A monthly refund total tells you almost nothing. The same number split by set, condition, and price band tells you exactly where to fix your process first.
Ask most sellers what their refund rate is and they'll give you a percentage: refunds divided by orders, for the month. That number is real, but it's also useless for deciding what to do next. It doesn't say whether the refunds are concentrated in one recent set, one condition grade, or one price band. It just says "some amount of money came back."
A dispute rate broken down by set, condition, and price band is a different kind of number. It points at a cause instead of just reporting an outcome.
A lump refund total hides at least three separate questions. Is a particular set driving the disputes, because it's a recent high-demand release where buyers are pickier and grading is harder to get right consistently? Is a particular condition grade the problem, because near-mint and lightly played are being called too loosely at pack time? And is the dollar exposure concentrated in a price band that makes each miss expensive, or spread thin across cheap cards where it barely matters? TCGplayer's own seller reporting shows refunds as one lump total for the period, with no way to slice by any of the three.
Grading near-mint vs. lightly played consistently and documenting condition at pack time both help you fix a grading problem once you know you have one. Finding out you have one, and where, is the step before either of those.
Condition-dispute resolutions typically land as a partial refund scaled to the size of the condition gap. That proportion is the same whether the card is a bulk common or a chase rare, but the dollar cost of the same grading error is not. A 20% refund on a $2 card costs $0.40. The same grading error on a $200 card costs $40, a hundred times the exposure from an identical mistake.
That's the reason a set-and-condition breakdown isn't enough on its own. A high dispute count in your bulk bins might be a rounding error in dollar terms. A handful of disputes concentrated in your $100-plus band, even at a low count, can be the majority of your actual refund cost for the month. Price band is what turns a dispute count into a dollar priority.
Put the three dimensions together and a pattern usually shows up fast. A set that just released, at a specific condition grade, above a specific price point, is a common shape: new inventory, graded under time pressure, on cards expensive enough that a miss actually costs something. Once you can see that combination instead of a single monthly total, the fix is specific: re-grade that set's remaining stock, slow down on that price band, or route high-value listings through a second check before they go live.
Without the breakdown, the same fix gets applied everywhere, which wastes effort on the parts of the catalog that were never the problem.
Hoard's refund tracking keeps disputes attached to the order, the product line, and the buyer, month over month, instead of collapsing everything into a single figure. That's the mechanism for actually running the breakdown above on your own inventory: pull the refund history, group it by set, condition, and price band, and see which combination is doing the damage. TCGplayer's own reporting stops at the lump total; Hoard's refund tracking is what makes the breakdown above possible.
If you'd rather see how your own numbers compare to the wider seller base once that data exists, the TCGplayer refund rate benchmark covers what we're building and why it isn't published yet. And if you're weighing whether to hand dispute handling to TCGplayer Direct entirely instead of fixing the source, is TCGplayer Direct worth it just to avoid disputes? runs through that tradeoff.
It hides three separate questions: whether disputes are concentrated in one set, one condition grade, or one price band. It just reports that some amount of money came back, without pointing at a cause.
TCGplayer's condition-dispute refunds scale in the same percentage steps regardless of price, but the dollar cost doesn't — a 20% refund on a $2 card is $0.40, the same miss on a $200 card is $40. A handful of disputes in a high price band can be most of a month's actual refund cost.
Pull the refund history and group it by set, condition, and price band, keeping each refund attached to the order, product line, and buyer instead of a single collapsed monthly figure.
The fix gets specific — re-grade the remaining stock in a set that's driving disputes, slow down on the price band where misses are expensive, or route high-value listings through a second check before they go live.
Hoard keeps every refund attached to the order and the product line, so you can see exactly where your dispute rate is coming from, not just what it adds up to.
Get started free