About Guardrail
George I.. Payments industry, 20+ years.
Who runs this
I spent two decades in transaction banking and payments infrastructure. Clearing operations at a major European bank, FX product management at a multinational broker, and the last several years advising fintechs on cross-border payment architecture.
Most of that work came down to one question: when does the money actually arrive? Not when it was sent, not when the invoice was raised, but the day it lands and clears. A lot of my recent work has been on freelancer payment workflows, including Upwork, Fiverr, Toptal and direct invoicing, and on receiving corridors in markets like Pakistan, Bangladesh, Nigeria and Egypt, where settlement is slow and the arrival date is genuinely hard to predict.
That is the background I brought to this. Most freelancers and contractors I have worked around do not have an income problem. They have a timing problem. The invoice gets paid. Just not on the day the rent leaves.
Why Guardrail exists
Budgeting tools mostly look backwards. They sort what you already spent into categories, which works well enough if the same amount arrives on the same day every month. If you are paid per project, per invoice or per shift, that framing answers a question you were not asking.
The question is not what you spent on groceries in March. It is whether the balance survives until the next payment lands, and what happens if that payment is two weeks late. Guardrail answers that and nothing else. You enter your balance, what you expect to come in, and what has to go out. It projects the balance forward one day at a time and gives you the first date it drops below zero, then suggests one change that pushes that date further out.
How the projection is calculated
Guardrail takes manual input only. It does not connect to your bank, and it does not use Plaid or any other account-linking service. You type the figures in yourself, and you can delete them whenever you want.
The arithmetic is deliberately plain:
- Start from the balance you enter.
- Add each income event on the date you expect it to land, not the date you invoiced it.
- Subtract each obligation on the date it leaves the account.
- Step forward a day at a time, and record the first date the balance goes below zero along with the lowest point it reaches.
No smoothing, no averaging into a tidy monthly figure, no assumed growth. I have left it simple on purpose, because the failure mode of a clever forecast is that you stop trusting it and stop opening it.
The projection is only as good as the dates you give it. Tell it an invoice arrives on the 1st when the client habitually pays on the 20th, and it will be wrong, in the direction that hurts. It also cannot know about money you did not enter. Those are the real limits and they are worth stating plainly rather than burying.
What Guardrail is not
It is not financial advice, and I am not a financial adviser. Twenty years in payments infrastructure teaches you how money moves between institutions. It does not qualify me to tell you what to do with yours.
Guardrail does arithmetic on numbers you typed in. It holds no view on whether you should take the contract, move house, or clear the card before the loan. Where it suggests an action, such as shifting a payment date, that is a mechanical output of the projection, not a recommendation about your circumstances. If you are making a decision that matters, talk to someone qualified and regulated to advise on it.
Standards I hold this to
Same rule I apply to everything I publish: on a site dealing with people's actual income, a wrong number is worse than a missing one.
- The method above is the entire method. There is no hidden model behind it.
- Anything that is an estimate is labelled as an estimate.
- Pages carry a date showing when they were last reviewed.
- If you find an error in the arithmetic, tell me. I will fix it and say what changed.
Why a pen name
I publish under the same name across the sites I run. That started with review work where I had prior working relationships with some of the companies being compared, and keeping a pen name kept those relationships out of the conclusions.
It matters less here, because Guardrail reviews nobody and sells nobody's product but its own. The principle still holds though. The method is written out above in full so you can check it against your own numbers on paper. It either adds up or it does not, and that does not depend on who wrote it.
Last reviewed . Something wrong on this page? How we handle corrections.