Thrift store jewelry is priced by volunteers working fast, not by appraisers, which is exactly why real pieces end up in the tangle bin next to the plastic beads. The trick is having a short, repeatable check you can run standing at the shelf, in bad lighting, with a queue behind you and nothing in your pocket but a phone and a magnet.
The fastest reliable sequence is marks, magnet, weight, and wear points. Read the stamp on a clasp or inner band first, check that nothing pulls on a magnet, feel whether the piece is heavier than it looks, then examine the high-wear edges for a different metal showing through. Four checks, under a minute, and they screen out most of the shelf.
Each check takes seconds, and each has a point past which it stops telling you anything. Both halves matter.
Start with the marks, because they cost you nothing
Under the FTC’s jewelry guides, precious-metal claims made on a piece are supposed to reflect what it actually is, which makes any mark the single most informative thing on a secondhand piece. Marks hide in predictable places: the inside of a ring band, the flat tab beside a necklace clasp, the post or butterfly back of an earring, the underside of a bracelet’s box clasp, the back plate of a brooch near the pin.
What you are looking for, in rough order of how much it changes the price you would pay:
- A karat or fineness mark alone. 10K, 14K, 18K, or a fineness number with no letters after it, points to solid gold. The fineness numbers are parts per thousand: the Assay Office lists 375 as 9ct, 585 as 14ct, 750 as 18ct, and 916 as 22ct, so a 375 mark is 9 karat rather than 10, which matters when you are estimating what the metal is worth. 925 or “sterling” points to sterling silver, which the FTC defines as containing 92.5 percent pure silver. PLAT, PT, or 950 points to platinum.
- A karat mark followed by letters. GP, GEP, RGP, HGE, or GF changes the answer entirely. These are coating and overlay marks, and the International Gem Society notes that many plated pieces carry no weight or content indication at all, so the letters are the only clue you get.
- A maker’s mark or signature. On a costume piece this can matter more than the metal, and it is worth photographing even when the metal reads as base. Signed fashion jewelry has its own market, which is the point of is costume jewelry worth anything.
- Nothing at all. Unmarked is genuinely ambiguous. Marks wear off over decades, and plenty of older and imported pieces were never marked. Unmarked pushes you onto the physical checks rather than settling anything.
A magnifier helps enormously here, and your phone camera is one. Photograph the mark and zoom into the image rather than squinting at the piece. That habit is the whole basis of photographing jewelry for identification, and it works just as well standing in a shop aisle.
The magnet test, and what it can and cannot rule out
Carry a small strong magnet. Gold, silver, platinum, copper, and brass are not attracted to it, so a piece that pulls toward the magnet has steel or iron somewhere in it and is not solid precious metal throughout.
Two honest limits keep this from being a verdict:
- Steel hides in the working parts. Spring clasps, pin mechanisms, and watch parts often contain steel even on genuine precious-metal pieces. Test the body of the piece, not the clasp, and if only the clasp pulls, that tells you about the clasp.
- Passing is not proof. Brass, copper, lead, aluminum, and most plating base metals are not magnetic either, so plenty of costume jewelry passes the magnet test comfortably. It rules things out; it never rules them in.
Weigh it in your hand, then look at the edges
Precious metals are dense, and hands are surprisingly good at density once you have handled a few real pieces. Gold and platinum feel noticeably heavier than they look. Silver is lighter than gold but still denser than the base metals used under most costume plating. Hollow construction muddies this badly, so treat weight as a nudge rather than a test: a chunky chain that feels weightless is either hollow or base metal, and you cannot tell which by hand.
Then look at the places the piece rubs, because that is where a coating fails first:
- Ring bands wear on the underside and at the edges.
- Chains wear where links rub, especially near the clasp.
- Earring posts and backs wear where they pass through and clip.
- Brooch pins wear on the pin itself and the catch. Vintage Christmas tree pins are a common thrift find in this group, and their makers’ marks are covered in vintage Christmas tree pins.
Yellow-pink or grey metal peeking through a gold-colored surface means a coating over something else, usually brass, copper, nickel, or steel. On a silver-colored piece, a warmer tone under the surface means the same thing. This is one of the few checks that distinguishes plated from solid without any tools at all.
Look past tarnish, and past dirt
The single most common thrifting mistake is walking past a genuinely good piece because it looks dull. Silver tarnishes to a dark grey or near black over years in a drawer, and that tarnish is a surface reaction that polishes off completely. A blackened, unloved-looking chain with a 925 mark is a normal find, not a damaged one.
The same goes for grime in the settings of an old ring, cloudy stones that are simply dirty, and the greenish film that copper alloys leave on skin and metal. None of these say anything about what the piece is made of. What genuinely does matter for condition is structural: cracked shanks, missing stones, sprung clasps, snapped pin catches, and repairs that were done badly. Those cost money to fix and should come off what you are willing to pay.
Meanwhile, be more suspicious of things that look too good. A bright, unworn, heavy “gold” chain at costume prices, with no mark anywhere, is exactly what a well-made plated piece looks like.
Where the shelf checks stop
Two questions cannot be answered standing in a shop, and it is worth being clear about them so you price your risk properly.
Exact karat. A mark is a claim, and thrifted pieces have been through decades of hands. Confirming solid gold and its actual karat takes acid testing on a touchstone or electronic testing, which the International Gem Society describes as the standard laboratory approach for exactly this reason. That is a jeweler’s bench job, not an aisle job. The full method is in how to tell if gold is real.
Gemstone identity. Nothing you can do at a shelf separates a natural stone from a synthetic or a well-cut glass. Color, sparkle, and setting quality shift the odds, and that is all they do.
If you are weighing thrift shelves against pre-sorted trays, estate sale jewelry covers how to shop those sales and what to check first.
That is the practical case for a photo-based first read while you are still deciding. A scan will not certify a stone or a karat either, but it gives you a structured opinion on the likely type, materials, and value range in the time it takes to walk to the till, which is a much better basis than a guess. What that kind of tool can and cannot honestly deliver is spelled out in can a jewelry identifier app tell you what your jewelry is worth.
A one-minute routine to memorize
- Find the mark, photograph it, zoom in.
- Magnet the body, not the clasp.
- Check the wear points for a second metal.
- Price the risk: pay costume money for anything unconfirmed, and treat a good mark as upside rather than as the reason to buy.
Scan anything you are undecided about with GemPeek before you commit. It flags the likely type, visible stones and materials, style, condition, and an estimated value range, which is enough to tell a considered gamble from a bad one, and you can take the good finds to a jeweler afterwards.