A few years ago, a friend of mine wrote the authentication module for a fintech platform. Twelve lines of Go. Those twelve lines sit between user credentials and roughly $2 billion in transaction volume every quarter. He makes $160,000 a year. The enterprise sales rep who closed the bank that processes most of that volume makes $340,000 in a good year. This isn’t an anomaly. It’s the entire story of how tech actually compensates people, and pretending otherwise is just mythology.
1. Markets Price What They Can Measure
Sales compensation is almost perfectly legible. A rep closes a $1.2 million contract. The commission structure spits out a number. Causation is right there on the invoice. Engineering compensation is priced against a market rate for a job title, because nobody has a clean way to say “this person’s work generated $X.”
This isn’t a conspiracy. It’s a measurement problem. The authentication module my friend wrote has never failed. There are no incidents attributed to it. In accounting terms, it has produced exactly zero events. The sales rep produced a signed contract, a number on a spreadsheet, a thing that shows up in the quarterly report. Companies reward what they can point to, and they can point to revenue with embarrassing ease compared to the absence of a catastrophic security breach.
2. Risk Is Priced Wrong, and Companies Know It
The economic logic of sales compensation is that you’re paying for demonstrated, attributable revenue. The economic logic of engineering compensation is that you’re paying for labor inputs, not outputs. This creates a systematic underpricing of risk.
If that authentication module fails, the company doesn’t lose a contract. It loses trust, regulatory standing, possibly its banking license. The downside is enormous. But because the downside hasn’t materialized, it doesn’t get capitalized into the engineer’s salary. Sales is rewarded for upside it directly creates. Engineering is not rewarded for downside it quietly prevents. The asymmetry isn’t fair, but it’s also not irrational from a CFO’s perspective. You pay for what you can account for. Nobody agrees what a software project actually cost, so how would you even begin to price the avoided cost of something that never happened?
3. Proximity to Revenue Is the Actual Job Title
Here is the uncomfortable truth about corporate compensation: your salary is largely a function of how many steps removed you are from a closed deal, not how important your work is to the company’s survival. Sales is zero steps. Marketing is one. Product is two or three. Infrastructure engineering is four or five, on a good day.
This is why a mid-level account executive at a SaaS company can out-earn a senior engineer, even when the SaaS company’s entire product is software. The salesperson is defined by revenue generation. The engineer is defined by cost center. Even companies that genuinely understand and respect engineering tend to compensate according to this structure, because compensation benchmarks are set by comparing to peer companies, and peer companies have the same structure.
4. The Explainability Premium Is Real and Underappreciated
The sales rep who “explained it to the bank” is doing something that isn’t as simple as it sounds. Enterprise sales into financial institutions is extraordinarily hard. Security reviews, compliance questionnaires, procurement committees, legal back-and-forth, relationship maintenance across personnel turnover. The rep who closed a $1 million ARR banking contract probably worked that deal for nine months.
There’s a genuine skill premium being paid there, and it’s the skill of navigating institutional complexity and making technical products legible to people whose job is to be skeptical of vendors. That skill is scarce and it’s valuable, and it’s not crazy that it commands high compensation. The problem isn’t that salespeople are overpaid. The problem is that the engineer who made the product worth selling is invisibly underpaid by comparison, and nobody in the org chart has any incentive to correct that.
5. Equity Was Supposed to Fix This, and It Mostly Didn’t
The standard response to “engineers are underpaid relative to their impact” has always been equity. Engineers get options. If the company wins, so do you. This argument has some validity at the very early stage, where a founding engineer might accumulate a meaningful stake. By the time a company is large enough to have an enterprise sales team, the equity math has usually stopped working in the engineer’s favor.
Late-stage options are often priced close to current valuation, with enough vesting cliffs and preference stacks above them that the actual payout on an exit can be disappointingly small. What you actually own when a startup sells you software applies equally to what you own when a startup sells you equity. The instrument matters as much as the number. Senior salespeople, meanwhile, increasingly negotiate for RSUs and bonuses that convert to cash on predictable timelines, regardless of exit events.
6. The Companies Getting This Right Are Doing Something Simple
A small number of companies have started compensating engineers on something closer to impact. Not perfectly, but meaningfully. They do it by connecting engineering work to product metrics that connect to revenue, then being honest about that chain in comp reviews. A payments infrastructure team that improves transaction success rate by a fraction of a percent gets credit for the revenue that margin represents. It requires someone in leadership who can hold both the technical and financial model in their head simultaneously, and is willing to fight internal benchmarking inertia.
This is genuinely rare. But it exists, and it works. The engineers don’t need to earn what the top salespeople earn. They need the gap to reflect something closer to actual value rather than organizational proximity to a signed contract. Getting there requires companies to solve the measurement problem, not just acknowledge it. Most won’t bother until they lose enough critical engineering talent to someone who did.
