How to Spot a Profitable Lot in 30 Seconds

You don't have hours to research every lot in a 400-item auction. Most professional resellers spend less than 30 seconds per lot deciding whether to dig deeper. Here's the fast-filter system they use.

The 30-second mental checklist

When you land on a lot listing, run through these five questions:

1. Is the title specific enough to price?

"Collection of items" = skip. "Sony WH-1000XM5 wireless headphones — tested working" = investigate.

If you can't extract a model number, brand + item type, or at least a clear product description, you can't price it in 30 seconds. Move on.

2. Is there clear demand on eBay?

Do a quick mental or actual eBay sold search. If there are fewer than 5 sold listings in the last 30 days, the market is thin. That's not necessarily bad (vintage collectibles trade slowly but at premium prices) but it means research will take longer.

3. Is the estimate realistic?

Auction house estimates are set by valuers. A lot estimated at £50–£80 that you know sells for £200 on eBay is a green light. A lot estimated at £150–£200 for something that moves for £90 on eBay means someone's hoping for a room full of uninformed bidders.

4. Is the condition workable?

"Spares or repair" electronics can be gold (test the screen, replace the battery) or a dead end (motherboard failure). "Collection of 50 mixed clothing items, grades A-C" means some unworn, some damaged — price the lot assuming 20–30% is unsellable.

5. Is there a clear sell path?

You should be able to answer "I would sell this on [platform] at approximately £[price]" within 30 seconds. If you can't, it's research work, not impulse buying.

The four lot types worth targeting

Type 1: Named brand + working condition

Examples: Nintendo Switch, Dyson vacuum, Barbour jacket, Apple AirPods, Stanley tools

These have deep sold-listing history, predictable demand, and fast sell-through. The risk is overbidding — they're popular, so competition in the auction room is real. Use the Max Bid Calculator rigidly.

Type 2: Vintage collectibles with specific demand

Examples: Vintage Levi 501s (specific wash/size), 1970s Wedgwood ceramics, 1980s band T-shirts, specific Lego sets

These have smaller buyer pools but much higher margin. A specific pair of 1995 orange tab 501s in 30x30 sells for £120–£180 on Depop. At auction, they might hammer for £15–£30 in a "mixed clothing lot."

Type 3: Bulk job lots with predictable salvage value

Examples: 50x mixed women's clothing, 100x books, 20x power tools

The key is being able to price the worst-case scenario. If 30% is unsellable waste, what's the remaining 70% worth? If the maths still works at that 70% yield, it's viable.

Type 4: Electronics with testable fault patterns

Examples: Laptops with "no power" (often just battery/charger), iPhones with "iCloud locked" (sometimes owner-resettable), consoles with "disc reader fault" (often cleanable)

These require basic technical knowledge but offer enormous margins. A MacBook Air "no power" that needs a new £35 charger and hammers for £80 is worth £350+ repaired or parts-sold.

Red flags to skip immediately

• "Untested, sold as seen" — Fine for small lots, a problem for anything electrical worth more than £50
• Condition photos showing water damage — Corrosion is almost never cost-effective to fix
• "From a house clearance, no documentation" — Fine for most items, major problem for anything requiring proof of ownership (watches, jewellery, high-end electronics)
• Heavy or large items with no collection option — Always check postage viability. A lot of 12 ceramic vases might profit £80 but cost £95 to courier
• Brand-sensitive items without authentication — Fake designer goods at auction will get your eBay account suspended

Using AI scoring as a pre-filter

Arbitrage AI's scanner pre-filters every lot using Claude's analysis. Anything scoring below 5/10 is automatically hidden. What the score reflects:

• 8–10: Clear demand, realistic estimate, known sell platform, good condition signals
• 6–7: Solid opportunity but research recommended — check sold comps before bidding
• 4–5: Borderline — depends on your specific knowledge of the category
• 1–3: Low demand, saturated, high risk, or insufficient information

The AI flags give you instant context:

• trending — demand is rising; sell quickly
• seasonal — Christmas toys in October = good; in March = hold
• saturated — active listings far exceed recent sold volume
• collectible — price depends heavily on exact variant/condition
• brand_sensitive — authentication may be required by the platform

---

Key takeaway: Most lots are not worth bidding on. Apply a fast mental filter — specific title, provable demand, workable condition, clear sell path. Let AI pre-scoring do the bulk filtering, then spend your research time only on the lots that pass the 30-second test.

→ See AI-scored lots from live auctions