RowGold
There is no universal list of the best junkyard parts to sell on eBay. There is a repeatable way to evaluate any part — demand, net margin, pull time, fitment, and shipping risk — before you spend a Saturday on it. This guide is that framework.
Get the part-evaluation worksheetRowGold scores modeled profit per minute from 90-day sold data — but a score never replaces verifying the part at the vehicle.
By RowGold · Reviewed October 5, 2026
Forget the listicle. There is no fixed "top 10 parts guaranteed to make money" — a part can carry a high apparent resale price and still be a bad pull because it is slow-selling, option-dependent, damaged, expensive to ship, hard to remove, or impossible to identify correctly.
What decides whether a pull pays is five things evaluated together: demand from sold listings, margin after every cost, pull time against your Saturday, fitment confidence, and shipping risk. This guide is that framework — which categories tend to fit eBay's national buyer pool, which usually don't, and how to turn the evaluation into a pull list.
New to the process? Start with our end-to-end guide to flipping junkyard parts. This page answers one deeper question: which categories are worth evaluating for eBay, and how do you run the evaluation?
Five conditions must hold at once: buyers are actually paying, the margin survives every cost, the pull fits your Saturday, the fitment is certain, the shipping won't eat the profit. The quick version is the table below.
Active listings are asking prices — wishes. Sold and completed listings are evidence of buyers paying. For the exact part, record the 90-day sold median and the sale count behind it: 200 sales is a market; 8 is a rumor with a number. Then check days of supply — active listings ÷ daily sales (90-day sales ÷ 90). Under ~60 days is healthy; over 200, skip. Full method in our pricing guide for used OEM parts.
The counter price is the start, not the cost. All-in = posted price + core + environmental fees + sales tax + your admission share — core deposits are sunk cost with no core to return. Fee basis, stated once: 13.6% eBay Motors Parts & Accessories final value fee (no store) + $0.40 per order, per eBay's fee page, checked Oct 4, 2026. Net on a conservative, below-median sold figure — margin that only exists at the optimistic asking price isn't margin. Full cost build in our used OEM parts pricing guide.
Net = sold price × (1 − 0.136) − $0.40 − counter outlay − shipping you pay
Counter outlay = posted price + core charge + environmental fees + sales tax + your share of admission. Worked examples on this page are ILLUSTRATIVE — hypothetical numbers used to show the arithmetic, never RowGold's data and never any yard's prices.
Your scarcest input is Saturday, not cash. Profit per minute — modeled net ÷ pull minutes — ranks one part against another: a $40 net over 90 minutes of wrestling loses to a $25 net in 10. Access decides the denominator — fastener reach, seized bolts, parts that must come off first.
A high resale price on the wrong option is a return, not revenue. Modules, lights, and mirrors split by option packages — engine, trim, lighting tech, camera, heated — and the wrong variant is not a close match. Confirm the part number off the part plus every option it depends on; photograph the OEM label, connectors, and mounting points.
Three questions before you pull: how big and heavy, how fragile, what does safe packing cost? Over 70 lbs can't go USPS Ground Advantage — it needs UPS/FedEx or freight. Glass, long trim, and big housings need double-boxing or arrive in pieces. If it can't ship cheaply and safely, it needs a local buyer plan first — our eBay vs Facebook Marketplace comparison breaks down which part types belong on which channel.
| Factor | Good sign | Walk-away sign |
|---|---|---|
| Demand | Recent, relevant sold evidence and plausible buyer demand | Numerous active listings with little recent sales evidence |
| Net margin | Conservative net remains worthwhile after every direct cost | Margin exists only if the optimistic asking price sells |
| Pull time | Accessible and removable within the planned time | Complex removal with unknown condition or uncertain resale |
| Fitment | Exact part number and option can be confirmed | "Looks like it fits" or broad year/model assumption |
| Shipping | Can be packed safely at a predictable cost | Fragile or oversized item with little local-sale backup |
Run every candidate through these five checks before you touch a wrench. Any failure means it stays on the donor.
Search eBay's completed/sold listings for the exact part: 90-day sold median plus the sale count behind it. Days of supply = active listings ÷ (90-day sales ÷ 90). Set your own minimum sale count and maximum days of supply.
All-in = posted price + core + environmental fees + sales tax + admission share. Core deposits are sunk cost — you have no core to return. Net on a below-median sold assumption, minus 13.6% + $0.40, minus all-in cost, minus shipping you pay.
ILLUSTRATIVE (hypothetical, not RowGold data): $200 sold, $40 all-in → $200 × (1−0.136) − $0.40 − $40 = $132.40 net. At a 20-minute pull: $132.40 ÷ 20 = $6.62/min.
Identify by the OEM number on the part — never from memory. Record year, make, model, trim, engine, drivetrain, and every dependent option; a different number is a different part. Photograph the label, connectors, and mounting points.
Inspect for corrosion, water/heat damage, cut wires, broken tabs, wear — photograph flaws honestly; they cost less than returns. Test what you can: continuity, bulbs, motors on a battery pack.
Measure before you pull. If it can't pack safely at a predictable cost, it needs a local buyer plan first — over 70 lbs rules out USPS Ground Advantage, and glass or long trim needs double-boxing or a local handoff. Decide the channel before the pull with our channel comparison.
Category guidance with the risks attached — not a promise any category pays.
Control modules, clusters, amplifiers: national reach, low-to-medium pull difficulty, low shipping burden, part-number searchable. The tradeoff — fitment and testing risk are both high. Pull only modules positively identified by the OEM part number on the part, including any strategy-code label, and test when the yard allows. Some are security-paired to the donor; know before you pull. Full method in our guide to avoiding fitment returns.
Window/mirror switches, HVAC heads, stalks: low pull difficulty, low shipping burden, national reach, medium fitment risk. Small, simple to photograph and pack. The risk is connector variants and trim-level option splits, plus wear on touch surfaces — photograph the connector and part number on every unit. See our fitment-returns guide.
Real national demand, manageable shipping burden — but high fitment risk. Heated, power-fold, signal, blind-spot, and camera variants make "looks the same" dangerous, and yard cars often carry damaged glass, housings, or tabs. Verify the side and every option. Our guide to avoiding fitment returns walks through option verification.
Strong demand on popular platforms, low-to-medium pull difficulty — but medium-to-high shipping burden (fragile, bulky, costly to pack) and high fitment risk: halogen/HID/LED, adaptive or not, left/right, ballast inclusion. Verify every option, confirm intact tabs, check the lens, photograph everything — the full method is in our fitment-returns guide.
The beginner-friendly corner: national reach, low pull difficulty, low shipping burden, low-to-medium fitment risk. Caps, lug hardware, TPMS components, emblems, small trim — cheap to ship, easy to photograph. Verify exact style, finish, and size; a "close" cap is a return.
Known-failure parts on popular vehicles — the parts owner forums complain about by name. That knowledge is a shopping list, but only with platform expertise: know which revision fails and which superseded part number fixes it. The strongest version is the NLA part — dealer-discontinued, so used is the only market. Verify "discontinued" (not "not in stock") in the dealer's parts catalog with a date.
A planning screen for comparing categories — not a price guide. Each row shows where a category's risks concentrate; the five-part screen still decides every pull.
| Category | Typical buyer reach | Pull difficulty | Shipping burden | Fitment risk | Condition/testing risk | Beginner suitability |
|---|---|---|---|---|---|---|
| Small OEM modules | National | Low to medium | Low | High | High | Low to medium |
| Switches/controls | National | Low | Low | Medium | Medium | Medium |
| Mirrors | National/local | Low to medium | Medium | High | Medium | Medium |
| Headlights/taillights | National/local | Low to medium | Medium to high | High | Medium | Medium |
| Center caps | National | Low | Low | Low to medium | Low | High |
| Wheels | Local/national | Medium | High | Medium | Medium | Medium |
| Seats/consoles | Local | Medium | High | Medium | Condition-dependent | Low to medium |
| Body panels/glass | Mostly local | Medium to high | Very high | High | High | Low |
Note: This table is a planning screen, not a valuation model. It does not represent all years, makes, or markets.
National reach, easy pull, cheap to ship — but high fitment and testing risk, hence low-to-medium beginner suitability. The pull is easy; the identification isn't. Only take modules positively identified by part number with sold evidence behind that exact number.
National or local reach, low-to-medium pull difficulty, medium-to-high shipping burden — fragile, bulky, costly to pack. High fitment risk: side, lighting tech, adaptive systems, ballast. Verify every option; photograph all tabs.
Like lights: national or local reach, low-to-medium pull difficulty, medium shipping burden, high fitment risk. Heated, power-fold, signal, blind-spot, camera — identical-looking mirrors can fit nothing in common. Confirm side and options.
Two categories in one. Caps, lug hardware, TPMS: national, easy, cheap to ship, beginner-friendly. Full wheels: medium pull difficulty, high shipping burden, usually local. TPMS sensors are shippable — mind age and battery life in the description.
Mostly local. Medium pull difficulty, high shipping burden, and condition is everything — wear, stains, tears, smells decide value more than the part number. Consoles and small trim can ship; full seats almost never do.
Small items — sensors, relays, coils, throttle bodies — pull easy, ship cheap, sell nationally with verified numbers. Heavy accessories — alternators, starters, compressors — carry core charges, real weight, and no yard testing. Small stuff with confidence; heavy stuff local-first.
Mostly local: medium-to-high pull difficulty, very high shipping burden, high fitment and condition risk, low beginner suitability. Glass and panels do sell — to someone who can pick them up; our channel comparison covers the local-first math. Pull only with a local buyer plan.
Where beginners lose money. Each of these looks like value at the yard and turns into cost at home.
The classic trap. A module you can't identify by part number can't be listed accurately, matched to a buyer, or returned cheaply. Label gone, code unreadable, application unconfirmed — that's a walk-away, not a discount.
A light or mirror with an unverified option package is a return in a box. Heated vs unheated, camera vs none, halogen vs LED — invisible in a yard glance, decisive in a listing. Verify the option or leave it.
Weight is only a problem without a local buyer. A transmission with a real local market is fine; the same part with only national eBay demand is a freight problem in a profit costume. Check local demand before the pull — our channel comparison covers the local-first call.
Active listings piling up with no sales means the price is wrong or demand is gone. Days of supply over ~200: walk away regardless of margin. Crowded pages clear when someone cuts the price — usually not you.
A complete headlight assembly, its ballast, and its igniter cannot automatically be treated as separate sellable profit on the same vehicle — harvest the loaded assembly and the components inside aren't separate inventory. Decide the extraction path first: sell loaded, or part out. Count only what can be harvested without consuming the same item twice. RowGold ranks the best non-overlapping set per car — the full logic is in how RowGold scores opportunities.
Airbags, seat belts, brake hydraulics, steering, fuel-system parts. The failure mode isn't a return — it's liability you can't price or insure away as a casual seller. The margin is never worth the downside. Leave them.
The framework only pays as a list you carry into the yard. One line per part: vehicle, part, what to verify, net, pull time, channel, decision.
Buyers shop by vehicle and platform, so research vehicle by vehicle: which donors are on the lot, and which parts on each have failure demand, sold evidence, and a clean pull path? Three well-researched cars usually hold most of a Saturday's profit — the full workflow is in our yard research guide.
Depth beats breadth. Cap at three to five vehicles per trip: sold comps by part number, days of supply, all-in cost, fitment splits, pull time. Twenty half-researched targets make a trunk of guesses; four fully researched ones make a list.
No universal number exists — your travel, admission, home labor, storage, and risk tolerance differ from everyone else's, so anyone's fixed minimum is a guess about your life. Set your own two thresholds and write them down: minimum conservative net per part, maximum pull time per part. Below the floor or above the ceiling, it doesn't make the list.
Every line gets a verification task: the part number to confirm, the option code to check, the test to run, the flaw to photograph. This stops "looks right" from becoming a return — if you can't complete the verification at the vehicle, the part stays.
Tag every candidate before you leave home. eBay: small, shippable, part-number-confirmed, national demand. Facebook/local: heavy, fragile, oversized, verified local demand. Walk-away: fails any screen. One part, one channel, decided in advance.
RowGold ranks candidate pulls by modeled net ÷ estimated pull time, from 90-day sold data and posted yard pricing — compare a 10-minute switch against a 45-minute assembly before you pay admission. Thin or missing data gets flagged low-confidence, never hidden. Full inputs are in our scoring methodology.
No model can see the car. You still confirm the part is present, the exact part number and option, its condition, and that nothing about the donor disqualifies it. The score tells you where to look first; your eyes close the deal.
A score built on 200 sales is a different animal from one built on 8. Treat low-confidence scores as research leads, not pull orders: verify harder, price conservatively, size the bet to what you can afford to be wrong about. Missing sold data doesn't mean skip — but the model can't vouch for demand, and neither should you.
The framework is the whole game: demand from sold data, net after every cost, pull time against your own thresholds, fitment you can prove, shipping you can survive. Categories change; the evaluation doesn't. For the end-to-end process — research, pull, list, ship — see our complete guide to flipping junkyard parts for profit.
Center caps, interior switches and controls, and small exterior trim: low pull difficulty, low shipping burden, national buyer reach, low-to-medium fitment risk. Verify the exact style, part number, and finish before listing. Skip anything option-dependent, untested, or hard to identify until you've run the five-part screen a few times.
Body panels, glass, seats, dashboards, and full wheel sets: oversized, fragile, or heavy enough to price out of standard parcel shipping. Over 70 lbs can't go USPS Ground Advantage — it needs UPS/FedEx or freight, which changes the math. Usually local-pickup parts; confirm local demand before you pull.
They can be: national demand, compact, part-number searchable, cheap to ship. But they carry the highest fitment and testing risk of any category — only pull modules positively identified by the OEM part number on the part, and test when possible or price the risk honestly.
Check eBay's completed/sold listings for the exact part: record the 90-day sold median and the sale count behind it. Then compute days of supply — active listings divided by daily sales (90-day sales ÷ 90). Under ~60 days is healthy turnover; over 200 is a crowded market to skip regardless of apparent margin. No recent sold evidence means no proven demand.
Decide the best extraction path before you pull, and count only what can be harvested without consuming the same item twice. A headlight assembly, its ballast, and its igniter can't automatically be three separate profits on one vehicle — harvesting the loaded assembly means the components inside aren't separate inventory. Rank the best non-overlapping set per car.
Often, with verified options. Headlights have real national demand, but fitment risk is high: halogen vs HID/LED, adaptive vs non-adaptive, left vs right, ballast inclusion, intact mounting tabs. Verify every option; photograph all tabs and the lens. Shipping is a medium-to-high burden — fragile and bulky — so pack cost and damage risk belong in your net math.
Usually locally. Wheels have recognizable demand but a high shipping burden — heavy, bulky, expensive to pack safely — which is why the scorecard marks them local/national. Center caps, lug hardware, and TPMS components are the national eBay side: small, light, part-number searchable. Sell full wheel sets via local pickup unless shipping quotes still leave real margin.
The exact OEM part number on the part (never from memory), the strategy-code or calibration label if present, connector condition, and any option the module depends on — engine, transmission, trim, audio system. Check for water, heat, or corrosion damage, and test it when the yard allows. Can't positively identify it? Leave it.
Reviewed October 5, 2026. Fee, tax, and yard-price figures change — verify current figures before relying on them.
Get the part-evaluation worksheet and join the early-access list for the RowGold yard dashboard — free while we're in beta.
Get the part-evaluation worksheet