How to tell, honestly, whether you've outgrown your spreadsheets and QuickBooks — the specific signals, why they show up once lots, batches, expiry, and audits pile on, and what it costs to keep waiting.
Real help from real food people
Chef Diego runs a real food plant. If this page didn't get you there, tell us — a person reads every message.
After this lesson you can look at your own operation and say, plainly, whether your spreadsheets and QuickBooks still match reality — or whether you've quietly outgrown them. This is not a shopping trip; you don't need to pick anything today, and this lesson recommends nothing. The whole job is to see clearly: to tell a rough week apart from a real wall, and to understand why the wall shows up when it does.
Most growing food operations hit a moment where the tools that got them here stop keeping up. It rarely arrives as one dramatic failure. It shows up as a slow drift — the numbers on the screen and the reality on the floor pulling a little further apart each month, until one day you trust the shelf over the spreadsheet.
That drift isn't a sign you did something wrong. Excel didn't get worse; you didn't get sloppy. The work got bigger and more particular, and a spreadsheet has a fixed shape it was never going to grow past. What follow are the specific, observable signs that you've reached that point. One of them on a bad week is nothing. Several of them, most weeks, is the wall.
The first tell is quiet. You open the inventory tab to check how much of an ingredient you have, and you already know not to fully believe it. Operators reach for the same phrase for the master sheet at this stage: about sixty percent accurate, at best. That's not a measured figure — it's how people talk about a sheet they've stopped trusting. When a count matters, you walk out and look at the shelf, because the shelf is the thing you actually believe.
That's the symptom: the spreadsheet is no longer your single source of truth — the one place everyone treats as authoritative for a given number. Once you check reality against the sheet instead of reading reality from it, the sheet is a rough guess you maintain by hand, not a record you can run on.
Nobody decides to outgrow their system all at once. It happens one tab at a time. The master sheet can't hold lot numbers well, so someone starts a lots tab. Expiry dates don't fit, so there's a separate expiry tracker. Costing lives in its own workbook. Now the truth is scattered across half a dozen files, and it gets messy fast, because no single one holds the whole picture.
The hidden cost is the reconciling. Every number that lives in two places has to be kept in agreement by a person — usually your most capable one. Operators have a name for what the team quietly turns into here: the integration layer, the people who copy numbers from one sheet to the next and settle the differences by hand. It works right up until that person is out sick or moves on.
The third sign you can put a date on, because it hurt. A run stalls because the ingredient the sheet promised isn't there — or it's there but past its date and can't go in. Now you've got people paid to stand around, a rush order at a rush price, and a schedule to rebuild around the gap.
A discrepancy that only lived on paper was an annoyance. One that stops a batch is a real cost, paid because the count you were running on wasn't true. When this has happened more than once, the sheet isn't just inaccurate. It's expensive.
The fourth sign shows up the day someone asks you to prove something. A customer wants every finished lot a suspect ingredient touched. An auditor picks a code off a shelf and asks you to trace it both ways. And your honest answer is: give me until tomorrow.
That pause is the sign. If proving means pulling paper batch sheets, cross-referencing receiving logs, and rebuilding the chain from memory, you don't really have traceability — you have the raw materials to reconstruct it under pressure. The gap between "I can produce this in minutes from what we already wrote down" and "give me a day and a pot of coffee" is exactly what this signals. An earlier course, in the lesson on lot fields versus real traceability, walks through why a stored lot number and an answerable trace are two different things.
The last sign is about the other tool in the stack. QuickBooks is accounting software, and it's good at that — it keeps your books, and in its inventory tiers it tracks how much of an item you have on hand and what it's worth. The strain starts when you ask it to run the plant.
The boundary is worth stating factually, because it's easy to blame the tool for the wrong thing. In QuickBooks Online there's no native way to record a lot number or an expiration date against your items or invoices, and no built-in units of measure — no telling it that a case is twelve jars, or that you buy peppers by the bag but pull them by the pound. (Checked against Intuit's own help documentation in July 2026. QuickBooks Desktop Enterprise's higher tiers add lot and expiration tracking as an inventory add-on, but that's still accounting software reaching toward the job, not a system built around lots and batches. Confirm the current feature list on Intuit's own help pages before you rely on it.)
So the sign isn't that QuickBooks is bad. It's a fine tool doing what it was built for. The sign is what you're forced to do around it: keeping lots and expiry in a side sheet, handling by hand because the books have no concept of it, and reckoning true batch cost across multiple units of measure in yet another spreadsheet. When the books and the plant need different things and one tool is stretched across both, you've found the edge of its lane.
These signs tend to arrive together, in the same season — and that's not a coincidence. A young operation makes a few products, holds little stock, and remembers most of what it needs to. A spreadsheet fits that perfectly.
What changes as you grow is that the work picks up structure: lots you have to trace, batch records linking inputs to outputs, expiry dates that decide what ships first, QA holds that have to stick, audits that expect a clean trail. Each is a shape a grid of cells was never built to hold. Genealogy is a link between records, not a value in a box. FEFO is a rule the system applies, not a column you sort by hand. A hold is a state something stays in until it's released. An audit trail is a history, not a snapshot. Nothing failed — the work simply took on a shape a spreadsheet can't carry, and it took it on all at once, which is why a year of drift suddenly feels like a wall. Reaching this point is a milestone, not a mistake. It means the operation grew.
It's fair to sit with the wall rather than react to it, and plenty of good operators do. What's worth being honest about is that waiting is a choice with a running cost.
Some of it you can see: the stalled runs, the rush freight, the hours your best person spends reconciling sheets instead of making product. Some you can't, because the tools that would show it are the ones straining — margin quietly leaking on batches whose true cost you estimate by hand, stock you overbuy because you don't trust the count. The largest cost is timing. Wait long enough and the decision gets made for you, in a bad week: a failed audit, a recall scare, a retail account you turn down because you can't prove what it needs. Choosing on your own schedule, before a crisis forces it, is worth a great deal — and it's the thing waiting slowly spends. This isn't a reason to panic or to buy tomorrow. It's a reason to name where you stand, so that if you wait, you wait on purpose.
If several of these signs are true most weeks, you've done the honest thing and named the wall. That's the entire job of this lesson — not to fix it, just to see it clearly.
The next thought is almost always the same, and it's a fair objection: fine, but isn't a real system overkill for an operation my size? You've heard the horror stories about heavy software built for companies ten times your size, and you're right to be wary. Whether you're actually stuck between a spreadsheet you've outgrown and a system that's too big is exactly where this course goes next.
Do this in your operation