The Great Flattening: Engineering Manager vs. Staff Engineer in 2026

For two decades, the standard advice for an ambitious senior engineer was simple: if you want more scope, more money, and more influence, become a manager. That advice is no longer safe. Between 2022 and 2026, the largest employers in tech systematically deleted management layers, widened every surviving manager's span of control, and redirected career growth toward the staff-plus individual contributor track.

If you are a senior engineer deciding which track to interview for -- or an engineering manager wondering whether your job category is quietly disappearing -- this is not a question of temperament anymore. It is a market question, and the market has data. This guide walks through what the flattening did to manager headcount, how the comp math between the tracks works at real companies, and how to run a search on either track -- or both -- in 2026.

Big Tech Is Deleting the Manager Layer

The flattening started as a Meta idiosyncrasy and became an industry consensus. In his March 2023 "Year of Efficiency" memo, Mark Zuckerberg announced Meta would "make our organization flatter by removing multiple layers of management" and asked "many managers to become individual contributors." His three-word justification -- "flatter is faster" -- became the template every other CEO would reach for.

They reached for it quickly. In September 2024, Amazon CEO Andy Jassy mandated that every organization increase its ratio of individual contributors to managers by at least 15% by the end of Q1 2025, writing that "having fewer managers will remove layers and flatten organizations." In December 2024, Sundar Pichai told a Google all-hands that the company had cut manager, director, and VP roles by 10% -- some laid off, others converted to individual contributor roles.

The pattern repeated across the industry. Microsoft's May 2025 layoff of roughly 6,000 people was explicitly designed to reduce management layers, followed by another 9,000 cuts in July. Intel CEO Lip-Bu Tan told employees in July 2025 that Intel had reduced management layers by roughly 50%. Dell set minimum spans of control -- at least 15 direct reports for VPs and 20 for directors. Amazon's October 2025 cut of 14,000 corporate roles -- which grew to 16,000 by January 2026 -- was framed by its HR chief as "reducing bureaucracy, removing layers."

And it is still accelerating. When Coinbase cut 14% of its workforce in May 2026, CEO Brian Armstrong explicitly replaced "pure managers" with "player-coaches," capped the org chart at five layers, and set a 15-to-1 employee-to-manager ratio. Gartner predicts that through 2026, 20% of organizations will use AI to flatten their structure, eliminating more than half of current middle-management positions in the process.

What Flattening Did to the EM Job Market

Company announcements are one thing. The aggregate labor-market data shows what they add up to -- and for anyone planning to interview for engineering management roles, the numbers are sobering.

The number of managers in the US workforce dropped 6.1% between May 2022 and May 2025, per Live Data Technologies. Middle managers went from roughly 22% of annual layoffs in 2018-2022 to 29% of all layoffs in 2024, and middle-management job openings have fallen more than 40% since 2022, per Revelio Labs data cited in the same Korn Ferry analysis. All of this sits inside a tech hiring market that remains frozen: Indeed's Hiring Lab reports US tech job postings were still 36% below early-2020 levels as of mid-2025.

The dynamic is not confined to Big Tech. Gusto's analysis of roughly 8,500 small and midsize businesses found the ratio of individual contributors per people manager roughly doubled from about three in 2019 to about six, with manager hiring down nearly half since 2022. And Gergely Orosz's The Pragmatic Engineer, surveying the 2026 market, concluded that "there are fewer engineering managers for each engineer across the industry, and fewer VP of engineering and director of engineering posts at Big Tech."

Read those numbers the way a hiring market reads them: more EM candidates than ever, competing for fewer seats than ever, in a market where the fallback -- taking an IC role -- usually involves a pay cut and a skeptical interview loop.

The Manager Job That Survived Is a Different Job

Deciding between the tracks requires understanding that the EM role you are comparing against is not the 2019 version of the job. The surviving role is wider, more technical, and measurably more stressful.

Spans of control widened at every level. Gallup found the average number of direct reports per manager jumped from 10.9 in 2024 to 12.1 in 2025. Pave's analysis of 257,000+ managers in its compensation database shows spans rising at every management level between late 2023 and late 2025. In engineering specifically, SignalFire's State of Talent report counts roughly 12 engineers per engineering manager at big tech companies (up 14% from 2019) and 15 at early-stage startups (up 34%). LeadDev's 2025 leadership survey found 65% of engineering leaders reporting expanded responsibilities and 40% managing more direct reports than before; only 3% saw their scope shrink.

The role also re-acquired a coding requirement. The "player-coach" framing that Coinbase made explicit -- "Every leader at Coinbase must also be a strong and active individual contributor" -- had been building for years: when The Pragmatic Engineer surveyed the post-ZIRP landscape, Lyft's VP of engineering put the new expectation bluntly: "Engineering managers are expected to architect, manage, debug, and code."

The cost of the widened role shows up in engagement data. Gallup found manager engagement fell from 30% to 27% in 2024 while IC engagement held flat, and more than half of managers report burnout per Microsoft's Work Trend Index. Meanwhile the people being managed feel the absence of management: in Korn Ferry's survey of 15,000+ professionals, 41% said their company had cut manager-level roles, and 37% said they felt directionless as a result. None of this means you should not pursue management -- it means the job you would be pursuing is bigger, more technical, and less supported than the one your current manager signed up for.

The Comp Math: Where the Manager Premium Actually Lives

Here is the part most senior engineers get wrong: the flattening did not erase the manager pay premium at the top of the market -- but it changed where that premium lives and how reliable it is.

Start with the aggregate numbers. Levels.fyi's 2025 End of Year Pay Report -- built on 245,000+ submissions across 5,000+ companies -- put US Staff Engineer median total compensation at $457K, up 7.5% year over year, the largest jump on the IC ladder. Engineering manager compensation grew even faster, at 9.6%. Scarcity cuts both ways: fewer manager seats also means the survivors are paid well.

The level-matched comparison at specific companies is where the decision gets concrete:

Three observations from that table. First, at a matched level, the manager still earns a premium -- roughly 14% at Google L6, and about $110K at Meta M1 versus E6. Second, that premium is almost entirely equity, not salary: Google L6 EMs and Staff SWEs have nearly identical base pay ($271K vs. $284K), but the EM averages roughly $93K more in annual stock. The same pattern holds at Meta. Third, the premium is conditional on getting and keeping a manager seat -- the exact thing the market is making harder. A Staff or Principal IC clears $600K+ at these companies without betting on a shrinking role category, and SignalFire's 2026 report concludes that "top-of-band staff and principal packages at the Tech Majors now rival or beat director pay, reversing the management premium that defined tech careers for two decades."

For calibration outside the Levels.fyi bubble: the Bureau of Labor Statistics puts the median for computer and information systems managers at $171,200 and software developers at $133,080 (May 2024), with the top 10% of developers above $211,450 and developer employment projected to grow 15% through 2034. Even on the government's conservative numbers, a top-decile IC out-earns the median IT manager.

How to Decide Which Track to Interview For

Charity Majors' essay "The Engineer/Manager Pendulum" remains the sharpest framing of this decision: "management is NOT a promotion... management is a change of profession." Her observation that the best line managers are "never more than 2-3 years removed from hands-on work" has aged from contrarian advice into literal company policy -- it is essentially what Coinbase now mandates.

The pendulum framing converts a career-defining decision into a two-to-three-year decision, which is the right size. You are not choosing your identity; you are choosing your next tour. Concretely, weigh four factors:

• Market liquidity. Staff-plus openings are growing as a share of engineering hires while EM openings shrink. If you optimize for optionality and search speed in 2026, the IC track has more doors. The promotion-versus-job-hop math compounds this: fewer external EM seats means fewer chances to re-price yourself by moving.
• What the surviving EM job actually is. Twelve-plus reports, an expectation that you still architect and debug, and less support above you. If you want management because you imagine escaping the editor, 2026 is the wrong decade for that plan.
• Energy, honestly audited. The pendulum question is not "which title do I want?" but "which Tuesday do I want?" -- six one-on-ones and a calibration meeting, or six hours inside a gnarly design problem. The next generation has voted: in Robert Walters' survey, 52% of Gen Z professionals say they do not want to be middle managers, and 72% would prefer an individual route to advance.
• Reversibility. Swinging from IC to EM is structurally easy -- companies hand out management to strong seniors when a team needs it. The reverse swing is harder and gets harder every year you wait, which is the subject of the next section.

Making the EM-to-Staff Move Without Getting Down-Leveled

Thousands of engineering managers are making the return trip -- some by choice, many because the market chose for them. The move is entirely doable, but it has a failure mode: getting mapped to a lower IC level than your management scope deserves, a dynamic we cover in our guide to down-leveling.

In her follow-up essay on the pendulum, Majors names the fear directly -- "Can I still perform as an engineer? Will anyone hire me?" -- and offers the honest answer: "After two or three years of management, it's still pretty easy to go back to engineering," but the window narrows each additional year. If you have been managing for five-plus years and want back on the IC track, the deciding factor in your loop will be evidence of recent technical work: code you shipped, designs you authored, incidents you debugged -- not code you reviewed or roadmaps you approved.

Positioning matters as much as recency. The instinct is to present management experience apologetically. Invert that:

There is also a genuine tailwind worth citing in interviews: the agentic-coding era has made ex-manager skills newly valuable in IC roles. LeadDev reports that former engineering managers are in demand as individual contributors because breaking down work, delegating, and reviewing output -- the core of management -- maps directly onto orchestrating AI coding agents. One AI startup deliberately staffed its launch team entirely with former EMs. Decomposing ambiguous problems and verifying delegated work is now a hands-on technical skill, and you have years of it.

How the Interview Loops Differ

If you run searches on both tracks simultaneously -- which is increasingly common and entirely reasonable -- you need to prepare for two different loops with overlapping but distinct evaluation criteria.

Staff-plus loops are unstandardized, and that is your first problem to solve. Will Larson's StaffEng guide is blunt about it: the loop "might be the same exact interview you'd get for a Senior engineer role. It might be an engineering manager loop with a programming question added. It might be something else entirely." Before any staff-plus loop, get your recruiter to answer three questions: what formats and evaluation criteria apply, which interviews need specific preparation, and who will be interviewing you. A panel of mostly mid-level engineers rarely produces a Staff offer -- know that going in.

Larson's deeper critique is that most staff-plus loops are miscalibrated: they screen for "a senior engineer who's really fast at solving problems," even though staff engineers program less than seniors and are often slower at rote coding tasks. The well-designed loops evaluate self-awareness, judgment, collaboration, communication, and how you develop other engineers. Prepare for both realities: keep your algorithmic skills warm enough to survive a miscalibrated screen, and prepare the staff-level artifacts -- architecture narratives, influence-without-authority stories, technical strategy -- that the good loops actually score, as laid out in the senior interview rubric.

EM loops in 2026 assume the player-coach role. Expect a coding or architecture round that would have been rare in a 2019 EM loop -- if EMs are "expected to architect, manage, debug, and code," the loop tests all four. Expect deep scenario questions about running wide teams (12+ reports), operating without a supporting management layer above you, and doing more with flat headcount. The candidates winning scarce EM seats are the ones who can demonstrate both halves of the hybrid credibly.

Negotiating Across Tracks

Whichever track you land on, the flattening changes the negotiation.

• Anchor to level-matched data, not blended medians. The gap between the all-levels EM median ($352K) and level-matched figures at top companies ($732K at Google L6) is enormous. Pull the specific company-and-level cut from Levels.fyi and pay-transparency postings -- our guide to reading salary bands covers the mechanics.
• Negotiate equity first. The manager premium -- and most of the spread within staff bands -- lives in the stock grant, not base. Companies hold base salaries in tight bands but have far more flexibility on equity. An extra $90-100K per year of stock is a normal level-matched delta, which makes it a defensible ask.
• Cross-track offers are still competing offers. A Staff offer is real leverage in an EM negotiation and vice versa -- it proves your market value and your alternatives. The competing-offers playbook applies across tracks, not just within one.
• Price the risk, not just the number. An EM offer carries category risk a Staff offer does not: if the next reorg raises the IC-to-manager ratio, the manager layer is where the cut lands, and 32% of managers who fall back to IC roles take a pay cut. It is reasonable to weigh a slightly smaller Staff offer above a larger EM offer on risk-adjusted terms.