How Do I Know My Business Is Making Money?
What I found when I finally built something to check
Somebody asks you this at a party and you say yes — because the doors are open, Friday was busy, and there’s money in the account. Twenty-three years in, that’s not a lie. But it isn’t an answer either. It’s a feeling with a number stapled to it. So I finally built something to cross-check the software I was already paying for, and every meaningful thing I found had the same shape: the software was quietly wrong, and only building something else to check it surfaced that. I found at least ten. This is the dated record — June through early September 2026, pulled out of logs rather than out of my head — and by the end the question had changed into a more useful one: how would I know if I weren’t?
I wrote this about my own shop. I’m the same person who runs Keep the Dough — Amore Pizzeria, Boone Village, Zionsville, since 2003. Every number in it came out of my books, not somebody else’s study.
Chris Simone · Owner, Amore Pizzeria · Zionsville, Indiana
Where should I send it?
The PDF opens the moment you submit. Name and email, that’s all I ask for.
Here it is.
Download the PDF (34 pages)The follow-up comes from Chris personally, not a sequence. If you hear from me, it’s because I sat down and wrote to you.
Seven things my own software got wrong.
Every one of these was discoverable by any competent person who sat down and compared two systems that both claimed to be right. The reason nobody had done it isn’t that it’s hard. It’s that it’s relentless — it has to happen every day, on every number, forever, and most carefully on the days you’re busiest.
- Revenue overstated, because customer tips were being counted as sales. A year-over-year number I’d been reading as −3.8% was really −7.0%. The business was doing worse than the dashboard said.
- A month of books never closed at all — the bookkeeping bot died in May and the alert about it died too, silently, for almost a month. Then weeks later the books closed twice, 100 seconds apart, off two different half-finished snapshots. July net income overstated by roughly $17,000.
- Real advertising spend invisible, because a delivery platform nets its marketing fees out before the money ever reaches the bank. There’s no transaction to categorize. Actual July ad spend turned out to be roughly double what the books showed.
- Food-cost tracking completely dead for two months, with no alert of any kind. A model version got retired upstream, the call started failing, and nobody was reading the error. A stale number and a current number look the same on a screen.
- Best-selling menu items undercounted by about a third, because the register had two legitimate doors into the same product and the reporting only ever looked at one. It corrupted every decision downstream of “how much of this do we sell.”
- Labor cost wrong in both directions, six days apart. A terminal that never clocked out turned five-hour lunch shifts into seventeen hours, pushing a true 24.8% up to a false 29.8%. Then the scheduling tool’s cost estimate came in 16% low — it can’t model a salaried employee at all.
- Three separate systems that each claim to know an employee’s pay rate — and disagree with each other. Only payroll was authoritative. Payroll’s the one that actually cuts the check.
Honest scope. This is months, not years. There are no customer testimonials, because I haven’t earned any yet, and there’s no before-and-after profit chart, because that isn’t what this was — it was about measurement accuracy, not a turnaround. Seeing accurately doesn’t make a slow month into a good one. My corrected sales number was worse than my wrong one.