Skip to content
allanninal.dev/build
Collaborate

Part 5 of 7 · Petty cash tracker series ~5 min read

How the petty cash month closes

Everything so far has been for the people who spend the cash. This post is for the person who has to turn it into bookkeeping, and it rests on one claim: doing the categorisation once, at the end, by one person, is both faster and more accurate than doing it forty times at the point of spend.

Key takeaways

  • Categorising happens once, at month end, by the person who knows the chart of accounts.
  • The bookkeeper gets one screen: every receipt, its photo, and a category dropdown.
  • Previous categorisations pre-fill by vendor, so the second month is much faster than the first.
  • The export is one journal per float per period, and it balances by construction.
  • An uncategorised receipt blocks the export, which is the one hard rule at month end.

Why categorise at the end

The argument for categorising at the point of spend is that the person knows what they bought. That is true and it is not the constraint. The constraints are that they do not know your chart of accounts, they are standing in a shop, and they will pick whichever option is first in the list about a third of the time.

A bookkeeper looking at forty receipts in one sitting has the chart of accounts in their head, can see patterns across the month, and takes about ten minutes. Forty people interrupted individually take forty interruptions and produce a set of categories that has to be corrected anyway. The photo carries everything needed to decide, so nothing is lost by waiting.

How a petty cash month is categorised and exportedThree boxes across the top outside the AWS account. The Bookkeeper, working in one sitting on one screen. The Chart of accounts, holding your own codes. And the Accounting system, which receives a journal. Inside the account, three components. The Month view, showing every receipt with its photograph and amount. Pre-fill, which suggests a category by vendor based on previous months. And the Journal builder, which produces one journal per float per period. Arrows show the bookkeeper supplying categories, the chart of accounts supplying the codes, and a balanced journal going to the accounting system. A note says the pre-fill is why the second month takes four minutes rather than ten.AWS ACCOUNTBookkeeperone sitting, one screenChart of accountsyour own codesAccounting systemreceives a journalMonth viewevery receipt,photo and amountPre-fillby vendor, fromlast monthJournal builderone journal perfloat per periodcategoriesthe codesa balanced journalThe pre-fill is why the second month takes four minutes rather than ten.
Fig 1. The month-end path. The pre-fill by vendor is the part that compounds: after two or three months most receipts arrive already categorised correctly.

The month view

One scrollable list. Each row is a receipt: the photograph as a thumbnail that expands, the amount, the date, the vendor if it was read, who photographed it, and a category dropdown. Rows where the vendor has been categorised before arrive pre-filled and highlighted, which the bookkeeper scans rather than sets.

The pre-fill is by vendor and float together, not by vendor alone. The same supermarket is kitchen supplies from the office tin and site consumables from the van tin, and treating those as the same is how a pre-fill becomes something that has to be double-checked rather than something that saves time.

The journal

Petty cash — Office tin — July 2026

  Kitchen supplies          42.80  Dr
  Postage                   18.60  Dr
  Travel                    31.00  Dr
  Sundry                     9.40  Dr
  Petty cash (float)               101.80  Cr
                           ———  ———
                           101.80  101.80

Top-ups in period are a separate journal: bank Cr, float Dr.
A count adjustment posts to the difference account named in the sheet.

It balances by construction, because the credit side is the sum of the debits rather than a separately computed figure. That sounds obvious and is the difference between a journal that imports cleanly and one that produces a suspense entry every month.

The one hard rule

An uncategorised receipt blocks the export. This is the only place in the entire system where something is blocked rather than flagged, and it earns the exception because the alternative — exporting it to a sundry or suspense code — is how petty cash ends up as a single unhelpful line that grows every month and gets queried at year end.

What the period leaves behind

One month of a petty cash float summarised in five numbersA horizontal row of five boxes. Receipts: thirty-eight photographed. Spent: one hundred and one pounds eighty. Counts: four, all matched. Adjustments: none. Categorising: six minutes. A note says the third and fourth numbers are the control and the fifth is why anybody keeps using it.ONE MONTH OF ONE TINReceipts38 photographedSpent£101.80Counts4, all matchedAdjustments0Categorising6 minutesThe third and fourth numbers are the control. The fifth is why anybody keeps using it.
Fig 2. A month of one float in five numbers. The count and adjustment figures are the assurance; the categorising time is what determines whether the system is still in use next year.
  • Machine learning
  • Management
  • Front-end & mobile

That last number is the one to watch over a year. A month-end that takes six minutes stays done. One that creeps up to forty because the pre-fill is not working or receipts are arriving unread will quietly stop happening, and the first sign will be a tin that has not been counted since March.

Next: what all of this costs to run.

All posts