The question comes up in the shop almost every day, usually phrased politely. How do you know it is real?
It is the right question. It is also one where a great deal of the advice online is out of date, because the brands themselves changed the rules in 2021. Here is what still holds, what no longer does, and where the honest limits are.
How big the problem actually is
Counterfeit goods accounted for USD 467 billion of world trade in 2021, or 2.3 percent of everything traded, according to the OECD and the EU Intellectual Property Office in their 2025 report. Imports of counterfeits into the EU alone came to EUR 99 billion, 4.7 percent of all EU imports. Clothing, footwear and leather goods are the largest counterfeited category group.
EU customs detained 112 million counterfeit items in 2024, worth EUR 3.8 billion. That was the highest value ever recorded, on a lower volume than the year before, because fakes now carry higher unit prices.
The resale numbers are more useful than the trade numbers. Entrupy, which runs a microscope-based authentication service, publishes its failure rate by seller type. In 2023, items it could not confirm as genuine ran at 1.4 percent from wholesalers, 5 percent on consumer-to-consumer marketplaces, 7.7 percent from online resellers, 10.1 percent from offline resellers and 15 percent at pawn shops. Two caveats: Entrupy is reporting on its own scans, and unidentified is not the same as counterfeit. The direction is still clear. Who you buy from moves the odds more than any single detail on the item.
What a label proves, and what it does not
A care label is a date marker. Permanent care labels have been mandatory in the United States since 1972 under the Federal Trade Commission Care Labeling Rule. An American garment without one is very probably older than that. In Europe care labelling is voluntary even today. The symbols come from GINETEX, founded in Paris on 26 March 1963 and registered with WIPO in 1969; the international standard ISO 3758 first appeared in 1991, with the current edition published in December 2023. Since 2006 the symbols must run in a fixed order: washing, bleaching, drying, ironing, professional care.
Made in West Germany is the cleanest bracket in vintage. The Federal Republic existed from 23 May 1949 until reunification on 3 October 1990. Nothing carrying that label was manufactured after October 1990. Labelled stock kept selling afterwards, so read it as a ceiling on manufacture, not on sale.
A union label narrows the decade. The ILGWU merged into UNITE in 1995, so no garment with an ILGWU label was made after that year. The red, white and blue version reading Made in the U.S.A. runs from 1974 to 1995. The AFL-CIO label with large lettering is 1955 to 1963, the smaller redesign 1963 to 1974.
An RN number proves less than people think. It is a code the FTC issues to companies based in the United States so they can print it instead of the company name. It is not required by law, it is never issued to a business outside the US, and a company keeps the same number for decades. It tells you which company registered the label. It does not tell you the year, the country of manufacture, or whether the piece is genuine. The arithmetic formula that circulates for dating RN numbers is community folklore, and experts at the Vintage Fashion Guild warn against it directly.
Three brand systems that changed in 2021
This is where most guides online are now simply wrong.
Louis Vuitton stopped using date codes in March 2021 and moved to an embedded microchip. A current bag with no date code is normal, not suspicious. Older codes are readable: from 1990 to 2006 the format is two letters plus four digits, where the first and third digits give the month and the second and fourth the year, so FL1003 is October 2003. From 2007 until early 2021 those same positions give the week instead. A code that decodes impossibly, week 55, or a country that never made that model, is a disqualifier. A code that decodes correctly proves nothing at all, because counterfeiters have been copying valid codes for decades.
Chanel replaced the serial sticker and the authenticity card with an NFC metal plate in April 2021, from the 21A collection onward. One rule follows from that and it is absolute: a genuine bag never carries both a serial sticker and a microchip. If you see both, stop. You also cannot read the chip yourself, because the technology is proprietary and only a boutique has the reader. And a missing card on an older bag means very little, since cards were always separable from the bag and are trivial to forge.
Hermes uses a blind stamp, not a serial number. It is embossed without ink and encodes the year plus the workshop. A bare letter means 1945 to 1970, a letter inside a circle 1971 to 1996, a letter inside a square 1997 to 2014, and from 2015 a bare letter again on a non-sequential system. The letters and numbers beside it identify the craftsman, not the date.
Saint Laurent gives you a hard line. The ready-to-wear line was renamed Saint Laurent Paris in 2012, so that label cannot be older than that. Yves Saint Laurent Rive Gauche belongs to the years 1966 to 2012. The Yves Saint Laurent name was kept for accessories, so a recent handbag can legitimately still carry it.
What the construction tells you
Zippers date a garment better than almost anything else. In the 1960s, seven of every ten zippers made anywhere in the world were Talon. The company took that name in 1937 and closed its American plants in 1998, so a Talon metal zip points firmly at American manufacture between roughly 1937 and 1980. YKK adopted its trademark in 1946 and went overseas from 1959, which makes a YKK zip in a dress presented as 1940s an anachronism. Lampo in Italy secured the Italian rights to the Ruhrmann patent in 1955 and became the Italian standard; riri in Switzerland has made them since 1936. The coil zipper was invented in 1940 and only became common in the early 1960s, so a nylon coil in a 1940s garment is wrong.
Stitching is about consistency, not counting. The checkable rule is that a symmetrical component must match itself. The handle tabs on a Louis Vuitton Alma, Speedy or Neverfull carry exactly five stitches each, identical on both sides. Be sceptical of anyone quoting a stitches-per-inch figure for a luxury house, because no house publishes one.
Edge finishing is the tell that is hardest to fake cheaply. At Hermes a thick edge, built from as many as eight layers of leather, takes nine or ten layers of finish: dyed with a colour-matched pigment, scored with a grooved heated iron, smoothed with a flat iron, polished with sandpaper, then sealed with melted wax. Done properly the edge is thick, slightly domed, glassy and continuous, and it sits on the leather stack rather than being painted onto it. A single thin coat, a flat edge, visible layer lines or chipping at the corners is the economical version.
The pattern must run through the seam. On monogrammed canvas the print continues uninterrupted across a seam. A pattern that jumps, mirrors or truncates is a cutting shortcut that saves material and no luxury house takes it.
What no photograph can do
This part is worth being blunt about, because it applies to every online purchase including ours.
A camera cannot capture the weight of hardware or how fast it warms in your hand. A plated zinc clasp photographs exactly like solid brass and feels nothing like it. It cannot capture how leather bends and springs back, or the smell of vegetable against chrome tanning, which is among the fastest tells an experienced hand has. It cannot record how deeply a stitch is set or how tightly the thread is pulled, only how the stitches are spaced. And it cannot read a microchip.
Even the machines have this limit. Entrupy cannot authenticate from images you send in; the photographs have to be taken through its own app with a microscope attachment at the moment of scanning. It reports accuracy of 99.1 percent in its published report and 99.86 percent on its FAQ, covers more than forty brands, and supports only selected styles within them. Rare, vintage and niche pieces fall outside the supported models entirely, which is exactly the territory a vintage shop trades in.
And the counterfeits have moved. A handmade fake Birkin now sells for over 6,000 dollars, against roughly 25 dollars for a fake bag twenty years ago. At that price the maker can afford the correct hardware weight, the correct leather and the correct stitching, which is to say everything a photograph would once have caught.
What that means in practice
Buy from someone who has the piece in their hands. Ask what the hardware weighs against something known, what the edges look like under raking light, what the lining smells like. A seller who has handled the piece can answer that in a sentence. One who is forwarding photographs cannot.
Treat a certificate as evidence, not as proof. Authentication is a professional opinion and experts do disagree. The same bag has been called genuine by one authenticator and counterfeit by another. A platform guarantee is real protection, not a verdict.
Make the evidence agree with itself. A care label that fits the decade, a zip from the right maker and the right country, a label style that matches those years, edges that are built rather than painted, and a price that sits where it should. Any one of these can be faked. All of them together, rarely are.
Every piece at Berlin Vintage Corner is inspected in the shop before it reaches the rail or the website. If you want to see the label, the stamp, the zip pull or the edge of something you are looking at online, write to us and we will photograph it that day. If you are in Berlin, come and handle it yourself. That is still the version that works best.
Once the question of whether a piece is real is settled, the next one is usually whether it will fit. We wrote a separate guide to vintage sizing and why the number on the label is meaningless.
Berlin Vintage Corner
Elisabethkirchstr. 1, 10115 Berlin, Mitte
Monday to Saturday 12:00 to 19:00, Sunday 12:00 to 17:00