How an RV Repair Business Sorted Two Months of Books in One Evening

An RV repair owner in Georgia signed up, worked through two months of backlog in a single evening, and downloaded a Schedule C worksheet with nothing left to review. These are the numbers from our own product data, anonymised. No quote, no name, no star rating.

Most bookkeeping stories are about keeping up. This one is about catching up, which is the harder problem and the more common one.

A mobile RV repair business in Georgia had two months of 2026 sitting untouched: January and February. Parts bought from a dozen suppliers, fuel from whichever station was closest to the job, payment processing fees landing in ones and twos, and a shoebox worth of receipts. The kind of backlog that does not get cleared on a Tuesday lunch break.

They cleared it in one evening.

What actually went in

Across that evening and the following morning, 106 transactions were entered:

• 85 expenses, each one with a receipt photographed and attached
• 20 income entries
• 1 personal purchase, marked as not deductible so it stayed off the return entirely

The work happened in one sitting. The first expense was logged in the early evening and the last one a little over five hours later. Everything was entered through the web app, by hand and by camera. No bank connection, no spreadsheet import.

Where it all landed on Schedule C

The 85 expenses sorted into six Schedule C expense categories:

• Supplies took the largest share, 33 rows, which is what you would expect for a repair business buying parts and materials
• Bank and Merchant Fees, 43 rows, mostly small payment processing charges
• Car and Truck, 6 rows of fuel
• Office Expense, Software and Subscriptions, and Taxes and Licenses, one row each

That spread is the useful part. A repair business that dumps everything into one bucket gets a Schedule C with a single large number on Line 27 and no real answer if anyone asks what it was. Split across Supplies, Car and Truck, and Taxes and Licenses, each total sits on the line it belongs on.

The one personal purchase matters as much as the rest. Marking it not deductible is what stops it quietly inflating a deduction.

Nothing was left for later

When the evening ended, the review queue was empty. Of 106 transactions:

• 0 were left flagged for review
• 0 were left without a category
• Average confidence on the category assignments was 97.8%

This is the difference between a book that is entered and a book that is finished. Plenty of tools will take 106 rows off your hands and hand back 106 rows that still need a decision each. Here the decisions were made as the rows went in.

The receipts did not all agree, and that is the point

85 receipts went through. 58 matched cleanly. 27 raised a question, because the figure read off the receipt did not agree with the amount on the row. Two amounts were corrected as a result.

A third of the receipts asking a question sounds like a lot until you think about what the alternative is. A receipt reader that never disagrees with you is not reading the receipt. Fuel receipts with a pre-authorisation hold, invoices with a promotion applied after the fact, totals that include a deposit: these genuinely do not match, and booking them silently is how a return ends up wrong in a way nobody catches until it matters.

Those 27 were put back to the owner as a question instead of being booked on the system's own reading of the paper. Two amounts were corrected as a result.

From signing up to a worksheet in hand

The account was created in the small hours. The Schedule C worksheet was downloaded about 22 hours later, with the bulk of the work in the hours just before it.

They also sent 176 chat messages along the way, which tells you this was not a silent click-through. Questions got asked. For a first pass through two months of unfamiliar categories, that is roughly what you would hope to see.

What this does and does not show

Being straight about the limits, because a use case that overstates itself is worthless:

• This was two months, not a full year. January and February. A full year would be roughly six times the volume, and we are not going to pretend we measured that.
• It is one business. One case is an existence proof, not an average. It shows the thing is possible, not that it is typical.
• The worksheet is a worksheet. It arranges your totals by Schedule C line so they can be copied onto the form or handed to whoever files for you. ClaryBook does not file your return.
• This is our own measurement, not a testimonial. Every number here comes from the product database. The owner was not asked for a quote and did not give one, and we have not used their name, their business name, or anything they bought.

Why we published the numbers instead of a quote

The easy version of this page is a smiling photo and a sentence in quotation marks. We do not have one, so we are not going to write one. What we do have is what the product recorded: how many rows, which categories, how many receipts disagreed, how long it took, and what was left over at the end.

If you run a repair business, a trade, or anything else where the receipts pile up faster than the paperwork, that is probably more useful to you anyway.