Sooner or later, every print shop sorts its customers. It doesn’t happen formally, and usually not on paper. It happens in the schedule, in who gets called back first, in which job the shop will work a weekend for. Almost everywhere, the sorting runs on one number, which is how much the customer bills.

That number is easy to get. Every system reports it. It’s on the year-end summary and it’s in the owner’s head. It’s also the one number that can’t answer the question the owner is actually asking, which is whether the account is worth what the shop gives up for it.

Why Sorting Customers by Sales Hides the Answer

The advice every owner has heard is to take care of the top twenty percent. What that advice never says is to subtract.

Nobody’s version of it asks what the biggest account costs to keep. The account most likely to be quietly unprofitable is the one sitting at the top of the revenue list, because size is exactly what buys the accommodations nobody charges for. Somebody works the weekend, the rush shipping gets absorbed rather than billed, and three other jobs move to make room. None of that reaches the estimate, and all of it comes out of the same pocket.

Meanwhile the steady account that orders every few weeks and never asks for anything gets filed as a house account, which is the polite way of saying nobody needs to think about it. It’s a strange system. The customer who causes the least trouble earns the least attention, and the reward for being easy to work with is being forgotten.

One Prompt That Puts a Number on One Job

Start smaller than the whole customer list. Start with one job, and use the AI tool of your choice, and ask your AI:

We just ran a big job for our biggest customer. I know what it billed and what my estimate said we would make, but we also worked a weekend, rushed some shipping, and reprinted part of the run. Help me figure out what that job actually earned us.

What matters is what it doesn’t ask for. It needs no export, no report, nothing that has to come out of the accounting system. It runs on three numbers an owner already knows, because they lived the job.

That prompt is the start, and only the start. It answers one job. It’ll show the gap between the profit the estimate promised and the profit that survived the week, and that’s genuinely useful, but it stops there. It can’t tell you whether the same thing happened last year, and it can’t compare that account to any other.

The upgraded prompt is where that changes. It takes two whole accounts across three years, the biggest by sales and one quiet steady one, and works out what each one actually earned. That’s the version that can reverse a ranking. The full workflow goes further again, sorting an entire customer list and flagging the accounts whose ordering has shifted before the change reaches the profit and loss.

Compare Two Customers This Week

Run the simple prompt on the last job that made the shop work a Saturday. It takes ten minutes, and the number will either reassure you or it won’t.

Then use the form on this page to request the upgraded prompt. It arrives in your inbox, and it’s the one that compares two accounts across three years and tells you which has actually earned more. Pick the customer everyone can name, and one nobody ever mentions.

If the comparison surprises you, the AI For Printers Workshop is where the full workflow runs live on a whole customer list. It’s called What Your Biggest Customer Really Costs You, and it’s free.