Ask an operator what their bookkeeper does and you'll usually get a shrug and a number: "a few thousand a month, they handle the books." Ask what "the books" includes and the answer starts to sprawl — categorizing transactions, sure, but also keying vendor invoices, cutting a few checks, chasing a rep about a credit, and pulling a report at month-end. Somewhere in that sprawl, two completely different jobs got fused into one line item. One of them is bookkeeping. The other is your back office. You're almost certainly paying a premium for the second one and getting the worst version of it.
The distinction sounds academic until you see what it costs. So let's draw it cleanly.
The clean line: recording vs moving
A bookkeeper records what already happened. That's the whole discipline, and it's a real one: every dollar in and out gets categorized, the accounts get reconciled, the month gets closed, and you end up with financial statements that are accurate and defensible. It is backward-looking by design — its job is to describe the past correctly so your CPA can file taxes and you can see where the business stands. A good cannabis bookkeeper who understands 280E is worth keeping.
The back office does the thing the bookkeeper later records. It pays the vendors. It runs the consignment settlements. It recovers the credits vendors owe you. It collects the invoices your buyers slid past terms on. It is forward-looking and it moves real money now — this week, this cycle — not next month in a report. The bookkeeper writes down that a check went out. The back office is what decided the amount, matched it to the delivery, caught the overcharge, and cut it.
Bookkeeping is the record of the money. The back office is the movement of the money. One describes the past for compliance; the other changes your bank balance today.
Neither replaces the other. The mistake isn't hiring a bookkeeper — it's asking the person built to record money to also be the one who moves it, then wondering why the moving part is slow, expensive, and leaky.
Where the two jobs actually diverge
Line the tasks up side by side and the split is obvious. A ledger tool and a bookkeeper are built for the left column. Moving money against Metrc, vendor agreements, and the settlement cycle is the right column — and it's a different competency, not a harder version of the same one.
| The job | Bookkeeper | Back office (ShelfSpace) |
|---|---|---|
| Records & categorizes transactions | ✓ | Feeds it clean, pre-coded |
| Month-end close & reports | ✓ | ✗ |
| Files 280E & taxes | ✓ | ✗ |
| Pays vendors (matched, real checks) | Data entry only | ✓ |
| Runs consignment settlements | ✗ | ✓ |
| Recovers vendor credits | ✗ | ✓ |
| Collects overdue AR | ✗ | ✓ |
| Matches invoices to Metrc | ✗ | ✓ |
Notice the two rows in the middle where the bookkeeper gets a note instead of a check or an X. Those are the seams where operators fuse the jobs — the bookkeeper does the invoice data entry and cuts the checks, so it looks like the back office is covered. It isn't. Keying and printing is the mechanical shell of accounts payable. The part that saves money — matching every invoice against the Metrc manifest and the delivery, catching the duplicate line and the phantom charge, and systematically pulling back the credits your vendors owe — never happens, because a ledger isn't built to do it and a bookkeeper isn't paid to hunt for it.
The most expensive way to do the cheapest work
Here's the math nobody puts on the invoice. Your bookkeeping or AP firm bills a few thousand dollars a month. A meaningful slice of those hours is data entry — retyping invoices, cutting checks, reconciling a settlement spreadsheet. That's clerical work billed at a professional's rate. Meanwhile, the back-office labor that doesn't land on the firm — vendor email, credit chasing, matching deliveries — lands on an owner or an office manager instead, and that runs 10 to 15 hours a week of someone's time that should be on the floor or the assortment.
So you're paying twice: a premium rate for the cheap half of the work, and your own team's hours for the rest. And the expensive part — the money you're actually owed — still slips through, because recovering credits was never in anyone's job description.
10–15 hrs
a week your team spends on the back-office work the firm doesn't touch
$200K+
a year in vendor credits that go unrecovered at a multi-location operator
$8K–$25K
a month in recoverable credits found at one Massachusetts client
The $200K+ a year in untracked credits at a multi-location operator, and the $8K to $25K a month we surfaced at one Massachusetts client, aren't line items your bookkeeper missed out of carelessness. They're line items a general ledger and a compliance-minded firm are structurally not built to catch. A firm keeping your books to be accurate and tax-ready is doing exactly the job you hired it for. Finding the aging credit a vendor owes you on a case of expired product is a different job entirely — and if you also wholesale, so is chasing the receivable a buyer let drift from net-30 to net-90.
There's a timing problem baked into the fusion, too. Bookkeeping runs on a monthly rhythm — the close happens after the period ends, which is the correct cadence for a record. But money moves weekly. Vendors expect payment on delivery terms, consignment settles on a weekly cycle, and a credit you don't claim inside the return window can lapse before the month is even closed. When the two jobs share one owner working to a month-end calendar, the movement inherits the record's pace. By the time a reconciliation surfaces that you overpaid a duplicate invoice or missed a co-marketing credit, the check has cleared and the window has passed. The information arrives accurate and too late — which is fine for taxes and expensive for cash.
That lag is the hidden third cost, after the premium labor rate and the unrecovered credits. A backward-looking function will always tell you what went wrong after it's already gone wrong. Moving money well means catching the overcharge before the check is cut and claiming the credit while it's still live — which only happens when the back office runs on its own clock, against real delivery and Metrc data, instead of waiting for someone to get to it between closes.
What the back office looks like when it's handled
Put the movement of money on a system built for it, and the same week stops eating your team.
Bookkeeper doing double duty
A few thousand a month
Invoices retyped by hand, a few checks cut, a settlement spreadsheet reconciled after the fact. The books are accurate. The overcharges aren't caught, the credits aren't recovered, and your team still runs the vendor email.
Back office on ShelfSpace
A few minutes to approve
The platform reads Metrc and your POS, matches every invoice, runs the settlements, surfaces the credits vendors owe you, and drafts the vendor email. You approve. Checks go out. Every entry lands in QuickBooks pre-coded.
Under the hood, ShelfSpace connects to Metrc and reads your POS. It three-way matches each invoice against the manifest and the delivery so you never overpay a phantom line, then pays the vendor with a Check 21 check the card rails can't handle. It runs weekly consignment settlements against real sell-through and generates a report both sides trust. It builds monthly credit memos for returns, expirations, and co-marketing. And when you wholesale, it works the overdue receivables instead of letting a collection agency take a 25% to 50% contingency cut of money that was always yours. ShelfiQ, the AI layer, drafts and answers the routine vendor email so your team isn't the help desk — that's the engine that keeps it affordable, not the pitch.
The two connect through QuickBooks
This is the part that makes the split practical instead of a turf war. You don't fire your bookkeeper to put the money movement on ShelfSpace — you feed them better. Every bill, payment, settlement, and credit ShelfSpace moves writes straight into QuickBooks with the categories your accountant needs. The QuickBooks sync means your bookkeeper opens the month to find the transactions already coded and reconciled against Metrc, not a shoebox of invoices and a POS export to rebuild from memory. The back office moves the money; the ledger records it clean; the CPA files on top of both. Each does the job it's actually built for.
A bookkeeper tells you what happened to your money. A back office decides what happens to it. Keep the first for the ledger and the taxes — put the second on a system built to move it, and stop paying a premium to do the cheap work while the real money walks out the door.
Who should redraw the line
Operators paying a firm for "the books" who can't say how much of that bill is data entry versus actual accounting — the answer is usually more than they'd guess, and none of it is recovering credits. Single-location operators whose bookkeeper does the AP by hand and whose credits have never been systematically claimed. And multi-location groups where the back office doesn't scale: every new store multiplies the vendors, the settlements, and the money nobody's chasing, while the ledger just gets longer.
If that's your setup, the honest first move isn't switching bookkeepers — it's seeing the number. We'll connect to Metrc and your POS and show you what you're owed in unrecovered credits, how much of your receivables is overdue, and where your invoices don't match Metrc. It's free, it's specific to your operation, and it tends to surface far more than it costs to fix. If you're a bookkeeper yourself, we built a partner path so the money movement feeds your ledger instead of fighting it — the same way an outsourced finance function or a full outsourced back office would, without giving up the accountant you trust.