Budget

At a small company every purchase passes through one person. The founder reads each request, approves it, and that is the entire control system. With ten people it's also the most accurate system available, because that one person knows what each request is for.

Past forty people the same arrangement breaks in two places at once. Approvals back up first. Twenty requests a day arrive in front of someone who lacks the context to judge most of them, so in practice they read the amount and wave it through. Speed goes at the same time. Waiting four days for sign-off on a $50 collaboration tool becomes normal, and the delay costs more in stalled work than the tool costs to buy.

The larger failure shows up when individually sound decisions land together. Marketing decides to put more behind an ad that's performing. Engineering decides to add servers because response times are slipping. Sales decides to take a booth at an industry event. Each of those is defensible on its own merits. Land them in the same quarter and the month's outflow runs nearly $75,000 above what anyone pictured, and that fact surfaces when next month's statements arrive. Checking burn rate after the fact can't call money back.

A budget moves the judgment forward in time. Instead of asking about each purchase, it cuts the period and the amount in advance and stops asking inside those lines. It isn't more control. It's the same control exercised once instead of twenty times a day.

Set a spending ceiling per team for a quarter or a year, let the team spend freely under that line, and require approval only when crossing it. Setting the ceiling usually takes five steps.

1. Decide what the whole company will let out during the period. The starting point isn't the revenue forecast but the cash on hand and the runway you intend to hold. With $2.7 million in the bank and eighteen months to cover, the ceiling is $150,000 a month, or $450,000 a quarter.
2. Split the total into categories. Payroll, software and tools, advertising, servers and infrastructure, events and travel is the usual set. At a software company payroll runs around 70 percent of everything.
3. Assign the categories to teams. This is where the total always comes up short. Add up what teams ask for and it typically exceeds the ceiling by about 30 percent, and closing that gap is most of what a budget meeting consists of.
4. Put actual spending next to the budget every month and read the gap. A team given $60,000 for the quarter that has spent $52,000 in two months has $8,000 left for the third. Do this comparison only at quarter end and the moment to act has already passed.
5. Decide in advance what happens to what's left. Unless you settle whether unspent money carries over or disappears, December fills up with purchases nobody could otherwise justify.

The path for crossing the line gets defined alongside it. Approval thresholds by amount are the common form: a team lead signs off up to $1,500, their manager up to $15,000, the founder above that. Without that table every request eventually routes back to one person again.

A yearly budget gets set while the annual plan is being written. Goals come first and the money attaches to them, which is why reading a budget tells you what a company decided to do this year. Advertising doubled and hiring flat means this is a year for selling what already exists rather than building more of it.

Payroll often gets managed in people rather than money, and headcount is what holds that slot. Salaries vary enough that an amount doesn't tell you how many more people you can hire, and unlike a tool subscription a hire keeps costing the same amount every month for years.

Inside a team it becomes the test for adding a tool. A design lead weighing a second prototyping subscription checks the remaining line before checking whether the tool is any good. With budget left, the approval is a formality; without it, the purchase waits a quarter or something already in use gets cancelled to make room. When conditions shift sharply mid-quarter, the budget itself gets redrawn, and the document sitting next to it during that work is the financial forecast.

"That's not in this quarter's budget, so it either waits or something else comes out."

The total is fixed, so anything new has to displace something of the same size.

"We have the budget, we just don't have the headcount."

The money exists but a separate cap on people blocks the hire.

"We've been over budget two months running, so it's worth a look."

Actual spending is above plan and the gap isn't closing on its own.

Pagedeck runs on a quarterly budget of $690,000. Payroll takes $480,000, marketing $60,000, servers and infrastructure $46,000, and the rest covers tools, events, and travel.

Halfway through last quarter Ravi brought a proposal to increase ad spend, since the ads were performing. By then $52,000 of the $60,000 marketing line was already gone. Rather than raise the total, Grace chose between pulling next quarter's allocation forward and dropping one event sponsorship. Without a budget, that conversation would have started a month later, in front of a statement.

The most common belief is that a budget has to be spent in full. It comes from real experience: leave money on the table and next period's allocation gets cut by roughly that amount. But a budget is a ceiling, not a target. A team that was given $60,000, spent $46,000, and got the same result has added $14,000 back to the runway. Leave the belief in place and the last month of every quarter fills with thinly justified spending, and the next budget gets built on those inflated numbers.

The second mistake is assuming that cutting the budget extends runway by the amount cut. The arithmetic feels obvious — money not spent stays in the bank. In practice most of the total is payroll, and cancelling every tool, ad, and event outside it reaches only about 30 percent of the number. Saving $210,000 out of a $690,000 quarter buys roughly six more weeks. Not knowing this, an organization spends weeks trimming small line items while the conversation that actually moves the number, the pace of hiring, keeps getting deferred.