a founder spent four hours on a saturday going through every saas subscription his team had ever signed up for. notion, slack, linear, intercom, segment, an ATS, a CRM, a feedback tool, a calendly seat for every employee. he canceled nine of them. he was proud. he told the team monday morning.
it saved $9k a month. about half a week of runway. meanwhile a senior engineer he'd hired in january was doing work the rest of the team didn't actually need yet. nobody had said anything because firing is hard and the engineer was good. the founder didn't look at that line.
this is the central evasion of early-stage burn management. everyone wants to cut the small thing. nobody wants to cut the big thing. the small thing is fun to cut because it doesn't hurt anyone. the big thing is the one that actually changes the math.
the math, with the noise stripped out
a typical $400k/month burn at a seed-or-series-a startup breaks down roughly like this. $300k is salaries plus benefits plus payroll taxes. $40k is infra — usually aws, sometimes gcp, plus a smattering of data warehouses and observability tools. $30k is software and tooling. $30k is everything else: legal, accounting, office, travel, contractors.
salaries and benefits are 75% of total burn. cut tooling by a third and you save $10k a month. cut one mid-level engineer at $200k fully loaded and you save $17k a month. one hire. one decision. nearly twice the impact of the entire saas spring cleaning.
four hours of work bought him three days. two days of work, plus the worst week of his year, would have bought him six weeks. that's not a marginal difference. that's the gap between making it to a series A and not.
why founders avoid the only line that matters
three reasons, in increasing order of honesty.
one — sunk cost. you spent three months recruiting that engineer. you flew them out, sold the vision, beat a competing offer. firing them six months in feels like admitting you were wrong. it is. that's fine. the company doesn't care about your recruiting effort, it cares about cash on hand.
two — identity. founders who got their first hires through hard work build emotional attachment to the team. cutting feels like cutting yourself. the trick most survivors learn: the team you have today is not the team that gets you to the next milestone. they overlap, but not perfectly. the question isn't "are these people good?" — they almost always are. the question is "do we need this exact composition right now?"
three — optics. cutting headcount looks bad. layoffs leak. recruiters know. it signals that something is wrong. founders avoid the signal until the signal is unavoidable, which is also the moment when the cuts become deep and sudden instead of small and surgical.
the discipline most founders only learn after they almost die
paul graham wrote in 2015 that the right time to fire is the moment you first think about it. the math underneath that is just headcount math. the longer you wait, the deeper the cut has to be, because each additional month of payroll is another month of cash gone.
the founders who survive 2026 aren't the ones who never make a wrong hire. they're the ones who reverse a wrong hire in month four instead of month ten. the cost of a six-month-late firing on a $200k role is roughly $100k in cash plus the morale tax of carrying someone everyone knew wasn't working. that's real money on a balance sheet that doesn't have a lot of it.
the same logic runs in reverse on hiring. every approved role is a decision to spend $250k of runway over the next twelve months. founders treat that decision like it's reversible and small. it's neither. by the time you realize the role wasn't the right one, you've burned three months of salary and you owe the person another two months of severance to do the right thing on the way out.
the saas audit is a way to feel productive without making a hard decision. the headcount review is the hard decision dressed up as a meeting.
what the right process actually looks like
three things, in order.
monthly headcount line review. on the first of the month, the founder looks at every name on payroll and asks: "if i were starting today with current revenue, would i hire this role next?" not "is this person good?" — they almost always are. "is this role the next dollar of payroll i'd spend?"
a 90-day moratorium on roles you can't justify in one sentence. if you can't describe what a role does in one sentence and connect it to revenue or product within 30 days, the role waits. most founders find half of their open reqs fail this test. the half that fail were the ones the team wanted, not the ones the business needed.
a written re-hire-this-person test for every role. before any cut, ask whether you'd hire this exact person, into this exact role, today. if the answer is yes, keep them and find the cut elsewhere. if the answer is no, the conversation is overdue.
how zift handles this
zift tracks headcount cost against runway every fifteen minutes, pulling directly from your payroll provider. on monday morning the briefing shows: total comp by function, change from last month, runway impact of each open req still being recruited. no spreadsheet. no waiting for accounting.
if you're a finance lead at a series A team and need this with multi-entity payroll consolidation, zift handles that too.
the saas audit feels like ops. the headcount review is ops. one extends runway by a rounding error, the other extends it by a quarter — the only people who confuse them are the ones who haven't had to make the call yet.