Every collectible gets priced against something. The question is whether the something you picked actually tells you what a buyer will pay, or just what another seller hoped for.
Active listings are hopes. They show you what people are asking, which is a different thing from what people are getting. I can list a common postcard at whatever I want, and it will sit there at that price for months, looking for all the world like a data point to the next seller who searches the same item and sees my listing first. That number is a wish, not a data point, and wishes are contagious. One optimistic listing produces two more priced against it, and now the category has a consensus price that nobody has ever paid.
Sold listings are the correction. They show what a buyer actually handed over, when, and for what condition. That is the comp. Everything in this piece is about how to pull comps well and how to read them honestly once you have them.
How to get to the sold filter
eBay gives you two roads to the same data.
The fastest is the search sidebar. Run a normal search for the item, then turn on the sold-items filter in the results filter panel. The results repopulate with completed sales only, each one stamped with the date it sold. On the mobile app the same filter lives behind the filter button at the top of any search result.
The other is Advanced Search. Select “Advanced” next to the search bar, enter your keywords, and check “Sold listings” before you search. eBay’s own pricing guidance describes this path plainly: navigate to Advanced Search, input descriptive keywords, choose Completed listings, and select Search. This route is better when you want to lock in a category and other constraints before you run it.
One distinction worth keeping straight: “Sold” shows only the listings that actually moved. “Completed” shows everything that ended, sold or not. Both are useful. The sold filter tells you what buyers paid. The completed filter, which includes the unsold listings alongside the sold ones, tells you what they passed on, and the gap between those two numbers often teaches you more than either one alone.
Ninety days, and the wall behind it
The standard sold-listing search covers roughly the previous ninety days. After that, the listings fall out of the public search results and you cannot find them by scrolling.
For a common item that sells regularly, ninety days is plenty. You will have enough comps to see a range, spot the outliers, and land somewhere in the middle with confidence.
For something scarce or seasonal, ninety days can be empty or misleading. A holiday-themed collectible that sells every November and December will show you nothing in June. A piece that changes hands once or twice a year may have zero comps in a ninety-day window, which does not mean it is worthless. It means the window is too short for the item.
eBay’s own Product Research tool, free to all sellers inside Seller Hub’s Research tab, reaches back up to three years. eBay describes it as providing “recent eBay supply, demand, and pricing data to help you determine what to sell, when to sell it, and at what price.” Where the standard search gives you the raw listings, Product Research gives you calculated metrics: average sold price, sold price range, sell-through rate, average shipping costs, and the number of sellers who moved the same item. You can search by keyword or by product identifiers like UPC or ISBN, and filter by category, condition, and other fields.
For a scarce piece, the three-year window is the one that matters. A single comp from two years ago is thin evidence, but it is real evidence, worth more than the zero comps you get from a ninety-day search on something that rarely trades.
The comp has to match what you are selling
This is where most pricing errors happen, and they happen because the seller found a sold listing for the same item and stopped reading.
Condition is the biggest variable in collectibles, and it moves the price more than most sellers expect. A near-mint example and a heavily worn example of the same piece are not the same comp. If you are holding a piece with a crease, a chip, a missing part, or visible wear, the comp you need is the one that sold in the same state, not the clean copy that went for the top of the range.
Read the comp listings themselves, not just the prices. Look at the photos. Read the condition descriptions. A sold price attached to “some shelf wear, light foxing” is a comp for your copy with shelf wear and light foxing. A sold price attached to “mint in original box, never displayed” is a comp for somebody else’s copy, not yours.
Variants matter the same way. A first printing and a later printing are different comps. A complete set and a partial set are different comps. A signed copy and an unsigned copy are different comps. The more specific you are about what you are actually holding, the closer your comp gets to the truth, and the less time you spend wondering why your item sat for weeks at a price that seemed right.
How many comps is enough
One is not enough. One sold listing is an anecdote. It might have been a sniped auction that ended low because nobody was watching, a best-offer acceptance well below the listed price, or a motivated seller who priced to clear. You do not know, and one number cannot tell you.
Three to five comps in similar condition start to show you a range, and the range is what you are after. Look at where the middle sits. If the prices cluster tightly, you have a market telling you something clear. If they spread wide, something is driving the gap, usually condition, completeness, or a variant you have not identified yet, and that is worth investigating before you price.
When you cannot find three comps in the ninety-day window, widen the search. Drop a keyword, broaden the category, or move to the three-year view in Product Research. When you still cannot find them, you are pricing a genuinely scarce item, and the honest move is to price it where the evidence lands and be willing to revise. A best-offer listing is built for exactly this situation.
Auction comps and fixed-price comps are different animals
An auction that ended with one bid tells you the starting price found one taker. It does not tell you the market price. An auction that ended with fifteen bids tells you something closer to what the item is worth to the buyers who happened to be watching.
Fixed-price comps are more stable but carry their own blind spot. A listing that sold on day one was probably priced below the market. One that sat for three months was probably priced at the ceiling. They tell you different things, and averaging them without noticing the difference flattens the signal.
Best-offer sales are the hardest to read. The sold price shown is the accepted offer, which is real, but you do not know how many offers came in before it or how long the listing ran at the original price. It is still a comp, just one with less context than the others.
Category quirks that change the window
Not every collectible category comps the same way.
Trading cards move fast and in volume. Ninety days is usually enough, the data is dense, and prices shift week to week with the sport’s season. A card that comped at a certain level in April may comp differently in August, and the three-year view can mislead if the market has moved structurally since then.
Vintage pottery, glass, and decorative items move slowly. A specific maker and pattern might produce two or three comps per year. The three-year view is the working window for these, and even that can be thin. The spread between a perfect example and a chipped one is wider here than in most other categories.
Records sit in between. A common pressing sells often enough that ninety days works. A specific rare pressing on a specific label from a specific year is a one-comp-per-year item, and you are back to the long window and the honest admission that you are pricing on thin evidence.
Paper, ephemera, and one-of-a-kind items are the hardest to comp, because there may not be another one. When the exact item has no comp, you are comping by analogy: similar age, similar type, similar condition, similar subject. That is weaker evidence, and the price should reflect the uncertainty.
What to do with the number once you have it
The comps give you a range, not a price. Where you land in that range is a decision, and it depends on how fast you want to move.
Price at the low end of the sold range and the item moves quickly. Price at the high end and it may sit, but it might catch the buyer who wants exactly that piece. Neither is wrong. It depends on your inventory, your storage, and whether the thing is appreciating or depreciating while it sits.
eBay shows you the median sold price of similar listings as you build a new listing, right there in the pricing step. That number is a reasonable starting point for a common item in average condition. It is the wrong starting point for a piece in unusually good or unusually poor shape, and it is a guess on anything scarce. Treat it as a check, not as the answer.
The comp is the foundation. Lay it every time, and pricing stops being a feeling and starts being a method.
How to write the listing that matches the price you set, starting with what goes into the title and what the item specifics panel will ask for, are their own pieces in this group.