No bars, and a line of people
Two things go wrong at a card show table and they go wrong together. The venue’s Wi-Fi stops working, and three people are waiting while you try to remember what you paid for the card in your hand. Almost everything written about dealer software describes a sync feature. This is about what the words have to mean before they mean anything, and both answers work on paper first.
Written for anyone keeping books at a show — a notebook, a spreadsheet, a notes app, any of it. Nothing in sections 1 to 7 needs particular software.
In this guide
- The two things that fail at a table
- What “works offline” has to mean
- The hazard is a duplicate, not a loss
- The three moments cost can be entered
- What one card out of a $600 box cost
- The costs that are not the purchase price
- The stock already sitting there with no cost
- Where the software comes in
- The five questions, answered short
1. The two things that fail at a table
A convention hall is a few hundred people with phones in a building with one uplink, and most of them are checking comps. By eleven in the morning the bars on your phone are a decoration. This is not an edge case and it is not the venue being cheap; it is what every busy show is like.
The second failure has nothing to do with the network. Somebody is holding cash out and there is a line behind them. Any step that takes more than a few seconds does not get done — not badly, not at all. You will tell yourself you will fix it tonight.
Both failures produce the same thing: a gap in the books that you fill in afterwards from memory. And the number you reconstruct on Sunday night is not a measurement, it is a guess that will be treated as a measurement by every report you run for the rest of the year.
The test for any system, software or not: can you record a sale, with the buyer still standing there, with no signal, in under ten seconds? If not, you do not have a bookkeeping system at a show. You have one for the hotel room afterwards, which is a different and much worse thing.
2. What “works offline” has to mean
“Works offline” gets used for several different things, and only one of them is any use to you. A page that loads from cache when you have no signal is not it — being able to read with no network is easy and nearly worthless, because what you need to do at a table is write.
Four questions, and a tool has to answer yes to all four:
| Question | Why it matters |
|---|---|
| Can you write with no signal at all? | Put the phone in airplane mode and record a sale. If anything spins, errors, or silently does nothing, you found out now instead of at the show. |
| Does the write survive the app closing? | Phones kill background tabs when memory is tight, and a hall full of people is when that happens. A write held only in memory is gone and nothing tells you. |
| Does it go up by itself when signal returns? | If sending requires you to remember to press something, it joins the list of things that do not get done. |
| Can you read your own stock while offline? | “What did I pay for this” is the question you need answered at the table, and it is useless if the answer needs the network. |
On paper all four are automatically yes, which is why paper is still at most tables and why it is not a stupid choice. What paper cannot do is answer question four in a second for a card you bought four months ago, which is the trade you are actually making.
3. The hazard is a duplicate, not a loss
Dealers worry about losing a row. In practice losing one is loud — the sale is not there, you notice, you add it. The quiet failure is the opposite: the same sale recorded twice, once in the app that was offline and once in the notebook you fell back to, or once on your phone and once on the tablet your partner was running.
A duplicate does not announce itself. It makes the day look better than it was, it double-counts the card out of stock so your quantity goes wrong, and you find it weeks later when something does not reconcile.
One device of record, per show. If two people are working the table, one of them owns the books and the other hands over slips of paper. Splitting entry across two devices at a busy show is how you get a day that has to be rebuilt from scratch. This costs nothing and prevents the expensive version.
The second rule is about the fallback. If you do drop to paper mid-show, do not re-enter those sales and keep the paper as a backup — enter them and strike each line through as you go. An unstruck line is one you have not entered. A pile of paper where you cannot tell is worse than no paper.
4. The three moments cost can be entered
A card’s cost can be written down at exactly three moments, and most dealers end up at the worst one by default.
| When | What it is like |
|---|---|
| At the purchase | Cheap. You are holding the money and the cards, and the number is in front of you. It takes no recall at all, because it is not a memory yet. |
| At the listing | Workable. You are already handling the card and thinking about price, so cost is a natural second field. But it is already recall, and the lot may be weeks old. |
| At the sale | Terrible. There is a customer in front of you and you are trying to remember a trade from March. Whatever you type is a guess, and a guess entered here is permanent — it becomes the margin on that sale forever. |
The whole discipline is: cost is a purchase-time field, not a sale-time field. If recording a purchase is harder than recording a sale in whatever you use, you will drift into entering cost at the sale, and your margins will be fiction with a decimal point.
This is also the honest reason a lot of dealers have no cost on anything. It is not laziness. It is that buying is the fun part and nobody wants to stop and do data entry in the middle of it. The fix is to make purchase entry shorter than the negotiation you just finished, not to resolve to try harder.
5. What one card out of a $600 box cost
You hand somebody $600 for a long box with maybe four hundred cards in it, five of which are the reason you bought it. What did any one of those cards cost? There is no true answer, only defensible ones. Pick a method and use the same one every time, because comparing two shows is meaningless if you changed the method in between.
| Method | How | When it is defensible |
|---|---|---|
| Equal split | $600 ÷ 400 = $1.50 a card, keys included. | Genuinely uniform bulk. On a box with keys in it, it makes the keys look like a miracle and the bulk look like a disaster. Both are wrong. |
| Value-weighted | Add up what you think each card is worth, then give each one that share of the $600. | The most accurate, and the slowest. Fine for a twenty-card lot, not for four hundred. |
| Keys and remainder | Price the keys at what you would have paid for them on their own — say $480 across five cards — then spread the remaining $120 across the other 395 at about 30¢ each. | What most dealers do, and it is sound: it says what the deal actually was. You bought five cards and got a box thrown in. |
The one rule that is not optional: whatever you allocate has to add back up to what you paid. $480 + $120 = $600. The moment the parts stop summing to the whole, you have invented money, and every profit number downstream inherits it.
A corollary worth saying out loud: it is fine for the remainder cards to be worth more than 30¢ each. Cost basis is what you gave up, not what something is worth. Mixing those two up is how a dealer ends up unable to explain why a good year felt like a bad one.
6. The costs that are not the purchase price
Money leaves your hands for a lot of reasons and only some of it belongs on a card. The split is simple once you say it: a cost belongs on a card if the card carries it when it changes hands. Otherwise it belongs on the show.
| On the card | On the show |
|---|---|
| What you paid for it | The table fee |
| Grading fees, if you sent it away | Gas, parking, the hotel |
| Shipping in, if you bought it online | Food, the badge, the help you paid for the day |
| The case or sleeve, if you count those | Shipping out to a buyer |
Grading is the one that gets left out most often, and it is the one that distorts the most. A $50 card plus $20 to send it away is a $70 card. If you work margin against the $50, the books flatter you by exactly the amount you spent — and the cards you grade are the expensive ones, so the error lands precisely on the decisions that matter.
Table fees go the other way. Spreading a $200 table across the day’s cards makes every card’s margin depend on how busy a Saturday was, which tells you nothing about the cards. Keep it against the show and ask the only question that matters there: did the day clear the table.
7. The stock already sitting there with no cost
Everything above describes a clean start, and you do not have one. There is a shelf of cards from before you kept records properly, and the honest answer for most of them is that you do not know.
Do not try to fix the whole backlog. It is a weekend you will not get back, and most of those cards will never sell.
Fix cost on the way out the door. The moment a card is going to be listed or sold is the moment its cost finally matters, and also the one moment you are already holding it. Put the best number you have on it then, and only then. The backlog drains itself in the order that actually matters, which is the order things sell.
Mark an estimate as an estimate. A guessed cost and a known cost look identical in a total, and six months later nothing tells you which was which. A flag in a notes field is enough — the point is that when a margin looks strange you can ask whether the cost was ever real.
And a zero is worse than a guess. A card with no cost on it does not show up as an unknown; it shows up as pure profit, silently, in every total that card touches. One sealed case entered at zero can carry a whole month’s numbers somewhere they do not belong.
8. Where the software comes in
Sections 1 to 7 hold with a notebook. Here is what Clody does about the same two questions, and what it does not.
Offline. Writes go into a database inside your browser first, then replay to the server when the connection comes back. That covers shows, cards, transactions, expenses, consignors — and the photos, which queue as their own uploads and attach themselves to the right row afterwards. The point is that a sale recorded with no bars is a recorded sale, not a spinner.
The limitation, stated plainly: nothing on screen counts what is still waiting to go up. The design assumes you should not have to think about it, and the price of that assumption is that you cannot check. If you want certainty before leaving a venue, the practical move is the same one it has always been: open the app once with signal and confirm the day’s last sale is there.
Cost goes in at the purchase, because the purchase is the entry. There is no separate step where you go back and fill in cost. Recording a buy either opens a new inventory lot or buys into one you already have; buy into an existing lot and the card’s cost becomes a running weighted average across every lot of it you have ever bought. Selling subtracts from the lot. That is section 4’s discipline built into the shape of the thing rather than left to willpower.
Grading rides into cost. A card’s unit cost is what you paid plus what you spent grading it, everywhere that number appears — the $50 card you spent $20 on is a $70 card on the dashboard and in the charts beside it, not just in one of them.
A foreign show converts once, at your rate. Buy at a show in Seattle with your books in Canadian dollars and the purchase converts at the exchange rate you typed when you created the show. That rate is not decoration for the report: a card outlives any one show, so its cost can only be meaningful in one currency. Standing at the table, the margin hint reads back in the money actually changing hands, because “you would lose $20 here” is only useful in the notes in front of you.
A sale can carry its own cost. If you pin what was actually given up on that particular trade, that wins over the running average. A three-way trade is not an average, and the books should say what happened.
What it does not do. It does not split a bulk lot across cards for you — section 5 is a judgement call about a deal you were in, and software guessing at it would produce confident nonsense. It does not look up market prices or reprice anything; a price in Clody is one you typed. And it is not a marketplace: no checkout, no buyer accounts. Selling still happens across a table or on Facebook, Instagram, Whatnot or eBay. Clody is where it lands afterwards.
9. The five questions, answered short
Sections 1 to 8 are the argument. These are the same answers at the length you would give them across a table — for anyone who arrived at one of these questions rather than at the top of the page.
Does card show bookkeeping work without an internet connection?
It has to. A busy convention hall is a few hundred people sharing one uplink, so the network is unreliable by the middle of the morning. “Works offline” is only useful if you can write with no signal, the write survives the app being closed, it uploads by itself when signal returns, and you can still read your own stock while offline. Being able to load a cached page is not the same thing.
When should a dealer enter the cost of a card?
At the purchase. That is the one moment the number is in front of you instead of being a memory. Entering cost at the listing still works but is already recall; entering it at the sale, with a customer waiting, means typing a guess that becomes that sale’s margin permanently.
How do you work out what one card from a bulk lot cost?
There is no true answer, only defensible ones: an equal split across the count, a value-weighted share, or keys-and-remainder — pricing the few cards you actually bought the lot for at what you would have paid for them alone, then spreading the rest of the money across everything else. Whichever you pick, the parts have to add back up to what you paid.
Do grading fees count as part of a card’s cost?
Yes. A $50 card plus $20 to send it away is a $70 card. Working out margin against the $50 flatters the books by exactly what you spent, and because the cards you grade are the expensive ones, the error lands on the decisions that matter most.
What costs belong on the card and what belongs on the show?
A cost belongs on the card if the card carries it when it changes hands — the purchase price, grading, shipping in. Table fees, gas, parking, food and shipping out belong on the show. Spreading a table fee across the day’s cards makes each card’s margin depend on how busy a Saturday was, which tells you nothing about the cards.
Next: how to track card show sales, which covers the other half of the table — what to write down during a show when there is no time to write anything down.