A full-time in-house bookkeeper costs a US ecommerce business roughly $63,000 to $72,000 a year once payroll taxes and benefits are loaded on. Automated marketplace accounting plus an outsourced monthly close runs between $6,000 and $14,000. That is not the comparison most sellers should make, because the two options do not do the same work. The honest comparison is between a bookkeeper who also handles the things software cannot, and software plus the fraction of a person you need.

What the salary number covers

The Bureau of Labor Statistics Occupational Outlook Handbook, last updated August 27, 2026, reports 2025 median pay for bookkeeping, accounting, and auditing clerks at $50,670 a year, or $24.36 an hour. The distribution is wide: the lowest 10 percent earned under $36,000 and the highest 10 percent earned more than $74,550. Retail trade pays below the median at $45,830, while construction pays $54,890 and professional, scientific, and technical services $51,880.

Salary is not cost. Employer payroll taxes, unemployment insurance, workers compensation, health contributions, software seats, and equipment typically add 25 to 40 percent. Take the median of $50,670, apply 30 percent, and you are at $65,871. Hire above median for someone who has closed marketplace books and you are near $80,000 all-in.

For context on the next rung up, BLS puts 2025 median pay for accountants and auditors at $83,680, and the median across all occupations at $50,980.

What the software number covers

Marketplace sync software prices between $29 and $600 a month depending on channels and depth. A2X starts at $29 for a single channel and reaches $229 a month on its multi-channel plan at 10,000 combined orders. ConnectBooks publishes starting prices of $149, $199, and $349 for its three tiers. Webgility runs from $69 a month annually at 300 orders to $599 for its highest published tier.

That covers the data movement. It does not cover the close. An outsourced monthly close from a service like Bookkeeper360 starts at $399 a month, Xendoo at $355 to $895 depending on monthly expense volume, and Webgility’s full service engagement at $649 a month with the platform included.

So a realistic automated stack for a multi-channel seller is $350 a month for sync plus $500 a month for the close, or roughly $10,200 a year. Add a CPA for the tax return and you are near $12,000.

The framework: what does each side do

Where software wins outright

Volume and repetition. A settlement report with 40,000 line items is not a human task. Neither is applying FIFO cost to every unit shipped, matching a 1099-K to twelve months of gross sales, or converting foreign currency transactions at the right daily rate.

BLS itself makes this point without editorializing. Its outlook section states that software innovations have automated many of the tasks performed by bookkeeping, accounting, and auditing clerks, and that as a result the same amount of work can be done with fewer employees. Employment in the occupation is projected to decline 6 percent from 2025 to 2035, a loss of 85,600 positions, even though about 144,100 openings a year are expected from turnover.

Where a person wins outright

Judgment, exceptions, and everything that touches other people. Chasing a supplier about a duplicate invoice. Deciding whether a returned pallet is sellable. Noticing that the 3PL bill went up 11 percent and asking why. Handling payroll questions. Talking to your bank.

A bookkeeper is also the person who catches the thing no rule anticipated. Software will faithfully post an incorrect landed cost for eighteen months. A person who touches the numbers weekly tends to notice when a margin moves for no reason.

BLS anticipates exactly this shift, projecting that with more automation of routine tasks these workers will take on a more analytical and advisory role, focusing on analyzing books and pointing out areas for efficiency gains rather than entering data by hand.

Where neither one covers you

Compliance judgment. IRS Publication 538 states that an inventory is necessary to clearly show income when the purchase or sale of merchandise is an income-producing factor, and that a business required to account for inventory must use an accrual method for purchases and sales. The small business taxpayer exception applies to filers averaging $26 million or less in annual gross receipts over the three prior tax years who are not tax shelters. Deciding whether you qualify, and whether a method change is worth filing for, is a CPA question either way.

A worked comparison

Take a seller doing $4.2 million across Amazon, Shopify, and Walmart, with 280 SKUs, two warehouses, and four employees.

Option A, in-house. One bookkeeper at $58,000 salary, loaded at 30 percent, is $75,400. Add QuickBooks Online Plus at roughly $1,200 a year and a marketplace sync tool at $3,000, since a person still cannot hand-enter 40,000 settlement lines. Add a CPA for the return at $3,500. Total: about $83,100.

Option B, automated plus outsourced. Sync software at $349 a month is $4,188. Outsourced monthly close at $599 is $7,188. CPA at $3,500. Total: about $14,876.

The gap is roughly $68,000 a year, which is the honest headline. What the gap buys you in Option A is 2,000 hours of someone’s attention, of which the close consumes maybe 300. The other 1,700 hours go to AP, AR, supplier communication, payroll support, and the exception handling nobody scopes in advance.

If those 1,700 hours are currently being absorbed by the founder at midnight, Option A is not a $68,000 cost. It is a $68,000 purchase of the founder’s time, and the question becomes what that time is worth deployed elsewhere.

The hybrid that most sellers land on

Sync software plus a part-time contractor at 10 to 15 hours a week. At the BLS hourly median of $24.36, twelve hours a week is $15,200 a year, with no benefits load and no hiring risk. Combined with $4,200 of software, that is under $20,000 and covers both the mechanical close and the human exception handling.

Category tooling has consolidated around making that hybrid workable. A2X, Webgility, Sellerboard, and ConnectBooks all approach the settlement problem from different angles, but the shared assumption is the same: move the mechanical work off a person so the person can do the parts that require deciding something.

The tie-breakers

Transaction volume. Under about 500 orders a month, a competent part-time bookkeeper with basic tooling is fine. Above 5,000, hand reconciliation stops being viable at any headcount.

How many things touch inventory. One warehouse, one channel, stable costs: a person can hold it. Three channels, two warehouses, and landed costs that move with freight rates: you need a system, and then you need a person to check the system.

Whether anyone is watching. A lender covenant, an investor, or an acquirer changes the calculation, because the standard shifts from “good enough for taxes” to “survives diligence.”

Turnover risk. A single in-house bookkeeper is a single point of failure who takes undocumented process knowledge with them. The IRS recordkeeping guidance notes that the responsibility to substantiate entries and deductions, known as the burden of proof, sits with you. It does not transfer to the person who left in March.

The version of this that goes wrong

Neither. What is wrong is hiring a bookkeeper and then not buying the software, on the theory that a person makes the tooling unnecessary. That produces a $75,000 employee spending two days a month reconciling settlement reports by hand, which is both the most expensive way to do the mechanical work and the least accurate.