How Much Cheaper Are Third-Party Marketplaces? We Measured
"Cheaper than Steam" gets thrown around by every skin site on the internet, usually without a number attached. We have two numbers, from two different instruments: a purchase ledger of our own automated buys, which has averaged ~36% below same-day Steam prices, and a daily snapshot of roughly 6,000 items priced across 10+ marketplaces at once. Here's what the data actually shows — including where the discount thins out.
Two instruments, one story
The ledger is the strict measure: real money spent, compared against the Steam price of the same item on the same day, weighted by dollars. It says ~36% average savings — a figure that updates live on our homepage as the bot keeps buying. The snapshot is the wide measure: every morning we pull the lowest listed price for ~6,000 items from each venue's public feed and line them up against Steam — the same data you can browse yourself, free, on the cs2stack marketplace. It can't capture execution (fees, fills), but it shows the shape of the discount across the whole market, not just what we happened to buy. The two agree where they overlap, which is what you want from two instruments.
Where the discount is deepest: commodities
Cases and other commodity items show the fattest, most persistent gaps — in the snapshot they sit routinely a third or more below Steam, day after day. A recent morning's examples: a case at $1.56 on Steam and about $0.97 at its cheapest venue; another at $0.36 and $0.27. The pattern holds across the category because commodity supply is deep and sellers dump these items in volume for cash. If your buying skews toward cases, expect your personal average to look better than 36% — the full commodity playbook is in the cases guide.
Where it thins: liquid, hyped skins
The most-traded rifle skins — the AK and M4 classics everyone prices constantly — show the narrowest gaps, often in the 10–20% band rather than 30%+. Makes sense: they're the items where cross-market attention is highest, so mispricings get arbitraged away fastest. Gloves and knives tend to sit in between, with wide item-to-item variance because each listing's float and pattern muddy the comparison. Treat these as observed tendencies from our snapshots, not laws — the dispersion within every category is large, which is precisely why per-item comparison beats any rule of thumb.
| Category | Typical gap vs Steam (tendency) |
|---|---|
| Cases & commodity items | Deepest — routinely ~30–40% |
| Knives & gloves | Middling, high variance per listing |
| Liquid meta skins | Thinnest — often ~10–20% |
Why the gap doesn't close
A discount this visible should be arbitraged to nothing — and it isn't, because the "arbitrage" is blocked by design. Steam wallet funds can never leave Steam, so buying cheap off-platform and selling on Steam yields credit, not profit. Sellers who want actual money must accept cash-market prices; buyers who don't compare keep paying Steam's. The gap is a toll on the locked wallet, and it persists for the same reason the wallet stays locked: it's the business model. The full structural breakdown is in why prices differ across sites.
What to do with the numbers
Don't chase the average — collect it. The 36% isn't a coupon you apply once; it's the residue of a habit: compare every item across venues, buy at the cheapest effective price, repeat. On any single purchase your mileage will be the category's, not the headline's. Over a hundred purchases, the ledger says the habit is worth roughly a third of everything you spend — and there's no fine print, because you can run the same system yourself.