How I’m building Runway

A few things I believe

Every number should show its work.

Tap any month and see exactly what moved — which income landed, which bill left, what the investments did. No black box, no hidden assumptions. If a number changed, you can point at the reason.

Answers are months, not percentages.

I don’t think “a 72% chance of success” helps anyone decide anything. “If the work stops, you have 2 years and 5 months” does. So Runway answers in months, and if you want to know what 2008 would do to that, you run 2008.

Trying something should never change anything.

A simulation is a question, not a decision. You can lose a client, move cities, take six months off — and close it, and nothing happened. If it changes your mind, you change your plan yourself, on purpose.

Your numbers stay on your phone.

No account, no cloud, no bank login. Runway works with nothing but what you type in. I never see your money, and there is no server that could leak it.

The past is never quietly rewritten.

Once a month closes, it stays what it was. A correction is a new entry with today’s date, not an edit to history. Your record should be something you can trust more the longer you keep it.

Nothing happens behind your back.

Runway computes, suggests, and explains. It does not decide. Anything that didn’t come from you — a suggestion, a default, someday a market price — arrives as a proposal with its source shown, and changes nothing until you accept it.

Plans should read like sentences.

“Emergency fund until it covers six months, then pay off the card, then invest the rest.” If a rule can’t be said out loud like that, it doesn’t belong in the app — and every rule you set can be read back to you the same way.

What it does today

A month-by-month timeline.

Your savings, investments, income, spending and debts, projected forward. The runway at the top is how long it all lasts if the income stopped — and it moves the moment anything else does.

A monthly check-in, not daily bookkeeping.

Once a month, Runway shows you what it expected and asks what actually happened. Confirm, correct, done — about ten minutes. In between, a one-off takes seconds to log. Skip a month and nothing breaks.

Named “what ifs”.

Lose the main income. A client leaves. A rent shock, a medical recovery, a sabbatical, a new baby. And the historical ones — 2008, the pandemic, the dot-com bust, the 1970s, the Great Depression — replayed against your numbers. Two dozen to start with, and you can write your own.

Goals that report back.

A buffer of six months, a number in the savings account, a debt at zero — each one tells you the month it arrives, and that month moves every time your plan does.

A plan for the money, in your words.

Where extra money goes when a month ends ahead — the emergency fund first, then the debt, then investing — and what covers a month that runs short. Keep more than one plan, and try each one inside a what-if.

What’s next

Planned, in roughly this order. None of it ships until it works.

Rules that know when to stop.

“Pay extra on the card until it’s gone, then invest.” “Build the buffer until it covers six months.” The rule ends itself the month the condition is met — and ends on a different month in every what-if, because the numbers are different.

More than one currency.

Runway stores currencies today but doesn’t convert them. Next: one reporting currency for the totals, with every account keeping its own.

Moving somewhere.

“Move to Barcelona in September” as a what-if: rent, groceries and transport repriced for the place, the moving cost, the income that changes — each line visible and editable, never a single verdict.

Buying something, and the day after.

A car, a flat, a laptop. Not just saving up for it — the insurance, the fuel, the rent that stops. “Buying in June leaves four months of runway; waiting until September leaves nine.”

Investments by holding.

Model the portfolio as what it actually holds, so “what if only my tech positions halve” becomes a question you can ask.