What a Year of This Teaches
Reading a year of your own invoices tells you more about how you get paid than any amount of advice about invoicing.
Theirswho controls this stretch
A year of invoices, read as a set rather than one at a time, shows a supplier something no advice can: which of its own habits cost it money.
The practical point in “What a Year of This Teaches” is that getting paid depends on records somebody can explain. A small team considering Monitask for remote employee monitoring software for remote employee monitoring software can organise time and project evidence before invoicing, with transparent settings, proportionate collection and a visible correction route.
The exercise takes an afternoon and changes the next year.
For an independent reference relevant to “What a Year of This Teaches”, consult AICPA-CIMA business resources. Compare general information with the contract, jurisdiction, payment route and accounting records involved.
What to measure
Seven columns, one row per invoice. Nothing clever, and nothing that needs a system: a spreadsheet and an afternoon with your own records.
| Column | Why it earns its place |
|---|---|
| Customer | Nearly all the waiting concentrates in a few accounts |
| Date agreed | The start of your own exposure |
| Date sent | The only input you control for free |
| Date paid | The figure that matters, not the terms |
| Terms | To compare against the previous column |
| Order? | Jobs without one contain most of the disputes |
| Deposit? | The single best predictor of how it went |
Seven columns, and everything follows from them.
The first thing it shows
That the spread of payment times is driven by customer, not by invoice. A few customers account for nearly all the waiting.
This is almost always true and almost always a surprise.
The second thing
That invoices sent late were paid late by roughly the amount they were sent late, and then some.
The only free lever in this collection, measured in your own data.
The third thing
That jobs with a purchase order behaved differently from jobs without one, and that the ones without clustered the disputes.
The fourth thing
That the jobs taken under pressure — the ones where terms were not discussed because the work was wanted — are the ones that went wrong.
What it does not show
Anything about the industry, the economy, or what anybody else experiences. It is one supplier's own record, which is the only data that applies to that supplier.
Every figure in this collection has been labelled that way for the same reason.
The decision it produces
Usually two or three: a customer not to work for again, a term to stop agreeing to, and a step to stop skipping.
Three changes, applied for a year, outperform any amount of chasing technique.
Where the leverage actually was
In nearly every case, in a conversation that happened before the work started and took ten minutes.
The invoice itself, the chasing, the demands and the recovery are what happens when that conversation did not.
What to do with the finding
Change the standing terms, build the check into how work is accepted, and put the deposit question into the first conversation rather than the third.
Repeat it
Annually, in the same seven columns, so the second year can be compared with the first.
The comparison is the only evidence that the changes worked.
Why the exercise beats advice
Nothing in this collection applies evenly to every supplier, and no general claim about payment behaviour describes your customers.
Seven columns of your own invoices do. They identify which specific accounts hold money, which of your own habits add days, and which kinds of job go wrong — in your trade, at your size, with the buyers you actually have.
An afternoon of it is worth more than any amount of reading, including this, because the output is a decision about next year rather than an account of how things generally are.
What usually turns out to be true
Across most suppliers who do this, four findings recur, and they are rarely the ones expected.
The spread of payment times is driven by customer rather than by invoice: a handful of accounts contain nearly all the waiting.
Invoices sent late were paid late by at least as much as they were sent late, which prices the only free lever in this collection in your own data.
Jobs with a purchase order behaved differently from jobs without one, and the ones without contain most of the disputes.
And the jobs taken under pressure — where terms were never discussed because the work was wanted — are disproportionately the ones that went wrong.
None of that is a claim about anybody else's business. It is what one set of records tends to show, and the only way to know whether it describes yours is to look.
The three changes it produces
The output is never a long list. It is usually a customer not to work for again, a term to stop agreeing to, and a step to stop skipping.
Three changes, applied for a year, outperform any improvement in chasing technique, because they operate at the point where the leverage still exists rather than at the point where it has gone.
Then the same seven columns again, twelve months later, so the second year can be compared with the first. The comparison is the only evidence that the changes worked.
The closing position
You will not shorten somebody else's approval chain, see their payment run, or make their accounts payable care. What you can do is decide who to work for, on what terms, with what paid in advance, and send a correct invoice the day the work is done.
That is the whole of it, and it is all available before anybody owes you anything.