eBay Product Research in 2026: How to Find Items That Actually Sell

Most eBay product research advice stops at "find trending items". That is not research, it is browsing. Research is elimination: you establish whether a product can be profitable for you, and you cut the ones that cannot.

I do this on Facebook and TikTok rather than eBay, but the discipline transfers exactly — the platform changes, the arithmetic does not. This guide covers the signals worth reading on eBay, the margin maths that stops most products before you list them, and where the free methods run out.

TL;DR — eBay product research in six lines

• Sell-through rate beats sold count. 50 sold from 60 listed is a market; 50 sold from 5,000 listed is a graveyard.

• Check completed listings, not active ones. Active listings show what people hope to sell. Completed listings show what actually sold, and for how much.

• Margin decides before demand does. After eBay fees, payment processing and shipping, a product needs roughly 3× its cost to be worth listing.

• Competition count matters more than category. A crowded category with a neglected sub-niche beats an empty category nobody searches.

• Free methods are enough to start. eBay's own filters and Terapeak cover the basics before any paid tool.

• Cross-platform demand is the leading signal. A product selling on TikTok today shows up on eBay weeks later.

The four signals that actually matter

1. Sell-through rate

This is the single most useful number on eBay, and the most skipped.

Sell-through rate = items sold ÷ items listed, over the same period. These thresholds are as observed on eBay's own sold-listing filters in August 2026.

A product with 200 sold looks strong until you see 8,000 active listings behind it. That is a 2.5% sell-through — you would be the 8,001st seller waiting. A product with 40 sold against 50 listings is an 80% sell-through: less glamorous, far more likely to move.

As a working floor, anything under 30% deserves a hard second look. Above 60%, demand is outstripping supply and the question becomes whether you can source it.

2. Completed listings, not active ones

Active listings tell you what sellers want to sell. Completed listings tell you what buyers actually paid.

Filter to completed and sold, then read three things: the realised price (not the asking price), how long items sat before selling, and whether the sellers are individuals or established shops. A category dominated by shops with thousands of feedback points is one where you compete on logistics, not on listing quality.

3. Margin after fees — where most products die

This is the step that eliminates most candidates, and it is the same arithmetic I apply before any product test.

The naive calculation is: sells for $30, costs $10, so $20 of margin. It is wrong. On eBay you subtract the final value fee, payment processing, shipping if you absorb it, and returns. What looks like $20 is closer to $10 — and if you paid for promoted listings, less again.

The practical rule: target a selling price of at least 3× the product cost, and count as margin only what survives fees, shipping and returns. In dropshipping the realistic net band is 20–25%, not the 60% a gross calculation suggests — I break that arithmetic down in my 2026 dropshipping explainer (French). That constraint is identical on eBay.

A product that cannot carry both a competitive price and a real margin is not a candidate, however well it sells for someone else.

4. Competition depth

Count the sellers, not the listings. Twenty listings from three sellers is a different market from twenty listings from twenty sellers — the first is a niche with room, the second is a commodity.

Look also at who they are. Established shops with tight supplier relationships can hold prices you cannot match. Saturation is relative to who you are competing against, not an absolute property of the product.

Minea

Top 100 best-performing products this month

Detected in real time by our AI from market signals: real sales, advertising statistics, and performance

See products

Free methods, in the order to use them

eBay's own search filters

Start here. Search your category, filter to Sold Items, then sort by most recent. You get realised prices, dates and volumes without any tool.

Two refinements make this far more useful: add a price floor to strip out low-value noise, and check the "Completed" filter alongside "Sold" to see the listings that expired unsold. The gap between those two lists is your sell-through rate, calculated by hand.

Terapeak

Included with most eBay seller accounts, and the fastest way to get sell-through without arithmetic. It reports sold counts, average price and sell-through over a chosen window.

Its limit is that it only knows eBay. It tells you what already sold there, not what is about to.

Cross-platform demand — the leading signal

This is where eBay research stops being retrospective.

A product's life usually starts on ad platforms: someone runs paid traffic to it on TikTok or Meta, it works, and the demand spills into marketplace search weeks later. By the time it ranks in eBay's sold data, the early sellers have been on it for a month.

Watching what is being advertised rather than what has sold moves you earlier in that cycle. I work through the filter sequence I use for this in this product-research walkthrough (French) — the tool is an ad spy, but the method is platform-agnostic: narrow to genuine e-commerce ads, then read engagement and longevity rather than raw view counts.

The criteria for judging what survives that filter are the same ones in winning dropshipping products.

When a paid tool earns its cost

Free methods take you a long way on eBay. A paid tool becomes worth it at one specific point: when the noise costs you more time than the subscription costs money.

The threshold question is the same one I apply everywhere — how many product tests does the tool need to save you to pay for itself? One avoided listing cycle, with its sourcing, photography and unsold inventory, usually costs more than a month of any research tool.

What a tool should buy you is earlier signal, not more data. More sold-item history is easy to get free. Knowing what is ramping right now is not.

Sourcing: the constraint nobody researches

Product research that ignores supply is half-finished. A product with perfect demand signals and a six-week lead time is a refund queue.

Check three things before committing: whether the product is reliably available, what the realistic delivery window is, and whether you can improve on what the current sellers offer — better packaging, a bundle, a variant they do not carry.

That last point is what separates a seller from a reseller. Undercutting on price is a race you lose to whoever has better volume. For the supplier side, dropshipping agents covers what changes when you move past marketplace sourcing.

A repeatable weekly routine

1. Pick one category you understand.

2. Pull completed + sold listings for the last 30 days.

3. Calculate sell-through; discard anything under 30%.

4. Run the margin test at 3× cost; discard what fails.

5. Count distinct sellers on what survives.

6. Cross-check the survivors against what is being advertised elsewhere.

7. List two or three, not ten. Learn from the ones that move.

FAQ

What is a good sell-through rate on eBay?

Above 60% is strong, 30–60% is workable, and below 30% means supply outstrips demand — thresholds measured against eBay sold-listing data in August 2026. Always calculate it over the same period for sold and unsold listings — comparing a month of sales against all-time listings inflates it badly.

How do I see what sold on eBay, not just what is listed?

Use the Sold Items filter in eBay search. It shows realised prices and dates. Comparing that against the "Completed" filter, which includes unsold listings, gives you the sell-through rate.

Do I need a paid eBay product research tool?

Not to start. eBay's own filters plus Terapeak cover demand and pricing. A paid tool is worth it when you need earlier signal than marketplace history provides, or when manual filtering costs more hours than the subscription costs money.

What margin should I target on eBay?

Price at roughly 3× product cost, then count as margin only what remains after final value fees, payment processing, shipping and returns. Expect a realistic net in the 20–25% range rather than the 50–60% a gross calculation implies, according to the margin breakdown in this video (French).

Is eBay product research different from dropshipping product research?

The signals differ, the discipline does not. eBay gives you retrospective sold data; ad platforms give you leading demand. Both are subject to the same margin and sourcing constraints, and both reward elimination over discovery.

How many products should I test at once?

Two or three. Testing ten means you learn nothing from any of them, because you cannot attribute what worked. One variable at a time is the rule on ads and it holds on marketplace listings too.