ยท 7 min read
The CFO and the CTO: the conversation that decides whether a right-sizing works.
A right-sizing works or fails in a single conversation, and it is usually the one that never properly happens: the CFO and the CTO, at one table, before any name is on a list. The CFO brings how much has to come out and by when. The CTO brings what each part of the spend actually buys. The cut works when those two views become one page. It fails when the CFO hands down a number and the CTO hands back an org chart, and the two are never reconciled.
What you have probably noticed
The target arrives as an email. Finance needs engineering down by some figure, or to some percentage of revenue, by the end of the quarter. You read it twice. Then one of two things happens. You build a list that hits the number, because that is what was asked, and you quietly carry the knowledge that two of the names on it are holding up the thing that ships. Or you write back with all the reasons the number is wrong, and finance reads that as engineering protecting itself, which is what engineering leaders always seem to do the moment a cut is proposed.
Either way the decision gets made in two separate heads and stapled together at the end. The CFO has a number with no map. You have a map with no deadline. Nobody put them on the same page, and the cut that comes out of it is a compromise between two people who never actually spoke.
What it actually means
A right-sizing is a translation between two currencies that do not convert cleanly. The CFO's currency is money over time: runway, monthly burn, the cash-out date, how much patience the board or the bank has left. The CTO's currency is delivery and risk over time: what ships this quarter, what breaks if a team goes, who leaves, how many months a rebuild costs if you cut the wrong thing.
Each number is real and each is blind. The CFO's target is correct about the cash and blind to which cut saves money cleanly and which buys a six-month hole in delivery. Your org knowledge is correct about the work and blind to the deadline that makes some of your preferred options impossible. A cut designed from only the finance side removes delivery capacity by accident. A cut designed from only the engineering side misses the date the company actually has to hit. The right-sizing works only when the two views are merged into one: the spend laid against what it buys, with the cash deadline drawn over the top.
What to do about it
- Agree the target and the deadline first, in one sentence, before any names. "EUR 120 000 out of the monthly run rate by the end of the quarter, because at the current burn the cash runs out in nine months" is something you can act on. "We need to reduce costs" is not. If the CFO cannot state it as a figure and a date, the conversation is not ready, and the cut will drift into a round of guessing. Settle whether the saving has to be permanent or one-off while you are there, because a severance bill and a lower run rate are different promises.
- Put the full org cost next to the delivery map on one page. The CFO brings the fully loaded cost of every team, not salaries: the number that includes employer charges, tooling, and the contractors nobody counts. You bring the same teams sorted by what they do, in the three piles a board already understands. Run keeps what exists alive. Deliver is what ships this quarter. Bet is capacity held against a future that has not arrived. That split, and how to price each pile, is the whole method behind the burn plan a board actually wants. The single page is usually where both of you discover you were arguing about different things.
- Cut the expired bets first, together. The spend you both agree is buying a future that is no longer coming is the cleanest money in the building and the smallest delivery cost. Naming those together, out loud, also sets the tone for everything after it: this is a joint decision about the company, not finance overruling engineering.
- Trade anything that touches Deliver in the open. You price the delay a given cut would cause. The CFO prices whether the cash position can absorb that delay. Then you decide together, trade by trade. A role that saves a set amount a month but adds months to a rebuild if the market turns is a decision both of you sign, not a line finance strikes and engineering discovers after the fact.
- Write down what you are not cutting, and why. The cut is one meeting. The pressure is permanent. Without a short record of the teams you protected and the reason you protected them, next quarter's target erodes them one role at a time, and in a year you have made a cut that nobody ever decided to make.
I have sat in both versions of this meeting. In one, the CFO had already settled the number and the CTO had already drawn the reorg, and the ninety minutes were two monologues that changed nothing; the cut that followed hit the target and lost two of the people the company most needed. In the other, the two of them built one page in the room, the finance cost on the left and the delivery map on the right, and by the end the cut was close to obvious and neither of them owned it alone. The difference was not the quality of the numbers. It was whether they were on the same page before anyone picked a name. What happens in the first 30 days after the cut decides whether it holds, but this conversation decides whether it was the right cut at all.
When not to reach for this
This conversation assumes the number is a real financial constraint. Sometimes it is not. If the target exists because the board wants a headline rather than because the cash demands it, you are not in a right-sizing conversation, you are in a message-management one, and treating it as financial will cost you credibility when the saved money goes nowhere visible. Say that plainly to the CFO instead of executing a cut you both know is theatre.
The conversation also fails when the CFO and the CTO disagree about where the company is going. If finance is planning for flat revenue and engineering is building for three times the traffic, no cut will hold, because the two of you are sizing different companies. That is a CEO conversation first, and no amount of spreadsheet detail substitutes for it. And if there is no CFO, which is the case in a lot of owner-led companies, the owner sits in that seat and the bar is the same: bring the real cash deadline, not a general sense that costs feel high. The honesty the conversation needs does not depend on anyone's title. When the shape of the org, not its size, is the problem, that is a restructuring question, answered differently from a one-off cut, and building the map itself is the work a structured diagnosis exists to do. It is the same ground under any honest right-sizing decision: you cannot size a team well until you can say, on one page, what its spend is buying.
Frequently asked
What should the CFO and the CTO agree on before a right-sizing?
One number and one date, before any names. The target has to be stated as a euro figure out of the monthly run rate and a deadline it must land by, tied to the real reason: extend runway to a milestone, hit a margin, or absorb a revenue miss. Only then do you put the fully loaded cost of every team next to a map of what each team ships. A right-sizing works when the finance view and the delivery view become one page before anyone is on a list, and fails when a number is handed down and an org chart is handed back without the two ever being reconciled.
What numbers does the CFO bring to an engineering cut, and what does the CTO bring?
The CFO brings the fully loaded cost of every team, not salaries: the figure including employer charges, tooling, and the contractors nobody counts, plus the target in euros, the deadline, and whether the saving has to be permanent or one-off. The CTO brings the delivery map: each team sorted into Run, Deliver, and Bet, the key-person risk you cannot lose without a hole, and the cost of each possible cut translated back into the CFO's currency, meaning how many months a rebuild takes if the cut turns out wrong. Neither set of numbers is enough on its own.
Why do engineering cuts set purely by a finance target often fail?
Because a finance target is correct about the cash and blind to what the money buys. A cut designed only to hit a number removes delivery capacity by accident, since finance cannot see which teams ship and which hold optionality the company no longer needs. The cut lands, the target is met on paper, and a quarter later delivery has a hole nobody chose. The fix is not to ignore the target. It is to merge it with the delivery map in one conversation, so every trade against shipping is made in the open by both people.
How do you right-size engineering without a CFO?
The owner sits in the CFO's seat and the bar does not change. Bring the real cash deadline, not a general sense that costs feel high: how much has to come out of the monthly run rate, by when, and why. Then lay it against a map of what each part of engineering actually ships. The conversation that decides whether a cut works is the same whether the other chair holds a CFO or the owner; the honesty it needs does not depend on a title.
Written from patterns I've seen across engagements. If any of it sounds too specific to be general, that's why.