At a glance
| Item | Detail |
|---|---|
| Client | Sports and trading-card analytics platform |
| Market | United States |
| Platform | eBay US |
| Scope | Sold listings for trading cards: price, date, card identity, grade, variant and sale format |
| Frequency | Daily |
| Delivery | Daily Parquet files and API |
| Service | eBay sold data scraping by WebDataScraping.us |
The client
The client "combines player analytics, market context and exact card intelligence to help collectors scout, value and invest." Its core need was simple to state: "eBay US sales data." Recent sold prices are the most widely used reference for trading-card values, and the client needed them structured by exact card, grade and variant.
The challenge
- Unstructured titles. Sellers write listing titles freely: "2023 Example Rookie Auto /99 PSA 10 GEM MINT!!" contains year, set, player, card type, serial numbering, grader and grade, all in one string.
- Exact card identity. Values differ greatly between base cards, parallels, autographs and serial-numbered versions.
- Grading. A grade 10 card can be worth many times a grade 9 of the same card.
- Sale formats. Auctions, fixed-price sales and "best offer accepted" sales need to be distinguished, as the final accepted offer price may not be shown.
The solution
Step 1: Define the card universe. WebDataScraping.us and the client agreed the sports, sets and years in scope, starting with major US sports and the most traded modern sets.
Step 2: Collect sold listings daily. Collectors captured publicly visible sold listings: title, sold price, sold date, sale format, shipping cost, item condition, item specifics (where the seller filled them) and listing ID.
Step 3: Parse titles and item specifics. A parsing model extracted year, set, player, card number, parallel or variant, autograph flag, serial numbering, grading company and grade. Item specifics were used first; titles filled gaps.
Step 4: Match to the client's card catalogue. Parsed sales were matched to the client's card IDs. Low-confidence matches were held for review rather than added automatically.
Step 5: Clean and flag. Outliers (for example, lots of multiple cards, or obviously mislabelled listings) were flagged. Best-offer sales were marked so the client could treat them separately.
Sample data delivered
The record below is illustrative.
{
"listing_id": "000000000000",
"sold_date": "2026-10-05",
"sale_format": "auction",
"sold_price_usd": 412.00,
"shipping_usd": 5.00,
"title": "2023 Example Set Rookie Auto /99 Sample Player PSA 10",
"year": 2023,
"set": "Example Set",
"player": "Sample Player",
"card_type": "rookie_autograph",
"serial_number": "/99",
"grader": "PSA",
"grade": "10",
"matched_card_id": "CARD-0048221-AU99",
"match_confidence": "high",
"is_lot": false
}| Card (illustrative) | Grade | Sales last 30 days | Median sold price | 30-day change |
|---|---|---|---|---|
| Rookie base | Raw | 410 | $18 | -4% |
| Rookie base | PSA 10 | 96 | $142 | +6% |
| Rookie auto /99 | PSA 10 | 7 | $405 | +11% |
| Rookie silver parallel | PSA 9 | 38 | $64 | -2% |
Results
| Metric | Outcome (illustrative) |
|---|---|
| Sold listings processed per day | 85,000 |
| Share matched to exact card and grade | 82% automatically, rising to 91% after review |
| Card IDs with a 30-day median price | 210,000 |
| Platform features launched | Price guide, player market index, grade premium tool |
The client launched card price guides and a player market index built on daily sold data. A "grade premium" tool showed collectors how much grading added to a card's value, helping them decide which cards to send for grading.
Market insights the platform now publishes
- Player indices: the combined value trend of a player's most traded cards, updated daily.
- Grade premiums: the ratio between graded and raw prices for each card, which guides grading decisions.
- Liquidity: how many sales a card sees per month, helping collectors understand how easily a card can be sold.
- Event reactions: how prices move after big games, injuries or awards, measured within days.
These features depend on clean, matched sold data at daily frequency, which a manual process could not provide.
Why it worked
- Sold prices, not asking prices, gave a true market reference.
- Structured parsing turned free-text titles into exact card identities.
- Review queues kept low-confidence matches out of published prices.
- Daily collection captured fast-moving markets around player performance.
How this differs from listing monitoring
WebDataScraping.us's sister brands also offer real-time new-listing feedsfor eBay, which focus on items as soon as they are listed. Sold data focuses on completed transactions and is better suited to valuation and price guides.
Power your valuations with sold data
WebDataScraping.us delivers compliance-first eBay sold data scraping for collectibles platforms, resale analytics and dealers. Share your categories and we will send a sample of last week's sold listings, parsed and matched.
Frequently asked questions
No. Only public listing and sale information is collected; no buyer data is gathered.
Yes. Coins, comics, sneakers, watches and other collectibles can be tracked with category-specific parsing.
Where the accepted offer price is not publicly shown, the sale is flagged so it can be handled separately.