Sold evidence
How Many Sold Comps Do You Need to Value an Item?
Learn how many sold comps to compare, why relevance beats quantity, and how medians, outliers and recent sales affect confidence in a resale estimate.
By Frank Cheung · · 7 min read
Aim for several relevant sales, not a magic number
Three to five closely matching sold comps can provide a useful initial indication of resale value. Six to ten can give a broader comparison, but no number guarantees an accurate price. Relevance, condition, recency and genuine sold evidence matter more than a large total of loosely related listings.
A comp is a comparable sale used as a pricing reference. The sold-comps guide explains the basics. Think of the sample as evidence to inspect, not a vote in which every listing deserves equal weight. Three sales of your exact model can be more informative than twenty accessories carrying its name.
What counts as a valid sold comp?
Start with the same product or model, comparable condition and similar specifications. Size, colour and edition matter when buyers value them differently. A large-capacity device, scarce trainer size or limited edition is not automatically interchangeable with a common version.
Confirm that the listing sold. A completed listing can have ended without a buyer, while an active listing is only an asking price. Public best-offer results may not reveal the final accepted amount. Keep uncertain amounts separate rather than treating the displayed figure as verified.
Use recent transactions in the relevant geographical market where possible. Check whether shipping is included, what accessories came with the item and whether the sale was for a bundle. A camera with lenses and a body-only camera belong in different comparisons unless you can reliably separate their values.
Practical guidelines for how many comps to use
The following bands are research guidelines, not statistical confidence levels. They assume that you have already excluded obvious mismatches and duplicate representations of the same listing. Moving into the next band does not automatically make your estimate dependable.
| Sold comps | Interpretation |
|---|---|
| 0 | No direct sold-price evidence |
| 1–2 | Very limited evidence |
| 3–5 | Useful initial indication |
| 6–10 | Broader comparison sample |
| 10+ | Potentially more robust, if genuinely comparable |
One or two sales: investigate rather than conclude
Inspect each listing closely and work with a wider possible price range. One bargain auction or one unusually high collector purchase can dominate a tiny sample. Seek more evidence before paying a price that only works if that result repeats.
Three or more: assess consistency
Check whether the prices cluster, whether condition differs and whether sales occurred over a meaningful period. A consistent group is helpful, but three near-identical listings from one seller may not represent the wider market. Do not count the same sale twice when it appears in multiple searches.
Quality matters more than quantity
Keep the correct brand, model and product category central to your comparison. Listings often mention compatible products in their titles: a replacement remote may name the television, and a phone case may name the handset. Keyword overlap does not make the accessory a comp for the whole device.
Condition differences can explain a wide spread. A pristine boxed example should not dominate an estimate for an incomplete, heavily used item. Use the condition and resale value guide to decide what to inspect and disclose.
Do not remove a low price just because it is inconvenient. Investigate whether it reflects damage, an auction, missing parts or ordinary variation. Likewise, a high price might include valuable accessories. Exclude a result because it is not comparable or its price is unreliable, not because it weakens your hoped-for profit.
Why recent sales matter
Demand, supply and buyer expectations change. Electronics may depreciate after a new model appears; seasonal clothing may command different prices at different times; collectibles can gain or lose interest. A year-old sale is evidence of what happened then, not necessarily what will happen now.
For frequently traded products, prioritise a recent window and compare similar dates. For infrequently sold items, older evidence may still help, but label the limitation. Avoid combining different seasons or market conditions into one average without looking at the dates.
Also consider the pace of sales. Ten transactions over a long period may provide price references without demonstrating quick turnover. Sold comps help with pricing, but sample size alone does not tell you how soon your listing will sell.
Worked example: median versus average sold prices
The average, or arithmetic mean, adds the prices and divides by the number of sales. The median sorts them and takes the middle price; with an even number, it averages the two middle prices. Neither calculation fixes a poor comparison sample.
Hypothetical example: five apparently similar items sold for £40, £42, £45, £48 and £150. The average is (£40 + £42 + £45 + £48 + £150) ÷ 5 = £65. The median is £45. The £150 result pulls the average well above the cluster.
Investigate that outlier. If it included a valuable accessory bundle, remove it from a standalone-product sample. The remaining average becomes £43.75 and the median becomes (£42 + £45) ÷ 2 = £43.50. If the £150 sale truly is comparable, keep it documented but do not assume you can reproduce it.
Review a sensible lower-to-upper range as well as the midpoint. For a buying decision, test the lower case after costs. These illustrative calculations describe a sample; they are not current market prices or a prediction for your next sale.
What to do when there are too few comps
Broaden the query without abandoning essential product details, check alternative marketplaces and look at older transactions cautiously. The no-sold-comps checklist covers this process in more detail.
Work with lower confidence and a wider potential price range. If your buying decision changes from profitable to unprofitable after a small reduction in assumed selling price, the margin may be too thin for uncertain evidence. A conservative budget protects you better than pretending a single result is precise.
Active listings can provide additional context, but keep them separate from sold prices. Adding unsold asking prices does not increase your sold-comp sample. Use the resale value guide to turn your researched range into a cautious purchase decision.
How FlipWorth uses sold comps
Normal FlipWorth analysis retrieves recent eBay sold evidence through its sold-data provider, then filters results using product relevance and the selected market’s currency. Duplicate listing IDs and publicly flagged best-offer sales are excluded from this sold valuation process. The internal relevance filter uses brand, product category, model and identifying text features; its score is not a statistical probability.
This workflow requires at least three qualifying sold comparables for a sold-based estimate. It uses the price distribution, including a median midpoint, rather than simply averaging every search result. If sold evidence is insufficient, matching active eBay listings may provide a separately labelled lower-confidence fallback, which also needs at least three qualifying listings to yield an estimate.
That minimum is an application rule, not proof of statistically reliable valuation. Automated relevance checks cannot establish every condition, size or completeness detail. The experimental Lucky workflow uses different visual-match and fair-value logic; the standard three-comp rule should not be assumed to describe every Lucky result.
Frequently asked questions
Are three sold comps enough?
They can be enough for a useful initial indication if they closely match. Three is not a guarantee of accuracy, especially when condition, dates or transaction prices are uncertain.
Are twenty sold listings always better than five?
No. Five recent, relevant sales can be more useful than twenty mixed models, accessories or bundles. Filter first, then judge how much evidence remains.
Should I use the average or median?
Inspect both, but the median is less affected by a single extreme price. Explain any large difference and investigate outliers rather than choosing whichever number produces the largest profit.
Can active listings count towards my sold comps?
No. They show asking prices, not completed sales. They may support a fallback scenario, but they cannot establish that those amounts were paid.
About the author
Frank Cheung
Frank Cheung, Founder of FlipWorth
Stop guessing resale prices. Check comparable sales with FlipWorth.
Identify your item and inspect the relevant comparison evidence before committing to a purchase.