Here's the arithmetic, and it's simple enough that it's slightly strange how rarely it gets stated.
A dollar of new revenue does not become a dollar of profit. It arrives net of what it cost to earn — the sales commission, the infrastructure to serve it, the support it generates. If the business runs at a 20% net margin, that dollar contributes twenty cents.
A dollar of eliminated cost contributes a whole dollar. There is nothing to subtract.
So at a 20% net margin, saving a dollar is worth what earning five is worth. The multiplier is just the reciprocal of the margin.
Net margin | Revenue needed to match $1 of saving | Where you see this |
|---|---|---|
50% | $2 | A mature, dominant software business |
20% | $5 | A healthy, profitable company |
10% | $10 | Most operating businesses |
5% | $20 | Retail, logistics, thin-margin services |
Negative | Undefined — every marginal sale deepens the loss | Most venture-funded companies before profitability |
Read the bottom rows and the uncomfortable implication appears: the multiplier is largest exactly where companies are least profitable. The businesses with the most to gain from cost discipline are the ones least likely to have anyone doing it, because everyone is focused on growth. The go-to-market half of a dollar saved is worth more than a dollar earned is handled in more depth by XenGrowth's revenue operations work.
In a company losing money, a marginal sale makes the loss bigger and a removed cost makes it smaller. The multiplier isn't just large there — it's the only lever pointing the right way.
Two ways this gets abused
Before using this argument, know how it's misused, because someone in your finance team will spot it instantly and you'll lose the room.
The first is comparing a one-off saving to recurring revenue. Deleting some orphaned storage and saving $4,000 once is not equivalent to $20,000 of annual revenue. The honest comparison is recurring against recurring: a saving that persists every month, against revenue that persists every month. Most infrastructure savings genuinely are recurring, which is what makes them strong — but say so explicitly rather than letting the ambiguity work in your favour.
The second is using net margin when gross margin is the relevant figure, or the reverse. If your saving is in cost of goods sold — hosting, egress, third-party APIs — it moves gross margin, and the cleanest statement is in those terms. Reaching for the more dramatic net-margin multiplier when a gross-margin one is the accurate frame is the sort of thing that gets your next proposal read more sceptically. If the operations side of this is the part you are stuck on, The XenGrowth resource library is the better reference.
A worked example, with the arithmetic shown
Take a company doing $10m of annual revenue at a 75% gross margin and a 10% net margin — a plausible mid-stage software business. Infrastructure is running at $1.2m a year, which is 12% of revenue and sits inside the 25% of revenue that cost of goods sold consumes.
An engineer finds $180,000 of recurring annual saving: a retention policy nobody set, a set of over-provisioned instances, and one access pattern generating avoidable egress. That is 15% off the infrastructure line. Stated as a dollar figure it sounds like a nice piece of housekeeping and will be received as one.
Stated properly it is 1.8 points of gross margin, moving the company from 75% to 76.8%, permanently, with no further spend. At a 10% net margin it contributes as much to profit as $1.8m of new revenue would — 18% of the entire company's current annual revenue, found by one person inside a cloud console. And unlike $1.8m of new revenue, it does not have to be won again next year, does not require hiring anyone, and carries no additional cost to serve.
None of those figures is exotic and none of the work is technically difficult. What makes the example unusual is only that somebody computed the second and third paragraphs rather than stopping at the first, which is the entire skill this post is about.
The honest limit of the argument
Cost reduction has a floor. You can save at most everything you currently spend, and long before that point you start removing things the business needs. Revenue has no ceiling, and it compounds — a customer retained this year expands next year, refers someone the year after, and Benchmarkit's data puts net revenue retention around 106% on average with top performers above 120%, which is growth arriving without any new acquisition at all.
So nobody saves their way to being a large company. The multiplier is an argument about relative effort for a given increment of profit, not a strategy. Anyone who reads it as "cost cutting beats growth" has taken it further than it goes, and the version of this post that says so is more useful than the version that doesn't. XenGrowth on AI agents and marketing automation approaches this from the AI agents and marketing automation side.
What it does establish is that the two are wildly asymmetric in effort. Earning $500,000 of new annual revenue takes a sales team, a quarter, a pipeline and some luck. Eliminating $100,000 of annual infrastructure cost at a 20% net margin does the same thing for profit and may take one engineer three weeks. Both are worth doing; only one of them is available to you personally on a Tuesday.
Why this is specifically an engineering argument
Because of where the cost sits. Infrastructure, third-party services, egress and support are cost of goods sold, which is above the gross profit line — and they are almost entirely determined by engineering decisions.
Which makes this the one line on a profit and loss statement that engineering can move without anyone else's cooperation. Sales cannot reduce your egress bill. Finance can ask about it but cannot change a retention policy. Marketing has no view. It is yours, and almost nobody treats it as an asset. XenGrowth on AI search, GEO and discovery covers the AI search, GEO and discovery side of this.
Recurring, so the saving persists every month without further work — unlike a sale, which has to be made again next year
Unilateral, requiring no other department's participation, budget or approval
Compounding in a modest way, since a lower cost base means the same margin holds as you grow rather than deteriorating
Fast, on the scale of weeks rather than the quarters a revenue initiative needs
Measurable precisely, in a currency the CFO already tracks weekly — which is more than can be said for most engineering outcomes
Why nobody is already doing this
Given arithmetic this favourable, the obvious question is why cost work is chronically under-resourced in most engineering organisations. The answer is not that people are unaware of the multiplier. It is that the incentives run the other way at every level.
Cost reduction produces no artifact. There is no demo, no feature in the release notes, no customer thanking anyone. A quarter spent on it looks, in every system that tracks engineering output, like a quarter in which less was delivered — and the benefit accrues to a line on a statement most of the team never sees. Compare that with shipping a feature, which is visible, attributable and immediately rewarded, and the allocation of attention stops being mysterious.
There is a second-order effect too. Because savings are invisible, they are also unclaimed: the engineer who removed $180,000 of recurring cost usually cannot point to it at review time, because nobody recorded the counterfactual. This is worth fixing for entirely self-interested reasons. Write the number down when you find it, state it in margin points, and send it to somebody in finance. That single email converts an invisible contribution into a documented one, and it is the cheapest career action available in this entire post.
How to present it
State the recurring annual figure, not the monthly one and not a one-off. Annualised recurring is the unit finance thinks in
Convert to margin points if you can get revenue. "This is 1.5 points of gross margin" is a stronger sentence than any dollar amount, because it names a metric the board already watches
Give the revenue equivalent using the company's actual margin, and say which margin you used. Showing your working is what stops it reading as a rhetorical trick
Be explicit that it recurs and does not need to be re-earned. This is the property that distinguishes it from a sales result and it is the one people forget
Say what it costs to get — three weeks of one engineer, with a named opportunity cost. A proposal with no cost attached reads as a claim rather than a decision
Weak framing | Strong framing |
|---|---|
"We could save some money on AWS" | "$96k recurring annually, about 1.4 points of gross margin" |
"This is wasteful" | "At our net margin this is equivalent to roughly $480k of new revenue" |
"It'd take a few sprints" | "Three weeks of one engineer, against a permanent recurring saving" |
"Costs are getting out of hand" | "Infrastructure has grown 40% while customers grew 15%" |
"We should do a cost sprint" | "Here are three recurring savings and the two I recommend skipping" |
The bottom row matters more than the arithmetic does. Bringing the items you are recommending against is what makes the rest of the list credible — and credibility is the actual scarce resource in these conversations, not analysis.
Further reading from XenGrowth
The XenGrowth resource library — what you'll learn: how the commercial side of this work is run, across search, automation and revenue operations.
XenGrowth on AI agents and marketing automation — what you'll learn: how the teams who own AI agents and marketing automation plan and measure it.
XenGrowth on AI search, GEO and discovery — what you'll learn: how the teams who own AI search, GEO and discovery plan and measure it.
Where this work meets go-to-market
The operational playbooks that sit alongside a dollar saved is worth more than a dollar earned live with the XenGrowth practice.
