RowGold
Pricing a used part is a repeatable method, not a guess: identify the exact part, collect truly comparable sold listings, set a conservative price from the cluster, verify the net, and run Best Offer on rules you set before you list.
Get the comp worksheetRowGold helps you research which vehicles and parts may be worth your time before you head into a self-service yard. Scores are modeled estimates — not guaranteed sales or income.
By RowGold · Reviewed October 5, 2026
The most expensive pricing mistake in used auto parts is copying the highest active listing and treating it as market value. It isn't — an active listing is an asking price, one seller's hypothesis. A completed sale is evidence: a buyer paid that number, for that part, in that condition, on that date.
A completed sale is only evidence for your part if the sold item is truly comparable — same part, fitment, condition tier, completeness, and shipping profile. This guide is the method: identify the exact part, collect comparable sales, price conservatively from the cluster, verify the net, and run Best Offer on rules set before you list. For the wider process, see our guide to flipping junkyard parts for profit.
One rule drives every pricing decision here: price from completed sales, never from active listings.
Active listings are survivorship-biased: fairly priced items sell and vanish; overpriced ones sit for months, inflating the average ask. They establish one useful number — the competition count. They never say what buyers pay.
A sale is comparable only if a buyer shopping for your part would have considered it instead: same part number or interchange, generation, position, engine, drivetrain, options, condition tier. Normalize shipping — $95 + $25 and $110 free-shipping are both $120 of buyer cost. Four matching sales beat thirty of "the same part" from the wrong generation.
Three close comps support a preliminary range — list with a wide Best Offer band. One comp is a data point, not a market; zero recent comps is a research condition, not proof of rarity. Five sales at $85–$105 define your market even if one sold for $190. Durable-demand categories are in our guide to the best junkyard parts to sell on eBay.
Pricing fails at the search bar: "2012–2016 sedan headlight" returns a mixed bucket of options, sides, conditions, and assemblies. Identify the exact part first, then search for its sales — see research the yard before you go.
The part number is the primary key of this business. Read it off the part — suffix and revision codes included. Search solds by part number first: it surfaces cross-model sales. Never guess a part number into a title.
Generation breaks matter more than model years: a refresh can change connectors and electronics while the model name stays put.
Position changes fitment and price: driver/passenger mirrors and left/right taillights are different parts with different markets. Record it as buyers search it.
Modules, sensors, and mounts are keyed to engine and drivetrain — record engine code, drivetrain, transmission, and body style. See our fitment-returns guide.
This is where pricing hides: heated vs. non-heated, halogen vs. HID/LED, blind-spot options — each a different part with its own sold history. Price unknown options to the lower option's comps.
Record what's included — brackets, connectors, hardware, intact tab count — and what's missing. Then condition, honestly: tested or untested (and how), cosmetic grade, cracks, corrosion.
Fill these before you search; every comp gets judged against this card.
With the identity card filled, collect sales of that exact part — eBay's sold/completed filter is the default source.
Use a 90-day window — long enough for a sample, short enough to reflect the market. Match condition tier first: tested-to-tested, clean-to-clean. Untested against tested comps? Price below by the market's testing discount, noted explicitly.
Never blend assembly and component sales — different buyers, different medians. An assembly you intend to break down is component inventory: commit to one.
Give every comp one row with identical fields. The two columns beginners skip — shipping charged and why comparable — prevent the most expensive errors.
| Sale date | Part number / fitment | Sale price | Shipping charged | Condition | Included items | Notes | Why comparable |
|---|---|---|---|---|---|---|---|
| Sep 12 (illustrative) | 12345-AB, 2013 sedan LH | $95 | $18 | Tested, good | Assembly + bracket | Options verified | Same PN + options, tested |
Template — fill one row per comp. The example row is illustrative, not market data.
Outliers are guilty until proven informative — verify the explanation across several sales, or exclude it from the midpoint and note it.
Write n and the date span next to every range: "n=7, 90 days, tight cluster" is defensible; "n=1, eight months ago" is a guess wearing a number. Thin sample? Widen the search before widening the price. Still empty? Conservative price with firm floors, or skip. Part categories with durable sold demand are in our guide to the best junkyard parts to sell on eBay.
Set a range, not a number: walk-away floor (Step 4) to list price (the sold cluster). The middle is negotiation room.
Sort comps by normalized total — price plus shipping — and drop unexplained outliers. The midpoint of the rest is your market reference; write it down with the n behind it. A $60–$140 "cluster" with no center means mixed comps — fix identification.
List at or slightly under the midpoint when your condition is equal to or worse than the comps — that's where half the buyers already bought, where velocity lives. Above the midpoint only with a concrete reason: testing the comps lacked, better condition, valuable accessories, thin competing supply.
Tested where comps are untested: price toward the top and state the test. Untested where comps are tested: price below by the market's testing discount. Missing pieces, cut pigtails, broken tabs move the price down and must be disclosed.
The sold cluster says what buyers paid; the active page says what you're competing with — days of supply. Thin page, steady sales: midpoint with patience. Crowded page, slow sales: list under it, sharpen the floors, or skip.
One comp or zero recent comps doesn't mean rare and valuable — it means you can't see the market. Answer with more research and a conservative price with firm floors. If research stays empty, treat the listing as a test — or pass.
What midpoint pricing does NOT mean: not undercutting the cheapest active listing; not using a damaged part's price for a clean one; not anchoring on one high sale; not ignoring shipping.
The list price is a market decision; the net is a business decision. Run the net first — it sets your Best Offer floors.
Your yard cost is the full counter outlay: posted price, core, fees, sales tax, plus your admission share. Cores are sunk cost — no core to return. Work it illustratively: $40 posted, $5 core, 9% tax, plus admission share. Chain-by-chain ticket breakdowns are in our guides to LKQ Pick Your Part pricing and Pull-A-Part pricing.
On eBay, model 13.6% final value fee for Motors Parts & Accessories (no store) plus $0.40 per order — per eBay's fee page, checked Oct 4, 2026. The percentage applies to the buyer's total; the $0.40 applies once per listing. This 13.6% + $0.40 basis is the fee line used in every net calculation on this page.
Decide the shipping model before pricing: buyer-paid calculated shipping keeps fee math clean and comparisons honest. Free shipping folds postage into your price and inflates the fee base. Price the box first — weigh and measure packed, check dimensional weight. When shipping eats the margin, see our eBay vs Facebook Marketplace guide.
Some sales go wrong: fitment claims, lost packages, partial refunds. Budget a reserve — a few percent of revenue, up for untested electronics, down for simple tested parts.
The minimum acceptable net is the smallest proceeds that make the pull-list-pack-ship cycle worth your time — written down before you list. Auto-accept sits at or above it; auto-decline above all-in cost plus fees plus reserve. If the cluster can't support a list above it, the method did its job: pass. Which part types clear this bar is in our guide to the best junkyard parts to sell on eBay.
Net = sold price × (1 − fee rate) − $0.40 − yard outlay − shipping you pay − return reserve
Yard outlay = posted price + core + fees + tax + admission share. The $0.40 applies once per listing.
Best Offer is a negotiation tool, not a discount button. Used right, it converts watchers at pre-approved prices. Used carelessly, it trains every buyer to open low and wait.
The list price anchors every offer: list at the midpoint and offers cluster near it; list far above and they cluster below. The list price isn't your target; the auto-accept floor is.
The auto-accept floor is a price you've already agreed to — derive it from math, not mood. It sits at or above your minimum acceptable net. Set it before listing; don't move it because an offer "felt close."
Below auto-accept, set the automatic-no line: any offer netting below all-in cost plus fees plus reserve. Between the two, a narrow manual-review band — wide bands let floors drift down one "close enough" at a time.
One counter per buyer, moved toward your floor by a fixed step — never a chase to wherever they are. Second offer still below the floor? Decline and move on. Never counter below your auto-accept floor.
Lower on triggers set before listing: X days of views with no offers, new comps below your price, supply doubling while sales stay flat. When one fires, re-run comps first. Lower once, decisively.
End and relist when a price cut can't fix staleness: comps moved, title/photos need rebuilding, or the part's age is the objection. Relist with a fresh midpoint, same floors. Two disciplined midpoint listings with no action means the problem is demand, not presentation.
One part, start to finish, every number shown. Every figure below is illustrative — a teaching example, not RowGold's costs, not your yard's prices. Substitute your own comps, counter prices, and fees.
Illustrative example: driver-side power heated mirror, 2011–2015 mid-size sedan, OEM 12345-AB (illustrative), blind-spot indicator. Good cosmetics, electronics untested; pigtail cut at the harness.
| Sale date | Part number / fitment | Sale price | Shipping | Condition | Included | Notes | Why comparable |
|---|---|---|---|---|---|---|---|
| Sep 12 | 12345-AB, 2013 sedan LH | $95 | $18 | Tested, good | Assembly + bracket | Options verified | Same PN + options, tested |
| Aug 28 | 12345-AC, 2014 sedan LH | $88 | Free | Untested, good | Assembly + bracket | -AC later revision, same fitment | Same fitment family, untested like ours |
| Aug 19 | 12345-AB, 2012 sedan LH | $110 | $15 | Tested, excellent | Assembly + bracket + cap | Clean example, fast sale | Same PN/options; condition better than ours |
| Jul 30 | 12345-AB, 2015 sedan LH | $82 | $20 | Untested, fair | Assembly only, 1 tab cracked | Damage disclosed | Same PN; condition worse — lower bound |
| Jul 11 | 12345-AB, 2011 sedan LH | $97 | $16 | Untested, good | Assembly + bracket | — | Same PN/options; closest comp |
| Jun 22 | 12345-AB, 2013 sedan RH | $105 | $14 | Tested, good | Assembly + bracket | Right side — excluded | Wrong position; shown to demonstrate exclusion |
Illustrative example — all figures invented to teach the method, not market data. n=5 usable comps after excluding the wrong-position sale.
Normalized totals: $113, $88, $125, $102, $113 (right-side sale excluded). The cracked-tab $102 is a lower bound, not the cluster. Cluster $88–$125; midpoint $113. Untested with a cut pigtail, our part lists at or under the midpoint: $109.99, buyer-paid shipping.
Illustrative example: $40 counter + $5 core (sunk) + 9% tax ≈ $49.05 outlay. Illustrative minimum net: $30. Auto-accept ≈ $92; auto-decline ≈ $60 list. Review band $60–$92: one counter toward $92, then decline.
| Line | Illustrative figure |
|---|---|
| List price | $109.99 |
| eBay fee (13.6%) | −$14.96 |
| Per-order fee | −$0.40 |
| Yard outlay (counter + core + tax, illustrative) | −$49.05 |
| Shipping | Buyer-paid |
| Net at list | $45.58 |
| Net at auto-accept ($92) | $30.04 |
Illustrative example — teaching figures only. Fee basis: 13.6% + $0.40 per order (Motors Parts & Accessories, no store).
Bench-test the mirror and the comp set shifts to tested rows. Forty active listings against two weekly sales makes the midpoint a ceiling — list under it or pass. A same-PN sale from another model at $70 re-midpoints the cluster.
Five errors with one root cause: pricing from something other than comparable sold evidence.
The highest active listing is the most visible number and the least informative — usually the stalest: priced high months ago, never sold. Pricing from it anchors your Best Offer floors to a fantasy. Thinking "but there's one listed at…"? Open the sold filter.
A $100 free-shipping sale and an $85 + $20-shipping sale are the same buyer cost — but only the normalized number prices your part. Ignoring the shipping column misprices you in whichever direction the comps leaned.
Condition and completeness are price tiers, not footnotes. Pricing a flawed part from clean comps buys views, lowball offers, and returns. Price the part you have, from comps that match it, and disclose every flaw.
"No longer available" describes supply, not demand. A discontinued part nobody needs is scarcity nobody pays for. The valuable NLA case is specific: dealer-discontinued (verified in the catalog, with a date) plus recent sold evidence. Without the sold evidence, NLA is a reason to research, not a reason to price high.
"Fits 2008–2016" on a part that fits two option packages doesn't expand your market — it expands your return rate. Broad claims attract wrong-car buyers and let you borrow comps your part doesn't serve. Price the exact fitment; list the exact fitment.
RowGold's scores are built on the same sold-data method this guide teaches — automated across the lot, with the uncertainty shown instead of hidden.
A 90-day sold median summarizes selected historical sale evidence: completed sales of comparable parts, aggregated into the center of what buyers recently paid. "Selected" matters — sales are filtered for comparability first, because a median of mixed parts is a median of nothing.
A median without a sample size is a rumor with formatting. RowGold shows the sale count and date range behind every valuation and flags thin or inferred data. Thin-data parts aren't hidden or inflated — they're flagged so you can apply this guide's thin-data rules.
A model summarizes the past; your listing faces the present. Use the valuation to rank what's worth researching, then verify part number, condition, and current comps yourself. The full methodology is at How RowGold Scores Opportunities.
Sold listings. Active listings are asking prices — hypotheses, biased toward the overpriced because fairly priced items sell and disappear. Sold listings are evidence of what buyers actually paid. Use active listings only to count competition (days of supply), never to set the price.
Three close comps can support a preliminary range — same part-number family, same fitment envelope, same condition tier. One comp is a data point, not a market. Zero recent comps is a research condition: widen the search (cross-model interchanges, longer date range), verify NLA status, and list conservatively with firm floors — or pass. A tight cluster of five beats one high outlier every time.
Yes — always normalize to total buyer cost: sale price plus shipping charged. A $95 sale with $25 shipping and a $110 free-shipping sale are both $120 of buyer cost. Compare item prices alone and you will systematically misprice against whichever shipping model the comps used. Match your shipping model to the comps as well.
Treat it as a research problem, not a pricing premium. Widen the search: longer date ranges, the same part number on other models, related part numbers, and the dealer's catalog for NLA verification. If evidence stays thin, price conservatively from the nearest comparable cluster, enable Best Offer with firm floors derived from your net math, and set a review date. No sales is not proof of rarity — it is absence of evidence.
For turn-and-burn inventory, yes — at or slightly under the recent sold-cluster midpoint when your condition is equal to or worse than the comps. That is where half the market already bought, which is where velocity lives. It does not mean undercutting the cheapest active listing, using a damaged part's price for a clean one, or anchoring on one high sale. Price above the midpoint only with a concrete reason: tested, better condition, valuable accessories included, or thin competing supply.
Derive them from net math, not feel. Auto-accept sits at or above your minimum acceptable net — the sale price that, after the 13.6% eBay fee, the $0.40 per-order fee, shipping, and your all-in yard cost, still clears your bar. Auto-decline sits below all-in cost plus fees plus return reserve — anything there is a loss. Between them, a narrow manual-review band: one counter per buyer, moved toward your floor by a fixed step, never below the floor. No universal dollar rule — the same arithmetic on different parts.
Not by itself. NLA describes supply, not demand — a discontinued part nobody needs is scarcity nobody pays for. The valuable case is NLA verified in the dealer's catalog (discontinued, with a date — not merely out of stock) plus recent sold evidence at real prices. "New old stock" on a used part is a contradiction; describe the part's actual condition instead. Without sold evidence, NLA is a research flag, not a price premium.
On triggers, not on a schedule. Set review triggers before you list: a fixed number of days with steady views but no offers, new sold comps printing below your list price, or active supply growing while sales stay flat. When a trigger fires, re-run the comps first — a stale price in a moved market is the usual cause. Lower once, decisively, to the new midpoint. Repeated small cuts bleed search standing and train watchers to wait. If two disciplined listings at the midpoint produce nothing, the problem is demand, not price.
Pricing is one step in the full flip: research the donor, validate demand, confirm fitment, price from sold evidence, move inventory fast. The complete process is in our guide to flipping junkyard parts for profit.
Last reviewed: October 5, 2026. Fee figures are dated in the text and worked-example numbers are illustrative — verify current figures before relying on them.
Join the early-access list and get the comp worksheet — the copy-paste identity card and comp table from this guide — plus the RowGold yard dashboard while it's in beta.
Get the comp worksheet