How to split a lot's cost across Whatnot sales
Pro rata by gross, the rule our engine uses: a $200 lot across $343 and $29 of sales carries $184.41 and $15.59. Worked in full.
How do you split a lot's cost across its sales?
Pro rata by gross. Each sale carries a share of the lot's cost equal to its own gross divided by the lot's total gross, times the cost. A card that brought in a quarter of the money the box produced carries a quarter of the cost per card. The shares always add back up to exactly what you paid, and nothing else gets estimated.
A lot here is one purchase: a hobby box, a case, a collection you bought at a show, a pallet. One price paid, many sales out of it, often across different weeks and different shows. Splitting that one price across those sales is the whole problem, and it is the calculation BreakCount exists to do.
Why pro rata by gross, and not the other two ways
The first alternative is guessing per item: deciding what each card out of the box was "really" worth and pricing the cost against that. It fails for a reason that has nothing to do with effort. The guesses are a second opinion layered on top of the first, nothing checks them, and they do not add back to what you paid unless you force them to. You end up reconciling your own estimates instead of your money.
The second is splitting evenly: a $200 box across 12 sales is $16.67 each. Clean, and wrong in a way that matters to breakers. It says the $343 auto and the $2 base card cost you the same amount to produce, so every big hit looks less profitable than it was and every dollar card looks like a loss. Run a season that way and your best lots look mediocre.
Pro rata by gross uses the one ranking in your data you did not invent: what buyers actually paid. It is not a claim about what each card was worth in the abstract, it is a claim that the sales which brought in the money should carry the cost of producing it. And because the shares are fractions of one total, they reconcile by construction.
A $200 lot across two sales, worked in full
Say a box cost you $200.00 and produced two sales: an auto that hammered at $343.00 and a single that sold for $29.00. The lot's total gross is $372.00.
The auto's share is $200.00 times $343.00 divided by $372.00, which is $184.4086. The single's share is $200.00 times $29.00 divided by $372.00, which is $15.5914. Round each down to the cent and you get $184.40 and $15.59, which sum to $199.99: one cent short of the cost. That last cent goes to the share with the largest fractional part, the auto's .86 against the single's .14.
Final answer: the auto carries $184.41 of cost, the single carries $15.59, and the two add to $200.00 exactly. Those are the numbers our allocator returns for that input, and a unit test pins them so this page and the engine can never drift apart.
Now the per-sale profit is real. The auto grossed $343.00 against $184.41 of cost, before its commission and payment processing fees come out. The single grossed $29.00 against $15.59, and after fees it is close to nothing, which is worth knowing before you list the next fifty like it.
Where does the leftover cent go?
Almost no split divides evenly, so every method has to decide what happens to the cents that fall off the end. Ours uses the largest remainder rule: floor every share to the cent first, then hand the leftover cents out one at a time, biggest fractional part first, ties going to the earliest sale. The number of leftover cents is always smaller than the number of sales, so this terminates and the total is always exact.
This sounds like pedantry until you have a case split across sixty sales. Round each share independently and the shares miss the cost by a few cents; do that on every lot for a year and your net profit is off by an amount you cannot explain to anyone, least of all yourself. The rule exists so that the sum of the parts is the price you paid, always, with no plug figure anywhere.
What if the lot is still selling?
The honest answer: the allocation updates as sales land, and until the lot is finished the shares are provisional. The whole cost is always spread across the sales made so far, so early sales out of a fresh box carry more than they will end up carrying. Sell the $343 auto on night one and it briefly carries the entire $200, because nothing else has sold to share it with. When the $29 single lands, the auto's share drops to $184.41.
That is the correct behaviour, not a rounding artifact, and it means one thing for how you read a part-sold lot: judge it by how much of its cost the sales have recovered so far, not by the net profit on any single sale inside it. BreakCount shows open lots that way on purpose, with a recovery bar against the cost, and re-spreads the whole cost every time a sale is added, removed or the cost itself is corrected.
If you are trying to work out what the rest of the box has to bring in to get you whole, that is the other calculation, and the free break-even calculator runs it: the lowest price a sale can go for and still cover its costs after fees.
Giveaways and refunds carry no share of the cost
Two kinds of rows are deliberately weighted at zero, so they never absorb any of a lot's cost.
A giveaway has a gross of zero. Pro rata by gross would hand it a zero share anyway, and that is right: it did not bring in money, so it cannot carry cost that gets recovered from money. What it does instead is raise everything else. The item left the box without a sale, so the same $200 is spread over a smaller pile of gross, and every remaining sale's share goes up. You paid for the giveaway either way; this is what paying for it looks like in the numbers. The rest of that math is in what giveaways really cost on Whatnot.
A refund is a reversal rather than a sale, and its amount is normally negative. Giving it a share of the cost would mean charging cost against money that came back out, and double counting it against the sale row that is still sitting there. The rule is written on the amount, not on the label: anything at or below zero carries no weight, so refunds get no share and the reversal shows up where it belongs, in the show's totals. An appeal that goes your way can land as a positive amount, and that one is treated as the money it is.
One edge case for completeness: if every row on a lot weighs zero, an all-giveaway promotional box for instance, the cost splits evenly rather than dividing by zero, remainder cents to the earliest rows. The cost is real and it still has to land somewhere.
Doing it yourself, and doing it at volume
All of this is arithmetic you can run in a spreadsheet, and our free profit and loss template gets you the lot-level half of it: each purchase lot's cost against the sales that came out of it, with a recovered or to-breakeven status. Where the spreadsheet route gets heavy is the per-sale half and its upkeep: a lot that sells across four weeks needs re-spreading every week, sales have to be matched to the lot they came from, and a corrected cost means redoing all of it.
Where this sits in the bigger picture is our pillar guide, how to track your profit on Whatnot, which compares every honest way to keep the numbers.
BreakCount does that part. Import the Weekly Order Report, tell it what each purchase lot cost once, and it links the sales, spreads the cost with the rule above, re-spreads it whenever anything changes, and gives each show a verdict with a real net profit. And while any lot cost is still missing, it says so rather than showing you a net computed over a blank.