Reconciliation · Trust
The figures are checked,
not asserted.
This product’s whole claim is that its numbers can be trusted, which means the numbers have to be checked against each other. Every load runs the identities that must hold to the cent, the coverage checks that catch a source measuring half a month, and the variances worth naming.
Measured, not estimated · EU-hosted · Read-only by default
- Identitieshold to the centor it is a fault
- Ex-VAT revenue€6,3447,940 − 167 − 1,429
- Net profit€2,521reconciled
- Coverageflaggedwhen partial
4 of 4 identities hold, to the cent
- ✓Revenue less refunds is net revenue€7,773.00should be€7,773.00is
- ✓Net revenue less VAT is ex-VAT revenue€6,344.00should be€6,344.00is
- ✓The cost lines add up to total costs€3,823.00should be€3,823.00is
- ✓Ex-VAT revenue less total costs is net profit€2,521.00should be€2,521.00is
Each identity is arithmetic that cannot be argued with: if the parts do not add to the whole, something is wrong and the screen says so instead of rounding it away.
Coverage is reported too — what share of orders carried a measured cost, a measured fee and a known tax treatment — because an identity that holds over half the data has only half held.
01 / The case for it
A dashboard that is quietly wrong looks exactly like one that is right.
Without it
Every analytics tool prints figures with the same confidence whether or not they add up. If an arithmetic error creeps in, nothing on the screen changes — which is the reason a store owner is right to keep the spreadsheet open beside it.
What Profflow does
The figures are checked against each other, on every load.
The identities that must hold to the cent, the coverage checks that catch a source measuring half a month, and the variances worth naming. A failure is raised as critical and phrased as a fault in Profflow, not as a question about your data.
02 / Worth understanding
Why a product that claims accuracy has to prove it
Trust is a claim until it is checked
This product’s whole argument is that its figures can be relied on. That obliges it to run the checks in public rather than asserting the result, which is what this screen is: the arithmetic that has to hold, and whether it did.
A difference is shown, never absorbed
When an identity fails — a payout that does not match its charges, a cost that appeared after a month closed — the difference is displayed with the rows behind it. Silently adjusting a figure to make a total work is the one thing an accounting tool must never do.
Coverage is part of the answer
An identity that holds across 68% of orders has held across 68% of orders. The share is reported beside the result so a figure is never read as more complete than it is.
03 / What it lets you do
What follows from getting Reconciliation right.
Identities to the cent
If one fails, something in the arithmetic is wrong and every figure is suspect. It is raised as critical rather than smoothed.
Coverage checks
A processor connected halfway through a month measures half of it. A fee total silently covering fifteen days of thirty is worse than an honest estimate covering all of them.
Named variances
The fee rate in Settings against what the processor actually charged — two real measurements that need not agree, with the difference stated.
Measured against estimated
Every figure carries which of the two it is, all the way through the model.
Faults owned
An arithmetic failure is written as a fault in the product, not as a question about your data.
Traceable to rows
Each check opens to the orders and charges it was run over.
04 / What it is built from
Reconciliation is assembled
from accounts you already run.
- 01Order totals
- 02Refunds
- 03VAT accrued
- 04Cost of goods
- 05Processor fee statements
- 06Settings fee rates
- Identities to the cent
- Coverage checks
- Named variances
- Measured against estimated
05 / How it is worked out
From the account
to the figure on the screen.
- 01AssertsThe identities: revenue less refunds less VAT, then less every cost, is net profit.
- 02CoversWhether each measurement exists for the whole period it claims to cover.
- 03ComparesTwo legitimate readings of the same quantity, and names the gap.
- 04ReportsA failed identity as critical, and phrased as our fault.
[ Asked of the figures ]
How do I know the dashboard is not quietly wrong?
Because it checks itself and shows the working. Revenue less refunds less VAT is €6,344; less goods, fees, ads and fixed costs it is €2,521. If either identity failed, the screen would say so and call it a fault in the product rather than in your store.
How Profflow reasons06 / How the figure is kept honest
The same four rules,
on every screen.
- Read-only, on every plan
Profflow observes the accounts it is connected to. It does not write an order, a price, a campaign or a payout back to any of them.
- Measured stays measured
A charge read from an account is labelled measured. Anything modelled — a forecast, a projection, a filled gap — is labelled as such on the screen it appears on.
- Every figure opens
Follow any number to the dated rows underneath it. A figure nobody can check is a figure nobody acts on.
- You approve
The product prepares a decision and shows the evidence. Nothing consequential happens until you say so.
About
Reconciliation.
Can’t find the answer to your question?
Contact us by clicking here.
It is shown at the top, marked critical, and phrased as a fault in Profflow. A tool whose arithmetic has broken should say so loudly rather than keep printing figures.
It asks whether a measurement exists for the whole period. If a payment processor was connected on the fifteenth, its fees cover half the month, and a total that quietly covers half is more misleading than an estimate that covers all of it.
Not necessarily. Two legitimate measurements of the same quantity can disagree — the rate you entered against what was actually charged. The point is that the gap is named rather than averaged away.
Yes, on every load, over the period on screen. It is not a report you have to remember to run.
See Reconciliation on your own figures.
This product’s whole claim is that its numbers can be trusted, which means the numbers have to be checked against each other. Every load runs the identities that must hold to the cent, the coverage checks that catch a source measuring half a month, and the variances worth naming.