's orca programme; SeaWorld ended its orca breeding programme in 2016.
The exotic-pet trade — reptiles, primates, big cats, parrots — is overwhelmingly driven by demand in higher-income countries and serviced by capture from the wild. World Animal Protection's investigations have repeatedly documented the welfare, conservation, and zoonotic-disease consequences. The COVID-19 pandemic intensified global attention on wildlife markets but did not, in most jurisdictions, structurally change them.
Three industries Feral Femme covers most closely — beauty, fast fashion, and consumer media — are female-dominated on both sides of the supply chain. They are marketed primarily to women, and produced primarily by women, with the value extracted by ownership and shareholders that are not.
In garment manufacturing, the ILO documents that roughly 80% of garment workers globally are women. The Asia Floor Wage Alliance and Clean Clothes Campaign have repeatedly documented gendered wage gaps within already-low garment wages, sexual harassment as a structural feature of supervision in many factories, and the targeting of pregnant workers for dismissal.
In beauty, marketing research from Dove and others has documented the manufactured-insecurity model. The Dove Self-Esteem Project's longitudinal research shows that body dissatisfaction is associated with measurable mental-health outcomes and that exposure to retouched media increases the effect. The UK Advertising Standards Authority has, since 2019, prohibited gender-stereotyped advertising likely to cause harm or serious offence under CAP and BCAP Code rule 4.9 — a notable regulatory acknowledgement.
The thread connects to corporate accountability: the Walk Free Foundation Global Slavery Index estimates around 50 million people are living in modern slavery worldwide, with women and girls disproportionately represented in forced marriage and commercial sexual exploitation. The fashion and beauty industries' supply chains intersect with these figures in ways their own marketing rarely acknowledges.
Three industries Feral Femme tracks closely — fashion, food, and electronics — together account for a large share of global environmental damage. The numbers are striking and the connecting thread is consumption pattern: rapid production cycles, designed obsolescence, and discarded volume.
Fashion's environmental footprint includes textile wastewater (the World Bank has attributed an estimated 17–20% of global industrial water pollution to textile dyeing and treatment), microplastic shedding from synthetic textiles (Ellen MacArthur Foundation and IUCN reports), and an emissions share variously estimated at 4–10% of global anthropogenic GHGs. The wide range reflects methodological disagreement — Quantis, McKinsey, and UNEP have all published figures, with the highest cited numbers often exceeding the most defensible peer-reviewed estimates.
Agricultural emissions — the UN FAO attributes roughly 14.5% of global GHGs to livestock specifically, and the broader food system accounts for around a third of global emissions according to Crippa et al. (Nature Food, 2021). Beef and dairy lead per-calorie emissions intensity; Poore & Nemecek (Science, 2018) remains the most comprehensive meta-analysis comparing food categories.
Electronics: the UN's Global E-waste Monitor 2024 documented 62 million tonnes of e-waste generated in 2022, with a global recycling rate of approximately 22%. The fastest-growing waste stream in the world.
Consumer technology has four overlapping ethical problems: the labour and environmental cost of raw material extraction, the e-waste generated by short replacement cycles, the data-extraction economy underwriting "free" services, and the repair barriers that lock consumers into the disposal cycle.
Cobalt — essential for lithium-ion batteries — is mined principally in the Democratic Republic of Congo. Amnesty International's 2016 report This Is What We Die For documented child labour and severe injury rates in artisanal cobalt mining, and traced the supply chain to major consumer-electronics brands. Subsequent reporting by the Washington Post and RAID (Rights and Accountability in Development) has continued to document conditions.
The UN Global E-waste Monitor 2024 reports 62 million tonnes generated in 2022. Only about 22% is documented as formally recycled; much of the remainder is exported — often informally — to countries including Ghana, India, and Nigeria, where workers extract precious metals under hazardous conditions. Agbogbloshie in Accra, Ghana has become an emblematic site, documented by Pure Earth/Blacksmith Institute and BAN (Basel Action Network).
The right-to-repair movement and the EU's 2024 Right to Repair Directive address the structural side: manufacturers' use of glued assemblies, software-tied parts pairing, and absent spare-parts availability that force replacement over repair. iFixit, Right to Repair Europe, and PIRG (US) have driven most of the policy progress.
On the data side, the Norwegian Consumer Council's Out of Control (2020) investigation documented how a single user's data is routinely shared with hundreds of third parties via real-time bidding ad infrastructure. EU GDPR enforcement, US state-level privacy laws (California's CPRA, Colorado, Virginia), and ongoing FTC actions continue to shape this terrain.
Consumer-rights work sits at the intersection of every other industry on this site. A garment-worker abuse, an animal-welfare violation, a chemical-safety concern, an environmental claim — each is also a consumer-information problem the moment a product enters a store and asks for a purchase. The question is what consumers can be expected to know, and what regulators can be expected to enforce.
Greenwashing — environmental claims unsupported by evidence — is now actively enforced in multiple jurisdictions. The EU's Green Claims Directive (2024 framework) and Empowering Consumers for the Green Transition Directive require substantiation of environmental claims. The UK CMA has issued the Green Claims Code and brought enforcement actions against ASOS, Boohoo, Asda, Unilever, and others. The US FTC's Green Guides are under active revision; the FTC brought enforcement actions including the Kohl's and Walmart "bamboo" textile case ($5.5M combined penalties, 2022).
Dark patterns — UI design that manipulates users into choices against their interests — are now formally targeted by the EU Digital Services Act, the EU Consumer Protection Cooperation network, and the US FTC. The Norwegian Consumer Council's Deceived by Design (2018) is the foundational documentation. The European Commission and US FTC have each subsequently brought guidance and enforcement frameworks.
Cruelty-free, natural, clean, eco-friendly — these terms have no agreed legal definition in most consumer markets. The Norwegian Consumer Council and BEUC (European Consumer Organisation) have repeatedly documented the regulatory vacuum that allows them to proliferate without consequence.
The structural problem of corporate accountability is the gap between a published ESG report — voluntary, brand-controlled, designed to be read by investors — and the verifiable conditions in the global supply chain. Voluntary corporate disclosures have proliferated over the last two decades; independent verification has not kept pace.
Three legal frameworks have begun to close the gap. The UK Modern Slavery Act 2015 requires large companies trading in the UK to publish annual statements describing the steps they have taken to address forced labour in their supply chains; subsequent reviews (the Independent Anti-Slavery Commissioner, Walk Free) have repeatedly documented poor compliance quality. France's Loi de Vigilance (2017) goes further, imposing a duty of vigilance with potential civil liability. The EU Corporate Sustainability Due Diligence Directive (CSDDD) — adopted in 2024 — establishes mandatory due-diligence obligations for large EU and EU-trading companies on both human-rights and environmental impacts.
The Business & Human Rights Resource Centre maintains the most comprehensive open database of alleged corporate abuses and company responses. Know The Chain benchmarks major companies in apparel, food, and ICT on forced-labour risk. The FAIRR Initiative does the same for animal-protein producers. CDP (formerly Carbon Disclosure Project) holds the largest dataset on corporate environmental disclosure.
The recurring finding across all of these is a wide variance: top performers genuinely transparent; the bottom of each benchmark routinely opaque. The 2023 collapse of the major ESG fund label "Pure ESG" in Europe — following coordinated regulatory action across France, Germany, and the Netherlands — is part of the consequence.
The media-and-advertising industry is the connective tissue that makes every other industry on this list profitable. Without it, the manufactured-insecurity model of beauty does not work; without it, the rapid replacement cycle of fast fashion does not move volume; without it, factory farming does not project comfort onto its packaging.
Three categories of harm recur in the documented evidence. Image retouching — particularly in advertising directed at women and girls — has been the subject of regulatory action in France (the 2017 "photoshopped" disclosure law, requiring "photographie retouchée" labelling on commercial images where a model's body shape has been digitally altered), Israel (2012), and Norway (2022). The UK Advertising Standards Authority has banned individual ads for "irresponsibly thin" models on a case-by-case basis but does not require general disclosure.
Influencer non-disclosure — paid promotion presented as personal recommendation — is now actively enforced. The US FTC's Endorsement Guides require clear and conspicuous disclosure of material connections. UK CMA / ASA joint enforcement issued an "Influencer's Guide" and brought actions against creators including those promoting cosmetic and fashion brands. The 2024 EU Digital Services Act extends disclosure obligations to very large online platforms.
Advertorial disguised as editorial — "native advertising" — is regulated by the FTC, ASA, and equivalent bodies, but compliance is uneven. The 2015 FTC Native Advertising guidance remains the foundational US document; the UK CAP Code rule 2.4 is the equivalent.
Beneath these specifics is the structural question: the attention economy's incentives reward emotional intensity, certainty, and outrage. The Reuters Institute Digital News Report tracks this annually across more than 40 countries; the picture is consistently one of declining trust in news while platform-driven content displaces it.
Most issues belong to several industries at once. A fur coat is a Fashion question, an Animal Welfare question, an Environment question, and a Women's Rights question simultaneously — fur retail is overwhelmingly marketed to women and overwhelmingly produced by low-paid female labour. We tag accordingly. The categories above are an entry point, not a fence.
Spot a figure that needs updating? An organisation we should add? Email info@feral-femme.co.
Four interactive tools that turn the editorial into action — log your wardrobe, plan your week of meals, decode an ingredient list, or look up what a brand’s sustainability claim actually means. All data stays on your device.
The 30-wear test: a garment isn’t truly purchased until it’s been worn at least 30 times. Log what you own, tap to count each wear, and see your real cost-per-wear. Items under 30 wears stay flagged — a reminder before buying again.
Map your week of meals. Tap each cell to cycle Plant → Vegetarian → Omnivore. The dashboard updates live with the water, carbon, and land you’d save compared to a fully omnivorous week.
Paste any ingredients list — from cosmetics, food, household products — and the decoder flags animal-derived ingredients, palm-oil derivatives, and controversial chemicals. Each flag explains what the ingredient is and why it’s flagged.
A reference for the language brands use when they want to suggest virtue without committing to it. Each term shows what brands typically claim, what the term actually means in regulation, and where the line is.
These tools use industry-average figures, not brand-specific data. A leather jacket from one tannery may have a footprint dramatically different from another — there is no honest way for a tool like this to know which. What we can offer are the published averages from peer-reviewed sources and let you make the comparison.
All data stays on your device. Your wardrobe items, plate logs, and ingredient scans never leave your browser — they’re stored locally and only you can see them. Clearing your browser data clears the tools.
The decoders are heuristics, not verdicts. An ingredient flagged as “potentially palm-derived” might be sourced from coconut. A “biodegradable” claim might be substantiated by certification we don’t list. The tools surface the question; they don’t answer it for you.
Editorial campaigns, magazine covers, and visual essays. Every image is a deliberate frame — beauty, fashion, food, tech industries placed on the same surgical table as the bodies they exploit.
All editorial works © 2026 Feral Femme · Ayden Rosemary Brown. Press & editorial enquiries: ayden@feral-femme.co
Animal cruelty across every industry, continent, and species — sourced from UN FAO, USDA, UK Home Office, EU statistical reports, Humane Society International, peer-reviewed journals, IUCN, CITES, and government databases.
All counters and time-scale breakdowns are calculated from verified annual figures by dividing by 31,536,000 seconds in a year. The numbers updating in real time are mathematical projections from peer-reviewed and government data, not live tracking. No organisation tracks animal cruelty in real time.
Mathematical projection from UN FAO 2024 data: 92.2 billion land animals slaughtered annually for food alone, plus 1.1–2.2 trillion fish, plus testing, fashion, fur, and exotic-trade industries.
All animals across all industries combined — calculated from verified annual data divided by exact time intervals.
| Time | Land Animals | + Fish | + Lab | + Fashion |
|---|---|---|---|---|
| Per second | 2,924 | 38,000+ | 3.66 | 63.4 |
| Per minute | 175,440 | 2.28M | 219 | 3,805 |
| Per hour | 10.5M | 137M | 13,162 | 228k |
| Per day | 252M | 3.4–6.5B | 316k | 5.5M |
| Per week | 1.77B | 23.9–45B | 2.2M | 38.5M |
| Per month | 7.68B | 100B+ | 9.6M | 167M |
| Per year | 92.2B | 1.2–2.3T | 115M | 2B+ |
Each continent has distinct dynamics — production, regulation, welfare standards, cultural practices. Tap a card to view detailed country-level data.