How to use sold comps without fooling yourself

Define the item before searching

Write a compact identity record: manufacturer, exact model or part number, size or capacity, generation, included accessories, cosmetic grade, tested status, and known defects. A family name is often too broad. Two products that look nearly identical may differ in battery, storage, voltage, material, compatibility, or generation enough to change price.

For an auction lot, record what cannot be confirmed. “Powers on” is not the same as fully tested. “Open box” is not the same as unused. Missing chargers, remotes, stands, keys, proprietary trays, or mounting hardware must be reflected in the comparable set or deducted separately.

Choose the channel you will really use

National online sold prices may not predict a local cash sale. Local furniture can avoid packing but face a smaller buyer pool. A niche industrial marketplace may achieve a stronger price but require expertise, longer hold time, or costly freight. Build comps from the channel available to you, then account for that channel’s fees, shipping, return policy, and labor.

If you plan multiple channels, use separate estimates. Do not take the best price from a specialist site and the lowest cost structure from a local sale.

Collect a comparable set—not a highlight reel

A useful starting set contains enough recent sales to reveal a range. Record at least the sale date, item price, buyer-paid shipping, condition, included accessories, sale format, and any visible defect. Exclude active listings: an asking price is not evidence that a buyer agreed to pay it.

Also investigate suspiciously high or low results. A high sale may include extra accessories or be a multipack. A low result may be parts-only, local pickup, an incomplete listing, or a different model. Removing a result because it is inconvenient is cherry-picking; removing it because it is not actually comparable is classification.

Normalize the transaction price

Depending on the marketplace, the visible sold price may exclude shipping paid by the buyer. For valuation, normalize to the total amount the buyer paid for the item and ordinary delivery, then separately calculate the seller’s shipping expense.

Suppose three identical items sold for $70 plus $15 shipping, $82 with free shipping, and $65 plus $20 shipping. Their normalized buyer-paid values are $85, $82, and $85. Treating only the item prices would create a false range from $65 to $82.

Do not confuse normalized value with net proceeds. Fees, promoted listing charges, refunds, payment costs, outbound shipping, and supplies still need to be deducted when estimating profit.

Match condition with explicit adjustments

Adjustments should be deductions from evidence, not guesses added to the price. If a missing accessory costs $18 delivered and introduces return risk, deduct more than $18 or classify the item separately.

Use the median before the average

The arithmetic average can be pulled upward by a rare high sale. Sort normalized comparable values and inspect the median, the middle of the set. For values of $58, $61, $63, $66, $68, $72, and $104, the median is $66 while the average is about $70.29. The $104 result deserves investigation before it influences the bid.

When the comparable set is small, a range is more honest than a precise number. A planning value might use the lower-middle result rather than the average, especially for untested auction inventory.

Worked comparable example

A buyer is evaluating six untested countertop appliances of one model. Nine recent sold records are found. Two are excluded: one is a newer generation and one includes an accessory bundle. The remaining normalized transaction values are $54, $58, $59, $62, $64, $67, and $70.

• Median buyer-paid value: $62
• Conservative condition adjustment for untested inventory: 20%
• Planning sale value: $62 × 80% = $49.60
• Expected sellable rate after testing: 75%
• Sellable-equivalent units: 6 × 75% = 4.5

The model uses fractional sellable-equivalent units as an expected value, not a promise that half an item will sell. For cash planning, round conservatively and retain a separate downside case.

Do not mix sell-through with price

A $62 comp does not mean all six units will sell at $62. Price and sellable rate are separate. So are sell-through and time: an item that eventually sells after eight months may not fit a 60-day inventory plan. Track both the amount and the likely time to sale when the information is available.

For quantity lots, additional supply can also push your realized price below historic single-item comps. Competing against your own listings, repeating identical inventory locally, or holding seasonal goods past their demand window deserves a further adjustment.

Create three planning numbers

1. Base price: a conservative value supported by the central comparable range and matched condition.
2. Downside price: a lower value for weaker condition, slower demand, or an adverse market.
3. Net price: the base or downside price after channel fees, shipping contribution, supplies, returns, and other selling costs.

Enter the net evidence into the appropriate auction calculator, not the most flattering sold result. When the comparable set is weak, lower the bid or require more margin. Uncertainty is a cost even when it does not arrive as an invoice.

A repeatable sold-comps checklist

• Exact identity and generation match
• Comparable tested status and completeness
• Recent completed sales from the intended channel
• Item price and buyer-paid shipping normalized
• Bundles, parts-only results, and false matches classified
• Median and range inspected before the average
• Fees, outbound shipping, returns, and supplies deducted
• Sellable rate and time-to-sale modeled separately

Good comp research rarely produces the highest bid. It produces a number you can explain after the excitement is gone.