What manual matching costs you
Four numbers. The second is the one people underestimate, because the payments that match on sight are the ones you remember.
Everything that lands, including the ones that are obvious.
Not the ones you cannot resolve at all — the ones that take more than a glance. Count for one month before guessing.
Including the ones you go back to twice, and the message you send to ask.
Use your billing rate. This work displaces billable work; it does not happen in spare time.
- Payments needing real work
- 36
- Hours a month
- 3.0 hours
- Monthly cost of the work
- R$240
- Over a year
- R$2,880
At 90 payments a month with 40% needing real work, that is 36 payments. Five minutes each is 3.0 hours a month. At R$80 an hour the work costs R$240 a month, or R$2,880 over a year — against a tool priced at R$39.
Assumptions behind these numbers
- The hourly rate is what the time would otherwise have earned, not a salary cost. For a one-person business those are the same thing and the number is high.
- It excludes the cost of getting it wrong. A misallocated payment surfaces weeks later as a wrongly chased client, and that costs far more than five minutes.
- It assumes the work happens. In practice it accumulates and is done in one painful session at month end, which is slower per item than the model suggests.
- It ignores the tax and accounting deadline pressure that makes the same work more expensive in some weeks than others.
Nothing is submitted anywhere. The calculation runs in your browser.
Why the second number is always wrong at first
Everybody guesses low on the share of payments that need work, and the reason is memorable: you remember the payments you resolved instantly and forget the ones you resolved after a two-minute detour. The detours do not feel like work because each is small.
Count for one month. Put a mark next to every payment that took more than a glance, including the ones you resolved in thirty seconds by opening a second window. The number is generally somewhere between a quarter and a half for a business receiving on a static key, and it is usually double what the same person guessed.
Where the break-even sits against a R$39 tool
Holding minutes at five and rate at R$80, varying only volume and difficulty.
| Receipts a month | 20% need work | 40% need work | 60% need work |
|---|---|---|---|
| 20 | R$13 | R$27 | R$40 |
| 50 | R$33 | R$67 | R$100 |
| 90 | R$60 | R$120 | R$180 |
| 150 | R$100 | R$200 | R$300 |
| 300 | R$200 | R$400 | R$600 |
Source: The formula on this page at each combination. Monthly cost of the manual work, before comparing it to any tool.
The top-left cell is the honest one. Twenty receipts a month with most of them obvious costs about thirteen reais of time, and buying anything to fix that would be silly. This table exists mostly to tell some readers not to buy.
Five ways to cut the number without buying anything
In descending order of how much they help.
- Keep a payer-name-to-client listTwo columns. It converts the weakest signal in the record into the strongest one, and it costs five minutes a month to maintain.
- Move new clients to requested paymentsEvery client onboarded from today can be paid through a code that carries your reference. It does nothing for the existing ones and compounds quietly.
- Stop using round numbers for everythingFour clients at R$500 collide constantly. Pricing at R$495 and R$510 sounds petty and eliminates an entire failure mode.
- Record the date you asked, not just the date you invoicedTiming is the second-strongest signal and it depends on knowing when the client was last prompted.
- Reconcile weekly rather than monthlySame total work, much less of it, because a payment from four days ago is far easier to place than one from three weeks ago.
The first and third of these are close to free and between them remove a good share of the problem. A page selling a subscription should say that.
If the number is bigger than you expected
In the context field: the figure the calculator gave you. It is the most useful thing anyone can tell me about this idea.