How ValU Estimates Work
When you scan an item, ValU returns an estimated value range. This page explains, in plain language, what goes into that range, how to read it, what can make it less accurate and why it is a starting point rather than a formal appraisal.
Last reviewed 2026-10-16 by the ValU team.
Step 1: identifying the item
ValU first tries to work out what the item is. It uses AI image recognition on your photo, along with any text it can read, such as brand names, model numbers, labels and maker's marks, and any description you type or say. The goal is a specific identification: not just a camera, but the maker and model; not just a trading card, but the year, set, card number and player.
The quality of everything that follows depends on this step. A clear, well-lit photo that shows identifying marks is the single biggest factor in a useful estimate.
Step 2: looking at comparable sales
Once an item is identified, ValU looks for information about what similar items have sold for. The principle is the same one experienced resellers use: completed sales are better evidence than asking prices. Our data sources page describes the marketplaces and references ValU draws on.
Comparable sales are rarely perfect matches. They differ in condition, completeness, version and timing. ValU uses the available information to estimate a typical range rather than a single precise figure.
Step 3: producing a range and a confidence level
The result is a low-to-high range that reflects how much prices for similar items vary. A narrow range usually means there are many consistent sales of a well-identified item. A wide range means prices vary a lot, the identification is less certain, or condition has a large effect on value.
Where ValU shows a confidence indicator, it reflects how sure the system is about the identification and how much comparable data was available. Lower confidence is a signal to check the identification and do more research before relying on the number.
What affects accuracy
- Photo quality: blur, glare, poor lighting or cropped marks make identification harder.
- Hidden details: perforations, watermarks, metal purity, gemstone quality, inner damage and authenticity usually cannot be judged from a photo.
- Condition: small flaws that matter greatly to collectors, such as a soft card corner or a hairline crack, may not be visible.
- Rarity: very rare items have few comparable sales, which makes any estimate less certain.
- Market changes: prices move, especially in fast-moving categories, and recent changes may not yet be reflected.
- Variants: items that look alike but differ in edition, printing or version can be confused.
How to get the best estimate
- Photograph the item in bright, even light against a plain background.
- Include close-ups of labels, marks, serial numbers, signatures and dates.
- Photograph the back, the underside and any damage.
- Add a short description with anything you know, such as size, weight or where it came from.
- Check the identification ValU returns. If it is wrong, the value will be wrong too.
What an estimate is not
A ValU estimate is an automated, informational starting point. It is not a certified appraisal, an authentication, a grade or an offer to buy. It should not be used for insurance, tax, legal or estate purposes. The actual price you can get depends on where and how you sell, the item's true condition and the buyers who happen to be looking at that moment.
For items that may be valuable, we recommend checking recent sold listings yourself, reading the relevant guide on this site, and, where the stakes are high, consulting a qualified appraiser or a recognized grading or authentication service.
Our editorial approach to guides
The guides on ValU are written to help people understand value, not to promote any particular item or marketplace. We aim to describe widely accepted practices from each collecting field, such as standard grading scales and identification methods, and to say clearly where opinions differ or where the answer depends on the specific item. We avoid quoting price figures that change quickly, and we do not publish invented statistics or testimonials.
If you spot an error or think something is unclear, please let us know through the contact page. We review guides periodically and update them when practices or information change; each guide shows the date it was last reviewed.
Reading an estimate in practice
Imagine you scan a piece of glassware at a thrift store and ValU returns a range with a fairly wide spread. That spread tells you something useful: prices for similar pieces vary, perhaps because of pattern, size, color or condition. Before buying to resell, check the identification and the specific pattern, look closely for chips or cloudiness, and compare the item with several recent completed sales. If your piece matches the better examples, the upper half of the range may be realistic; if it has flaws, expect the lower end.
Now imagine a scan returns a narrow range for a common, well-identified item. That usually means the market is consistent, and the estimate is a reasonable guide to what similar items sell for. In both cases, the estimate is most valuable as a quick filter that tells you where to spend your attention.
Your data
Photos and descriptions you submit are used to produce your estimate and to operate the service, as described in our privacy policy. Please do not include personal information such as addresses or documents in photos you scan.
Why estimates can change
If you scan the same item again later, the range may be different. That can happen because new sales have occurred, because the market for that category has moved, or because a different photo led to a more or less specific identification. Changes are a normal part of how markets work. For an important decision, note the date of your estimate and repeat your research close to the time you buy or sell.
We also improve the identification and pricing process over time, so the same photo may produce a more specific result after an update. Significant product changes are listed on our updates page.