The best pricing tools for online resellers

Terapeak Product Research has been free to all Seller Hub sellers since 2021 — no Store subscription needed. It gives you real eBay sold data with average sold price, sell-through rate, and seasonality, which makes it the best single answer to "what does this normally go for, and does it actually move?"

Two limits worth knowing. Sourcing Insights, the sibling tool for category-level demand, still requires a Basic Store subscription or above. And Terapeak reports the listing price on a sale, so items that closed through Best Offer look like they sold for more than the buyer paid.

This is the fix for Terapeak’s blind spot. When a listing sells via Best Offer, eBay’s own sold search shows the asking price, not the accepted number — so on categories where haggling is the norm, every comp you pull reads high. 130point surfaces the amount the buyer actually paid.

If you sell trading cards, comics, coins, or memorabilia, this changes your numbers materially, and it costs nothing. Use it as a sanity check on any category where a big share of sales close below list.

A condition-tiered price guide rather than a search tool: for a game you get loose, complete-in-box, and sealed; for cards you get raw and graded tiers. That structure is exactly what these categories need, because condition is most of the price and a flat "average sold" hides it.

Browsing prices is free. The $6/mo Collector tier adds deal alerts, a lot-value calculator, and grading recommendations. The $49/mo Legendary tier is the one a shop buys — it unlocks downloadable price lists and API access to the price data, which is what makes bulk trade-in pricing possible.

Card-specific and deep: verified sales history, market indexes, population reports for graded cards, price alerts, and a camera search that identifies a card from your phone. The free tier gets you the price guide, the daily sales recap, and the indexes.

Pro at $20/mo (or $200/yr) is where the all-time sales history and CL Value estimates live. If cards are a sideline, the free tier plus 130point covers you; if cards are the business, Pro pays for itself on one correctly-priced slab.

Where ordinary comps run out. A piece of unmarked pressed glass, an unsigned oil, a mid-century chair with no maker’s label — these barely appear in current marketplace data, and WorthPoint’s value is an archive of past auction results going back two decades, plus a Marks & Logos library for identifying what you actually have.

It is not an AI tool and it is not fast. Treat it as the reference you open for the item you cannot identify or cannot comp, not the tool you run on every listing. WorthPoint does not publish its rates on a public page, so confirm the current price in the signup flow before you commit.

Catalog-driven markets, so the data is unusually clean: a live bid and ask plus sales history for a specific SKU in a specific size. For deadstock pairs that is close to a spot price, and it updates constantly.

The trap is condition. Bid/ask is for new, authenticated pairs — read it as your ceiling and discount from there for anything worn. Size matters as much as model, so never comp a 9 off a 12.

The most underused pricing tool is the sold filter inside the app you already sell on. Poshmark, Grailed, Depop, Vinted and Mercari each show you what their own buyers paid — and for the same garment those numbers can sit well apart from eBay, because the audiences are different.

The catch is labor: doing it properly means five apps, five searches, per item. That is precisely the job the Pricing Engine automates, but if you sell one category on one platform, the native filter is free and it is the most relevant data there is.

Not a pricing tool, and it belongs on the list anyway: bad comps are usually an identification problem, not a pricing problem. Get the model, colorway, pressing, or year wrong and every number downstream is wrong with it.

Point Lens at the item, get a candidate model, then confirm against a tag, a stamp, or a serial before you comp. Thirty seconds here saves you from pricing a reissue off the original’s sold data.

Name the exact thing first

Model, colorway, size, year, pressing, edition. Comps are only as good as the identification behind them — and a reissue priced off an original is the most common expensive mistake in resale.

Pull sold comps, and ignore the live asks

An active listing is evidence that a price did not work. Sold rows are the only data that tells you what someone paid. If a tool only shows you what is currently listed, it is telling you about other sellers’ hopes.

Take the middle of the cluster, not the top comp

One $400 sale among nine $180 sales is noise — a bidding war, a mispriced auction, a different variant. Price off the median of a tight cluster; if the cluster is wide, that is information too, and it means you need a range and a markdown plan rather than a number.

Subtract fees and shipping before you decide

A $60 comp is not $60. Platform commission, payment processing, and the label all come out first, and the same item nets differently on every channel. Price to the number you keep.

Set the price drops when you list, not when you panic

Decide at listing time what happens at 21 days and at 45 days. Scheduled markdowns beat emotional ones, and they keep aging inventory from quietly becoming a death pile.

Set auto-accept and auto-decline off the comp range

Anchor the accept floor near the bottom of your sold cluster and decline below it automatically. Base the window on the cluster, not on a single high comp, and you stop negotiating the same item forty times.