A vendor payout spreadsheet does not fail all at once. It degrades, and the symptoms are recognisable: refunds adjusted by hand, a payout run that takes an afternoon, vendors emailing to ask how they are doing, and formulas only one person understands. Most Shopify consignment stores hit two or three of these at around twenty vendors — not because twenty is a limit, but because each symptom costs time every month and they compound.
The spreadsheet is fine, until it is not
Worth saying first: a spreadsheet is a legitimate tool for this. It multiplies perfectly, it is free, everyone can read it, and for a store with eight vendors and few returns it will outlast several apps.
The mistake is not using one. It is not noticing when it has stopped working, because the failure is gradual and every individual month feels manageable.
Symptom one: refunds are adjusted by hand
The first to appear and the only one that costs money rather than time.
A refund happens in Shopify. The payout happened in a spreadsheet. Nothing joins them, so the adjustment depends on somebody remembering that an order from two months ago has been returned and finding the row it affected.
When it is missed — and it is — you have overpaid, silently. There is no error and no report. Stores usually discover a run of these during an annual reconciliation, long after recovering them is realistic.
The test: take last month's refunds and check each one was reflected in a payout. If that takes more than a few minutes, or if you find one that was not, this symptom is present.
Symptom two: the payout run takes an afternoon
Not the arithmetic — that is instant. The time goes to everything around it: chasing a product with no vendor set, working out which commission rate applied in March, allocating an order-level discount across two vendors, checking whether that big order was refunded.
The useful measure is not how long the sheet takes to calculate but how long between opening it and being confident in the numbers. Under an hour is fine. An afternoon means the sheet is now a process rather than a tool.
Symptom three: vendors email to ask how they are doing
This one scales with vendor count, not with sales, which is why it sneaks up.
Twenty vendors asking once a month is twenty lookups, whether you sold ten items or a thousand. Each is a few minutes. None feels significant. Together they are most of a working day, every month, spent reading numbers aloud that the vendor could have read themselves.
A vendor portal converts this into a link. It is the symptom with the clearest fix.
Symptom four: only one person can run it
The expensive one, and it shows up at the worst time.
A spreadsheet that has been maintained for two years has accumulated history — a column added for a vendor with an unusual rate, a hard-coded adjustment nobody remembers, a formula that handles one edge case and silently mishandles another. Its author understands it. Nobody else does.
You find out when that person is on holiday, or leaves. The payout either waits or is done wrongly by someone working from the shape of last month's file.
The test: could a colleague run this month's payout from the sheet alone, without asking anybody? If not, the sheet is a person rather than a process.
Symptom five: a vendor disputes a number and you cannot reconstruct it
The one that ends the argument for spreadsheets.
A spreadsheet holds current state. It does not hold how it got there. When a vendor queries a payout from four months ago, you need the rate that applied then, the orders included, the refunds deducted and the balance carried — and a sheet that has been edited every month since may no longer contain any of it.
Version history helps if you have it and if somebody saved a copy per period. Most stores have neither, and the conversation becomes a disagreement about memory, which nobody wins.
The errors a spreadsheet makes that nobody catches
Beyond the symptoms, there is a class of quiet mistake specific to spreadsheets. Each is individually small and all of them favour nobody consistently, which is what makes them hard to notice.
- The range that stopped at the old last row. A SUM covering rows 2 to 400 on a sheet that now runs to 430. Thirty orders are simply not in the total, and the number still looks plausible.
- The pasted value. Someone pasted a figure over a formula to fix a one-off, and the cell has been static ever since.
- The vendor name with a trailing space. A lookup misses it, and that vendor's total is short by exactly the orders that landed on the invisible variant.
- The rate changed in place. Updating a vendor's percentage in the cell that historical months also reference silently restates every past period.
- The duplicated row. An order counted twice during a copy-paste, which favours the vendor and is the least likely to be reported.
None of these produce an error. They produce a number that is wrong by an amount too small to notice and too large to ignore once it has run for a year.
The reason a tool helps is not that it multiplies better. It is that the data has one source and the arithmetic is not re-authored each month by hand.
What actually changes when you move
Be clear about what a tool does and does not solve, because the gap is where disappointment lives.
It fixes: refunds connected to the payouts they affect, rates applied by order date, attribution that cannot be silently missing, statements a vendor can check, and the running "how am I doing" question.
It does not fix: an undefined commission base, a vendor agreement that never said who bears discounts, intake processes, or deciding what happens to unsold stock. Those are decisions, and software will apply whichever one you have made — including the wrong one, consistently.
Which is why the order matters: settle the agreement, then pick a tool. The setup guide covers what belongs in writing.
Moving without breaking a month
- Write down the rules the sheet implements, including the exceptions. This is the document that does not exist yet, and you will need it whatever you choose.
- Run one full period in parallel. Both the sheet and the tool, same data, reconciled at the end.
- Investigate every difference. Some will be the tool being wrong. More will be the sheet having been wrong for months.
- Keep the sheet archived, per period, as the historical record for anything predating the move.
- Then switch, and stop maintaining two.
Step three is the valuable one. The parallel period usually finds a product that has been unattributed since spring, or a vendor whose rate changed and was applied from the wrong month.
Payvenda's free tier covers five vendors and twenty synced orders a month, which is enough to run a parallel period on a sample without paying for the privilege of evaluating.
Where to go next
If two or more symptoms are present, the sheet is costing more than it looks. Start by writing the rules down — that is useful regardless. Then compare the tools on your own numbers rather than on anyone's feature table.
