How often should I count my card inventory?

Count the case and the show box after every show, since those moved the most. Rotate through the rest one slice a week. With six boxes on rotation, everything gets counted roughly every six weeks and your fastest-moving stock gets counted every show.

Do I need to close down to do a proper count?

No. That is the point of counting in slices. You only freeze the one box you are auditing, and only for as long as the count takes. Full shutdowns are why most dealers never count at all.

What do I do with cards that are out for grading?

Give them a status rather than removing them from your records. They are still your capital, still part of your inventory value, and they will read as missing in every count until something marks where they went.

Should I just adjust the numbers to match what I counted?

Adjust the record, but figure out the cause first. An unlogged sale needs to be logged at its real price so your margin stays accurate. Overwriting the count hides a leak that will show up again in the next audit.

Counting Inventory Without Shutting Down

Nobody wants to give up a Saturday to counting cards. But a stock list that quietly stopped matching the case costs more than the count ever would.

You know the moment. A buyer asks for a card your list says you have, you dig through the box twice, and it is not there. Maybe it sold at the last show and never got logged. Maybe it went out in a trade, or it is in an envelope waiting on grading. You cannot remember, so you apologize, and the sale walks to the table across the aisle.

Most dealers answer that by promising themselves a full count after the season. It never happens, because a full count means shutting down for a day and touching every card you own. The fix is not a bigger count. It is a smaller one, run often enough that the list never gets far from the truth.

Where the drift actually comes from

Inventory does not break all at once. It slips in ones and twos, on the days you were busiest and least able to stop and write something down. Before you count anything, know what you are counting against.

Sales that never got logged. Three deep in a line, cash changes hands, the card goes in a sleeve, and the record never catches up.
Trades. Two cards leave and four arrive, and the four are easier to remember than the two.
Cards pulled for grading. Still yours, still worth money, no longer in the box. Unflagged, every count reads them as missing.
Buys off the floor. A collection comes in on Sunday afternoon and goes into a box "for now." Now lasts a month.
Bundles and lots. You sell six cards as one deal at one price, and the record shows one line where six should have closed.
Consignment. Someone else's stock in your case, or yours in theirs, with nothing marking which is which.

Almost every discrepancy traces back to one of those six. That matters, because it means a count is not really a hunt for lost cards. It is a check on which of your habits is leaking.

Count in slices, not in one sitting

A rolling count treats your stock as a set of slices and audits one at a time on a schedule you can actually keep. Nothing closes, you never lose a selling day, and because each slice comes back around, an error has a short life before someone finds it.

Pick slices that match how you physically store cards, not how you think about them. Six boxes is six slices. A case organized by set counts one set at a time. The rule is simple: a slice is small enough to finish in one sitting, and it has a physical boundary you can point at.

A cadence that fits a show calendar

Tie the count to the rhythm you already have instead of inventing a new one.

After every show: count the case and the show box. These moved the most and are the most likely to be wrong.
Weekly, one slice: whichever box has gone longest without a count. Twenty minutes on a weeknight.
Before a big show: count the high-value slice, whatever you keep the chase cards in. Those are the ones a buyer will ask for by name.

Six boxes on a weekly rotation means every card gets looked at roughly every six weeks, and your highest-turnover stock gets looked at every show. That beats an annual count for a fraction of the pain.

How to run a slice count

The whole thing is four steps, and the order matters more than the speed.

Freeze the slice. Do not sell out of the box you are counting while you count it. If you are counting between shows this is free.
Count the cards first, on their own. Record what is in the box before you look at what the list says. If you read the list first, you will find what you expect to find. Every counter does. Counting blind is the one thing that makes a count worth running.
Then compare. Mark every line that does not agree, in both directions. Cards in the box the list does not have are as much of a problem as the reverse.
Do not fix anything mid-count. Write the differences down and keep going. Stopping to investigate turns a twenty minute count into a two hour one, and you will quit halfway with the slice half done.

If you are counting against an app rather than a spreadsheet, scanning beats typing and is less prone to a slip of the finger. CardOps recognizes a card on the device against a catalog of more than 20,000 Pokemon cards, so a slice count is mostly pointing the camera and moving on. It works with no signal, which matters if your storage is in a basement or the hall has bad reception.

Reconcile by cause, not by number

Now you have a list of differences. The temptation is to set every number to what you just counted and call it done. Resist it. Adjusting the number hides the reason, and the reason will happen again next month. Sort each discrepancy into a bucket and handle the bucket, not the line.

What you found	Likely cause	What to do
Card on the list, not in the box	Unlogged sale or trade	Check that day's sales. If it sold, log it at the real price so your margin stays honest. Write it off as lost only when nothing explains it.
Card in the box, not on the list	Buy that never got entered	Add it at the cost you actually paid. Invented cost basis is worse than none.
Right card, wrong condition	Graded on intake, sleeved later	Correct the condition and reprice. Condition moves the number more than most dealers allow for.
Missing, no explanation	Out for grading, consignment, or gone	Give it a status rather than deleting it. A card in a grading envelope is still capital.
Same card twice	Double entry during a busy intake	Merge, keeping the earlier cost basis if the two disagree.

When you are done, look at which bucket was biggest. That is your real finding. A pile of unlogged sales says the problem is at the table, not in the box, and no amount of counting will fix it.

Close the leak at the source

A count tells you where you are losing track. Fixing that is what makes the next count shorter.

Log the sale while the buyer is still standing there. This is the whole game. A sale you record ten seconds after it happens is accurate. One you reconstruct on Sunday night is a guess. Show Mode in CardOps exists for exactly that: sell, trade, and buy from the phone in seconds while the line keeps moving, with the show's profit totaling as you go.
Enter buys before the cards go in a box. The moment a collection lands in an unlabeled box for now, it is off the books until a count finds it.
Give trades both halves. Log what left and what arrived in the same sitting. Trades are the most common source of one-sided records.
Flag anything that leaves the building. Grading, consignment, a card lent out. Status, not deletion.
Keep one intake spot. One tray where everything new lands and nothing leaves until it is entered.

None of that takes discipline in the abstract. It takes the logging step being fast enough that you do it in the moment instead of promising to do it later. That is the only test worth applying to any system you use.

Keep a record of the count itself

Write down what you counted, when, and what you found. A date, the slice, the number of discrepancies, and the adjustments you made. Two lines is plenty.

That log earns its keep three ways. It shows whether drift is getting better or worse, which is the only way to know if a habit change worked. It tells you which slice is due. And if you are ever asked to explain your numbers, by an accountant at tax time or an insurer after a bad day, a dated series of counts is the difference between a defensible figure and your word for it. A card you adjusted out with a reason and a date is a record. One that silently disappeared is a hole.

Once your counts land clean, the number on your dashboard stops being an estimate. Total inventory value, cost basis, per-show margin: all of it rests on the count being right.

Start this week

Pick your worst box. Not the biggest one, the one you trust least. Count it blind, compare, sort the differences, and fix the habit behind the biggest bucket. That is one evening, and it will tell you more than a year-end count you never quite get to.