Product Businesses 2026-10-30 9 min read

Your Books Don't Start in QuickBooks

You know roughly what a bag costs to make. Your books don't — which is why they can't tell you whether the Thursday market is worth the booth fee.

Ask a granola maker what a bag costs to produce and you will usually get a decent answer — within a dollar or so, from memory, based on what the last ingredient order came to.

Ask the same question of their QuickBooks file and you get nothing at all.

That gap is not a bookkeeping failure. It is what happens when the only thing feeding your books is a bank feed.

What it's actually costing you

Here is the decision it blinds.

You run a Saturday market and a Thursday market. The Thursday one is smaller, but it fills a slow afternoon and the booth fee is cheaper, so it feels worth keeping. You have run it for two seasons on that feeling.

Your P&L cannot tell you whether that feeling is right. It knows the deposits. It does not know which market they came from, how many bags each one moved, what those bags cost to make, or what you paid the two people who worked the table. Every one of those numbers existed on the day. None of them survived the trip to the ledger.

So the question stays a hunch. Same with whether the wholesale account that looks like growth is priced above cost, which of six flavors earns its shelf space, and whether the four-pack should go up a dollar. Those are not accounting questions. They are the questions you actually run the business on, and the books are supposed to answer them.

Why your books can't tell you

Your bank feed is a record of money moving. That is a genuinely useful record, and it is not the same thing as a record of your business.

A Square payout is the clearest case. Saturday might have been $1,340 across ninety-odd transactions — some card, some cash, two comps and one refund. What lands in the bank on Tuesday is a single deposit, net of fees, possibly batched with Sunday. Categorize that deposit as income and your books now say you made one sale on Tuesday, for an amount that matches none of your receipts.

Repeat that for a season and the P&L is arithmetically fine and practically useless. It will tell you roughly how much money came in. It will not tell you which market was worth the booth fee.

The same thing happens on the cost side. An ingredient order is a payment to a supplier. By the time it reaches your books it is "Supplies — $412," which is true, and tells you nothing about how much of it ended up in the apple cinnamon.

Underneath both is one rule:

Your books have to be at least as detailed as the decisions you need to make — and the detail has to be captured when it is created, not reconstructed afterwards.

Saturday's split between cash and card, comps and refunds, is knowable on Saturday. By March it is gone. No amount of bookkeeping skill applied in March brings it back, because it was never written down. Same for how many pounds of oats went into a specific batch, or which market a case of jars left from.

That is also the honest test of whether any of this applies to you. If the only decision you make from your numbers is how much to set aside for taxes, one blended P&L is genuinely sufficient. Nothing is wrong with your books and you don't need the rest of this article.

What one Saturday looks like when you catch it

Here is the same market day with the detail kept. Real-shaped numbers for a maker selling a $14 bag.

Saturday, Ballard — 96 bags

Gross sales $1,340
Comps and discounts −$24
Refunds −$18
Net sales $1,298
Card processing fees −$36
Cost of goods (96 bags @ $4.10) −$394
Booth fee −$65
Crew (2 people × 6 hrs @ $22) −$264
Contribution $539

The deposit that shows up Tuesday is $1,262 — and that is the only one of those thirteen numbers a bank feed would have given you.

Now the Thursday market, same treatment:

Thursday — 31 bags

Net sales $410
Card processing fees −$11
Cost of goods (31 bags @ $4.10) −$127
Booth fee −$55
Crew (1 person × 5 hrs @ $22) −$110
Contribution $107

$539 against $107. Thursday still makes money, so it never looked like a problem. But it returns a hundred dollars for a crew shift and a full day of yours, and once that is on paper it is a different conversation — drop it, shorten it, raise prices, or send one person instead of working it yourself.

That is the whole argument. Not that your books are wrong. That with four more pieces of information per market day, they start answering the question you were going to decide anyway.

Note the one number that does no work in a bank feed and all the work here: $4.10 a bag. That comes from a batch log, and nothing else can produce it.

Five places to catch the detail

None of this needs software you don't already have. Roughly in order of effort:

1. A monthly sales summary, posted as one entry. Pull the month's sales report out of Square — gross sales, discounts, refunds, fees, tax collected — and post it as a single journal entry. Deposits then reconcile against your sales instead of standing in for them. This one change is the difference between books that approximate your revenue and books that record it.

2. Channel tagging at the source. Square lets you set up a location or device per selling point: one per regular market, one for online, one for the farm stand. Tag it once at setup and every future sale sorts itself. Done at the source it costs nothing per month. Done afterwards it can't be done at all.

3. A batch log. One line per batch: date, flavor, yield in units, ingredients used, hours worked. A notebook or a spreadsheet is fine. This is the only place a cost per unit can come from. It is also what turns a precursor — a mix or a reduction that is itself materials plus labor — into a number that rolls into the batches using it.

4. A month-end count sheet. A printed list of your SKUs with a blank column. Walk the shelves once a month, write the numbers, hand it over. That becomes one inventory adjustment and a real cost of goods figure for the month, instead of a guess trued up next January.

5. A wholesale invoice register. If you invoice accounts on terms, keep a running list of what went out, when it's due, and what's been paid. Summarized monthly, that is a receivables figure you can chase.

Each one is a small habit at the point where the information exists. Together they are the difference between books that reconcile and books that answer questions.

The details — or hand this to your bookkeeper

Two things usually come up at this point, and both are worth getting right before anyone spends money.

Upgrading QuickBooks will not do this for you. There is a persistent idea that if the books aren't answering your questions, the fix is a bigger accounting subscription. Usually it isn't. QuickBooks Online is a general ledger, and it is very good at being one — accounts, reconciliations, balances, and the record a tax return is built from. What it does not do is watch you work.

Concretely: QuickBooks Online's inventory tracking, available a tier up from the entry plan, counts finished goods. It does not hold a recipe. There is no bill of materials in the Online product — that feature lives in QuickBooks Desktop — so nothing tells it that one batch consumed eleven pounds of oats, two quarts of maple and three hours of labor. Item bundles come close enough to be misleading: they group products for sale, but they do not draw down the components underneath. Upgrade hoping to learn what a batch costs and you will get a count of bags on the shelf and an unchanged answer to your actual question.

That is not a knock on the software. A ledger records what you tell it. The work is in deciding what to tell it, and the $4.10 has to be worked out where the batch is made, then posted in.

Where these numbers land in the ledger. The monthly sales entry debits a clearing account for the gross and credits sales by channel, with separate lines for discounts, refunds and processing fees; the payout then clears against it, which is what makes the bank reconciliation meaningful instead of circular. Batch costs accumulate to work in process and release to cost of goods as units are sold, with the month-end count correcting the balance. Booth fees and market wages are direct costs of that channel, not general overhead — put them in overhead and the per-market contribution above dissolves back into an average.

If you have a bookkeeper, this paragraph is the one to send them. It is an afternoon of setup, not a system change.

What it gets you

With those habits running, an ordinary monthly close produces things a blended P&L can't:

  • Revenue and cost of goods split by channel, so contribution per market is a number instead of a feeling
  • Cost per unit from actual batches, so a price increase is a calculation
  • A real cost of goods every month, rather than one year-end adjustment
  • A year-end that is a review rather than a reconstruction

None of it requires a bigger subscription. It requires the detail to be caught on its way past.


If you make and sell a physical product in Washington and your books aren't answering these questions, that's the work we do. See bookkeeping for product businesses for how an engagement is structured, or book a fit call — if you're not sure whether you're a fit, that's a good conversation to have anyway.

Need Personalized Guidance?

We can help you apply these strategies to your specific situation.