Repair buys speed. Testing buys margin.
Our own repair-ticket data: tested units sell in about half the time of untested ones, and at a far higher margin.
Testing something before we sell it and repairing something before we sell it are not the same decision, and our own numbers say so plainly. Testing raises the price. Repair does not. Repair buys something else entirely: speed.
We pulled the real numbers behind three condition labels: tested, untested, and repaired, sold units only. Two other condition buckets exist in the ledger and are not in this comparison, for reasons below.
Tested, untested, repaired: what these mean here
“Tested” means a unit went through a bench check before it sold: powered on, run, confirmed doing what it is supposed to do, no repair work performed beyond that. “Untested” means it sold as-is, with no bench check behind the condition claim. “Repaired” means the unit has a completed repair ticket attached to it: something was actually fixed, not just checked.
The numbers
| Condition | Median margin | Median days to sell | Units sold (n) |
|---|---|---|---|
| Tested | 209.5% | 30 | 25 |
| Repaired (completed ticket) | 4.1% | 30 | 27 |
| Untested | 11.9% | 58 | 18 |
We are leaving two condition buckets out of this table on purpose.
“New” carries an average margin of 4777.6%, and that number is real but useless here. It is an artifact of how cost gets allocated across a bulk lot purchase, not a comparable price outcome. A handful of near-zero-cost-basis line items inside a shared lot can post a margin that high without telling you anything about pricing strategy.
“Other,” a catch-all condition label rather than a testing state, carries 39 sold units at a 155.0% median margin and a 65-day median sell time. It sits outside this comparison because it does not describe a testing decision the way tested, untested, and repaired do: it is where items land when the condition claim does not fit those three categories, not a fourth point on the same axis.
Worth being honest about the sample sizes too. The tested bucket has 234 units in it total, but only 25 of those have actually sold and closed, which is what the margin and days-to-sell figures are drawn from. Repaired sits at 27 sold, untested at 18. None of these are huge samples on their own, but they are the real, closed outcomes, not a guess, and the pattern across all three is consistent enough that we are comfortable acting on it.
The turn
Repair does not buy margin. Repaired units close at a median margin of 4.1%, which is 7.8 margin points worse than doing nothing and selling the same condition of unit untested. If the goal is a bigger check when the item sells, opening a repair ticket is, on our own numbers, the wrong move more often than not.
Testing is the opposite story. Tested units sell at a 209.5% median margin against 11.9% for untested, a gap wide enough that whether we checked it is doing more work on the final sale price than almost anything else we control.
What repair actually buys
Repaired units sell in a median of 30 days. Untested units take a median of 58. That is the same 28-day improvement tested units get over untested, and repaired units get it too, just without the margin lift that comes with testing.
So repair is a cash-flow decision, not a pricing decision. It moves capital out of inventory and back into the business faster. It does not make the sale bigger.
This makes intuitive sense once you say it out loud, even though it runs against the instinct that fixing something before you sell it should always be the safer, more profitable move. A repair ticket does not change how good the underlying item is compared to a similar unit that was never broken. It changes whether a unit that could not be sold at all becomes sellable. That is a real service, and it is worth doing. It is just a different service from the one a bench check provides.
What this changes
A repair ticket should get opened when the honest reason is turnover, not because we expect a completed repair to raise what the item sells for. If a unit needs to move and testing alone will not get it working, repair earns its keep on speed. If the goal is margin, the bench check is the thing that pays, and the wrench is optional.
We are not going to stop repairing things. Plenty of units in this business would not be sellable at any price without a completed repair ticket behind them, and getting capital back out of a broken unit in 30 days instead of letting it sit is worth something on its own. What changes is how we talk about why we are doing it. Repair is how we keep inventory moving. Testing is how we get paid for it.