Financial Forecast

A bank balance tells you what you have today. It says nothing about what you'll have four months from now. While a company is tiny that gap doesn't hurt: roughly the same money goes out every month, roughly the same comes in, and one division in your head gets you the answer.

The trouble starts the moment a decision gets added on top. At Pagedeck, Theo once asked to hire three more engineers and Grace couldn't answer in the room. The balance looked comfortable, but nobody had written down how much three salaries would add to the monthly outflow or how many months that would cut off the runway — the number of months the cash lasts at the current rate of spending. The call got made on instinct, and two months later the balance made it obvious that it had been a stretch.

A financial forecast is that calculation done in advance. Instead of making a decision and reading the result off the bank account later, you translate the decision into numbers and look at what happens in which month. The table isn't built to be right. It's built to let you compare options, and that distinction shapes everything else about it.

A forecast lays revenue, costs, headcount and cash across monthly columns — twelve of them, starting with the current month. The rows break down like this.

1. Revenue — money expected to arrive each month, split between what's already contracted and what still has to be won. For a subscription product each month is built from the previous one: last month's revenue, plus the customers arriving, minus the ones leaving.
2. Costs — outgoing money grouped by type. Salaries, marketing spend, infrastructure such as servers, plus office and software subscriptions. At an early company salaries run 60 to 70 percent of the total, which means the hiring plan is most of the cost forecast.
3. Headcount plan — which month each team adds which people. It gets its own rows because people are money: move one July start date to April and three months of that salary ripple across the whole table.
4. Cash balance — the opening balance for the month, plus what came in, minus what went out. This row is the conclusion of the table. It shows how far the balance dips, in which month, and whether it hits zero at all.
5. Plan versus actual — once a month closes, the real number goes in next to the estimate and the gap stays visible. Forecast $60,000 of revenue for April, collect $50,000, and that $10,000 is the evidence for revising the months ahead.

Many companies add one more thing: two or three versions of the same table, one for things going well and one for things going badly. Put a plan where revenue lands as expected next to one where it lands at 60 percent, and the bad column shows the month the balance runs out — which in turn fixes the deadline for doing something about it.

The founder or the finance lead builds it, and three groups read it: the leadership team, investors, and team leads. At a 45-person company like Pagedeck there's no finance organization yet, so Grace keeps the table herself. As the company grows, a financial planning and analysis (FP&A) function takes it over — a team whose job is building the plan and comparing it against results.

Each reader goes to a different row.

1. Leadership reads the cash balance first. How much is left in which month is the starting point for hiring, fundraising and cutting decisions.
2. Investors read the revenue line and the reasoning behind it. In a fundraise they press harder on where the numbers came from than on the numbers themselves.
3. Team leads read their own rows, because next quarter's budget and open positions are cut out of the forecast.

The normal revision cycle is monthly. When the month closes, the actuals go in and the remaining months get redrawn. Update only once a quarter and you end up making decisions from a table that's two months out of date.

"The forecast has runway dropping to five months in December. We need to open a round before that."

Using the cash balance row to fix a date for raising money.

"Put those three hires into the forecast and let's talk again."

Turn the request into numbers before arguing about whether it's a good idea.

"Q3 came in more than 20 percent off the forecast. I think the assumptions need another look."

The problem isn't being wrong — it's a signal that the thinking behind the table has gone stale.

At Pagedeck, Grace revises the forecast in the first week of every month. Last month's actual revenue and costs go in, the deals Sofia expects to close get reflected, and any hire whose start date slipped moves back a column. Then she turns the change in the cash balance row into a single page for the leadership team.

That page sets the range of decisions for the month. If the balance improved on last month, a held position gets released. If it got worse, marketing spend comes down or a hire moves out a quarter. Rather than remaking every decision from scratch, the team adjusts by exactly as much as the table moved.

The most common mistake is reading a forecast as a prediction that has to come true. There are numbers on the page, so judging whether the numbers were right feels like the obvious thing to do. But the value here is comparison, not accuracy. It's an instrument for putting a table with three extra hires next to a table without them and seeing whose balance runs out first. Start grading it on accuracy and whoever builds it will write down numbers that are easy to hit, which produces a safe table that supports no decision at all.

Treating it as a document you write once is just as common. It gets built carefully for a fundraise and then sits untouched for six months, mostly because it takes real work to assemble and nobody wants to touch it again. The payoff, though, comes from the moment it meets reality. Skip that comparison and an assumption that broke in April keeps inflating May and June, and half a year later the balance reveals that the table has been describing a different company since the spring.