What a used laptop is actually worth: buyback payout against eBay resale

Every buyback business runs on one spread: what it pays the seller for a used device, and what the device fetches when it is sold again. Everything else in the operation, from the inspection bench to the listing template, exists to protect that spread. Yet almost nobody publishes it, and the platforms that could measure it, an off-the-shelf storefront bolted to a marketplace account, usually cannot, because the payout and the resale live in two systems that never meet.

We run the platform behind Gadget Salvation, a laptop buyback service. Since June 2026 its ledger has recorded both halves of every device's life: the payout confirmed at inspection, and the eBay sale that closes the loop, with the marketplace fee and the inbound shipping cost attached. This note is what that ledger says. Figures are ratios and medians rather than dollar totals, because the absolute margin is the business's own affair; the shape of it is what is useful to anyone building or running a reverse-commerce operation.

The sample

• 72 devices with both a confirmed payout and a completed resale.
• All resold on eBay between February and September 2026, with 68 of the 72 sold between July 16 and September 22.
• Laptops dominate: 43 mainstream Windows laptops, 24 gaming laptops, 4 MacBooks, 1 phone.
• Payout is the amount actually paid to the seller after inspection, not the online quote. Resale is eBay's line item price after any promotional discount, before fees.

That is a small sample and a short window. Treat the medians as a first reading, not a benchmark. Devices that were bought but had not sold yet by September 22 are not in the set, and that exclusion favours faster-moving stock.

The headline ratio

The median device sold for 2.9 times what the seller was paid. Half of all devices fell between 2.3x and 3.7x. No device in the sample sold for less than its payout, and the worst case was 1.6x, which still left a gross margin of 37 percent before fees.

The aggregate tells the same story from the other side. Across all 72 devices, the payouts added up to 35 percent of the resale prices. The seller gets roughly a third. That is the number to remember.

Where a resale dollar goes

Two of these numbers deserve comment.

The eBay fee is lower than the headline rate. eBay's published final value fee for most categories is above 13 percent. On these devices, the marketplace fee reported by eBay's order API came to 7.4 percent of the line price, and the per-order distribution is tight: most lines were charged 7 percent, with smaller groups at 8 and 9. Laptops fall into a category with a reduced rate, and the seller's store subscription lowers it further. If you are modelling a marketplace channel, use the fee your category and store tier actually pay, not the number on the pricing page.

The inbound label is not a rounding error. Buyback sites advertise free shipping to the seller, and someone pays for it. On this sample the prepaid UPS label came to 5.5 percent of the eventual resale price, which is most of the size of the marketplace fee. It is also a cost incurred before anyone knows what the device is worth. The label is bought when the seller accepts the quote; the resale price is discovered weeks later.

What remains, about 52 cents on the dollar, has to cover the inspection technician's time, any parts and cleaning, outbound shipping where the listing offers it, the cost of the devices that get returned or cancelled, and the business's actual margin. That is a wide band because the ledger stops at the marketplace. It does not yet capture labour or parts per device, which is the obvious next thing to instrument.

How the multiple is distributed

Two thirds of devices land between 2x and 4x. That is the core of the business: a laptop bought at a third of its resale value, listed, and sold.

The tail above 6x is worth understanding, because it is not where the money is. Fourteen of the 72 devices had a payout under $50, and their average multiple was 8.6x, with a maximum of 23x. These are devices the calculator priced near zero, typically because the seller declared them defective, that turned out to have a working screen, a good battery, or a saleable motherboard. A device paid at a token amount and sold for parts money produces a spectacular ratio and a modest gross profit. The median is the honest figure; the average multiple of 4.0x is pulled up by this tail and should not be quoted on its own.

Margin is flat above the first price band

Once the payout clears $100, the payout share of the resale price settles at about 35 percent and stays there. The pricing engine is producing the same margin on a $150 device as on a $600 one. That is a useful check on a calculator: if the ratio drifted with price, it would mean the base values or the condition adjustments were miscalibrated at one end of the range.

The same flatness shows up when the sample is cut other ways. Mainstream Windows laptops and gaming laptops both returned a 65.5 percent gross margin before fees. Devices the seller described as Flawless and devices described as Good both returned 64.4 percent. The condition tiers on the calculator, which apply percentage adjustments to the quote, appear to be tracking the resale market closely enough that a Good device is bought at the same fraction of its eventual price as a Flawless one.

Brand is the one dimension where the sample shows a gap: Dell and Lenovo devices returned 67 to 68 percent, HP devices 58 percent. With 10 HP devices that is a hint, not a finding.

Turnover: the seller is paid, then the device sells

The median device sold four days after the seller was paid, and 69 percent sold within a week. Measured from the day the device arrived at the warehouse instead, the median is about a week and 93 percent had sold within 30 days.

The reason for the short gap is a process decision rather than luck. The eBay listing is created during inspection, from the same record that fixes the payout, so the device is on sale before the seller's payment clears. The payout is not a bet on a future price; by the time it is paid, the listing often already has watchers. A buyback operation that waits to list until after payment carries every device for the full length of its own back office.

The two devices that took longer than 30 days are old stock from the previous platform whose sales happened to be captured after the sync went live; they are in the sample because the match was exact, and they are why the mean of 24 days is not a number to use.

What the marketplace feed adds that the storefront cannot

A conventional storefront records what it sells. A buyback platform has to record what it buys, and then find out, from a system it does not control, what happened to each unit afterwards. The interesting engineering is in that join. Three decisions made this ledger possible.

The payout is frozen at inspection. The device record carries the online quote, the inspected value, and a snapshot of the calculator answers the seller gave. The payout is the inspected value, written once. Nothing downstream is allowed to change it, so the resale can always be compared to the amount actually paid, not to a live price.

Resale is pulled from eBay's order feed and matched by SKU. Every listing carries a custom label that encodes the internal item reference. A scheduled sync reads completed orders from eBay's Fulfillment API, parses each line's SKU back to an item, and records the sale price and date against it. Lines that do not match cleanly, because the SKU was edited by hand, the quantity is not one, or the currency is not dollars, are held for review rather than guessed. Of 212 lines pulled in the period, 77 recorded cleanly, 91 were listings for devices that predate the platform, 28 were cancelled orders, and the rest needed a human.

Fees and inbound shipping are captured per unit. eBay reports a marketplace fee per order; the sync allocates it across the order's lines in proportion to price, so each device carries its own fee. The inbound UPS label cost is written to the order when the shipping API confirms it. Both live on the same row as the payout and the resale, which is what makes a "share of the resale dollar" chart a query rather than a spreadsheet exercise.

One more figure from the feed that a storefront would never see: 28 of the 212 lines, 13 percent, were cancelled before completion, mostly buyer cancellations and unpaid orders. Their refunds came to roughly 90 percent of the cancelled line value. They are not in the 72-device sample, because the device was relisted and sold again, but they are real cost: relisting time, and days of inventory age that the payout-to-sale medians above do not include.

What to take from this

• Model a buyback business on roughly a third. On this sample the seller's payout was 35 percent of the resale price, stable across price bands from $100 up, across laptop types, and across the top two condition tiers.
• Use your real marketplace fee. 7.4 percent here, not the 13 percent headline. The gap is worth more than most conversion optimisations.
• Count the inbound label as cost of goods. At 5.5 percent of resale it is comparable to the marketplace fee, and it is spent before the device's value is known.
• List at inspection, not after payment. It is the single reason the payout-to-sale median is four days.
• Quote medians, never the average multiple. Cheap devices that resell for parts produce ratios of 10x and more, and they distort any mean.
• Instrument the join, not just the sale. The value of the ledger comes from three fields that had to be designed in: a frozen payout, a SKU that survives a round trip through the marketplace, and a per-unit fee allocation.

We will rerun this when the sample is a few hundred devices and see whether the third holds.