Nobody publishes a clean figure for how many hours a multi-marketplace seller loses to bookkeeping each month, because nobody is measuring it. What does exist is a set of official numbers that bound the problem from several directions: what the government estimates compliance costs, what bookkeeping labour is worth, and how much reconciliation volume the payout mechanics of each marketplace actually generate. Put those together and the picture gets uncomfortably specific.
Every figure below carries its source in the text. Where a number is an estimate rather than a measurement, it says so.
What the IRS says filing alone costs
The Internal Revenue Service publishes taxpayer burden estimates in the Paperwork Reduction Act notice attached to its form instructions. In the Instructions for Form 1120-S (2025), with estimates current as of December 2025, the IRS puts the average burden for Forms 1120-REIT, 1120-RIC, 1120-S and all related attachments at 60 hours and $4,800.
The neighbouring figures in the same notice are worth having. Partnerships filing Form 1065 and related attachments average about 60 hours and $5,300. Corporations filing Form 1120 and associated forms average about 90 hours and $8,600.
Two caveats the IRS itself attaches. These are national averages that do not necessarily reflect a typical case, and most taxpayers experience lower than average burden. The out-of-pocket figure includes preparation and submission fees, postage, photocopying and tax software.
What that number does not include is the monthly close. It is the annual filing event only. Whatever it costs you to keep books during the year sits entirely on top.
What the labour is worth
The Bureau of Labor Statistics Occupational Outlook Handbook puts the median annual wage for bookkeeping, accounting and auditing clerks at $49,210 in May 2024, which works out to $23.66 per hour. The lowest ten percent earned under $34,600 and the highest ten percent over $72,660.
By industry, the same BLS release gives May 2024 medians of $51,670 in construction, $50,180 in professional, scientific and technical services, $48,810 in both healthcare and wholesale trade, and $45,030 in retail trade. Retail sits at the bottom of that list.
The $23.66 figure is the useful one for arithmetic, and it is conservative in two ways. It excludes employer payroll taxes and benefits, and it is the rate for a clerk rather than for the founder who is often doing this work personally at three or four times the opportunity cost.
Where the labour goes
Reconciliation volume is a function of payout mechanics, and those are documented by the platforms themselves.
Shopify Payments offers a choice of daily, weekly or monthly payouts, according to Shopify’s own Help Center. In the United States the minimum settlement time is two to five business days with no minimum payout amount, and payments captured Friday, Saturday and Sunday are consolidated into a single payout. Shopify’s documentation works the example: a charge captured Friday under a three business day settlement processes the following Wednesday, because weekends and holidays do not count.
The consequence is arithmetic. A store on daily payouts generates roughly 22 settlement events a month. On weekly, four or five. Each event is a separate reconciliation: gross sales, refunds, processing fees, chargebacks, adjustments. Same revenue, five times the reconciliation work, purely from a scheduling choice made once during setup.
Amazon’s published referral fee schedule at sell.amazon.com is the other volume driver. It lists more than thirty product categories, and a meaningful share of them are tiered by price rather than flat. Clothing and Accessories runs 5 percent at or under $15.00, 10 percent above $15.00 through $20.00, and 17 percent above $20.00. Baby Products and Beauty, Health and Personal Care both switch from 8 percent to 15 percent at the $10.00 threshold. Grocery and Gourmet switches at $15.00. Electronics Accessories charges 15 percent on the portion up to $100.00 and 8 percent above. Jewelry charges 20 percent up to $250.00 and 5 percent above. Watches charge 16 percent up to $1,500.00 and 3 percent above. Media items carry 15 percent plus a $1.80 per item closing fee.
A seller who models Amazon fees as a single blended percentage will be wrong on every SKU that crosses a threshold, and wrong in a direction that changes as their price points move. Getting it right means fee logic at the unit level, which is not a spreadsheet task at volume.
The channel-count multiplier
Each additional marketplace adds its own settlement format, fee vocabulary and payout cadence. Amazon’s Professional selling plan costs $39.99 per month against $0.99 per item sold on the Individual plan, per Amazon’s own pricing page, with a referral fee minimum of $0.30 in most categories. TikTok Shop charges its own category-based referral fee on qualified transactions, which Shopify’s Help Center dates to April 16, 2024. None of these settlement reports look alike.
And the channel count keeps rising, because the newest channel is growing fast. Momentum Works, in its TikTok Shop in the U.S. 2025 report published with Tabcut in February 2026, put US TikTok Shop GMV at US$15.1 billion in 2025, up 68 percent from US$9 billion in 2024, with global GMV reaching US$64.3 billion across 16 markets. The same report counted 803,500 US stores, of which more than half recorded no sales at all, while more than 2,000 exceeded US$1 million in GMV.
That last split matters for anyone building a bookkeeping estimate. The median TikTok Shop store has no reconciliation problem because it has no sales. The 2,000 stores above a million dollars have a substantial one, and they almost certainly have it on two or three other marketplaces simultaneously.
Building an estimate you can actually check
There is no authoritative monthly hours figure, so build your own from components you can count rather than borrowing a vendor survey number.
Count your settlement events per month, per channel. A three-channel seller running Shopify on daily payouts, Amazon on a fourteen-day settlement cycle and Walmart on its own schedule is looking at roughly 22 plus 2 plus 4, or about 28 reconciliation events monthly. Time yourself on one of each, honestly, including the ones that do not balance on the first pass. Multiply.
Then add the fixed monthly work that does not scale with settlement count: the landed cost update when a supplier invoice or freight bill changes, the inventory adjustment, the review pass, and the inevitable investigation into why the clearing account is $340 out.
Cost it at the BLS median of $23.66 an hour if someone else does it, or at your own effective hourly rate if you do it yourself. Then set the total against the annual IRS filing burden of 60 hours and $4,800, which you owe regardless, and against the subscription cost of automating the settlement layer. Sync and reconciliation tools publish entry pricing from around US$29 per month for a single channel; platforms that add inventory valuation and SKU-level profit reporting, including ConnectBooks, price higher.
The number nobody quotes
The BLS projection is the most interesting figure in this entire piece and it rarely appears in discussions of bookkeeping cost. Employment of bookkeeping, accounting and auditing clerks is projected to decline 6 percent from 2024 to 2034, a loss of 94,300 positions from a 2024 base of 1,613,400, against 3 percent average growth across all occupations. BLS attributes the decline specifically to software automating routine tasks.
Despite the decline, BLS projects about 170,000 openings each year over the decade, all of them from replacement need rather than growth, as workers retire or move to other occupations.
Read those two together and the labour market answer becomes clear. The routine reconciliation work is being automated whether individual sellers participate or not, and the people remaining in the field are moving toward analysis. The question for a seller is not really how many hours manual bookkeeping costs. It is whether those hours are being spent on work that a settlement parser will do more accurately for a subscription fee, and whether the person doing them could be doing something the software cannot.
