Corpus Christi Entrepreneurs  ·  Blog

We Tried It: Finding Our Real Costs in One Afternoon

Prefer to listen? Hear this post read aloud (about 4 minutes).

Years ago I sat down and measured everything we make at our shop. Every ingredient, every container, every label. That spreadsheet became our cost of goods sheet, and in my head I knew exactly where every number lived.

Then life happened. Prices crept up. Recipes changed a little at a time. Portions drifted. The sheet was still there, but I couldn't tell you which numbers were still true.

So last week we tried something. While I put together a batch of product, I talked through every step out loud with our AI assistant, and it updated our cost sheet as I went. Here's how it went and what we found.

What we tried

We picked two of our product cups, the simple kind: a base item, a little liquid to wet it, and a dry coating. Simple products are the best place to start, because a small mistake in an inexpensive recipe hides easily.

How we tested it

  • I weighed what went in before and after. A kitchen scale does most of the work. Tare the bowl, tare the cup, weigh what's left at the end.
  • I said every amount out loud. The assistant wrote it into the recipe, priced it from our own receipts, and asked when something didn't add up.
  • We timed the batch. Labor is a real cost, and it's the one most owners leave out.

The numbers that surprised us

  • Our flavoring guess was eight times too high. The old sheet said two teaspoons per batch. I actually used about a quarter teaspoon. We'd been overestimating that line for months.
  • Our coating was less than we thought, too. I measured out 2.65 ounces, had 0.70 left over, and the batch really used 1.95.
  • The same cup isn't the same product. One of our cups holds about half the base of another, because its topping is thick and takes up room. Same cup, similar price, very different cost. Without weighing, we'd never have known.
  • Time lies, too. The clock said 23 minutes, but the Wi-Fi went out and I took my time at the end. The real working time was about 10. Time your batches on a normal day, not a distracted one.
  • Where it landed: both products came in healthy, and one turned out to be among our best for profit.

The verdict: worth it

One afternoon gave us real costs on two products, corrected two ingredient lines we'd been getting wrong, and told us which item deserves a bigger spot on the menu. It worked because the assistant did the bookkeeping while I did the work. I didn't stop to open a laptop or hunt for a file.

You don't need AI to do this. A notebook, a scale, and your receipts will get you there. The assistant just made it fast enough that I'll actually keep doing it.

What we'd do differently

  • Start with your three best sellers, not your simplest items.
  • Weigh several of the same item, not just one. When we weighed one of our molded products, pieces of the same kind varied by a third of an ounce, all from how full the mold was filled.
  • Set a standard and write it down. We now say "fill flat with the rim" instead of eyeballing it.

Try it yourself

Pick one product this week. Weigh what goes in, check your latest receipt for each ingredient, time one batch, and divide by how many you actually got. If the number surprises you, you're not alone. It surprised us.

If you want to see how a small business can use an AI assistant day to day, that's what we'll be writing about here.

More from the blog