Job Hopping in Tech: When It Helps, When It Hurts

The job-switching premium that defined tech careers from 2020 to 2023 has mostly evaporated. According to ADP Research, the pay gap between job changers and job stayers settled to roughly 2 percentage points by August 2025 and has held there since. In 2022, that gap was 8.2 points. That’s a significant shift, and it changes how engineers should think about when to leave.

I want to be careful here. The conventional wisdom in tech circles is still that you should switch every two years to maximize salary. That advice was mostly right from 2020 to 2023. Whether it holds now is less clear, and I think people are slow to update.

What the Tenure Data Actually Shows

The BLS Employee Tenure report from September 2024 puts the overall median tenure for U.S. workers at 3.9 years in January 2024, the lowest since 2002. For workers in professional and technical roles, it’s 4.8 years. Those numbers don’t break out software engineers specifically, so take the “2.3-year average tenure for SWEs” figure that circulates on LinkedIn with some skepticism. I haven’t found a clean primary source for it.

What we do know is that tenure is falling across the board and has been falling for years. Engineers in their 20s average closer to 2.7 years per role, per the same BLS cohort data. At growth-stage startups the churn is faster still, partly because the companies change shape so quickly that your original role often disappears before you decide to leave.

Recruiters I’ve heard from consistently say three or more roles lasting under 18 months raises questions. One short stint doesn’t register. Two is context-dependent. Three in sequence reads as a pattern, and patterns invite scrutiny because the hiring team is about to invest real onboarding time in someone who might leave in a year.

When Switching Still Makes Sense

Early in a career, switching is often the most efficient way to learn. Different companies teach different things. A seed-stage startup forces you to ship fast and own problems you wouldn’t touch at a large company for years. A mature engineering org teaches you code review culture, incident response, and what real systems debt looks like at scale. Neither environment fully substitutes for the other.

Beyond learning, the network effect is real. Most referrals come from former coworkers. Every company you work at expands that pool, and referrals still move faster through most hiring funnels than cold applications.

The salary case for switching is more complicated than it was two years ago. Job switchers still earn more on average when they move, but the gap is much thinner. SHRM reported in early 2025 that job changers were seeing 4.8% wage increases versus 4.6% for stayers. That’s barely above the margin of noise. If you’re leaving primarily for a pay bump, you’ll need a strong competing offer to make the math work. The expectation that switching inherently pays is no longer enough.

The Depth Problem Nobody Talks About Enough

The most underrated cost of frequent switching isn’t how your resume looks. It’s the work you never get to do.

A lot of the interesting engineering problems only appear after you’ve been somewhere 18 months or more. Designing systems for real scale, seeing the downstream consequences of an architecture call you made, mentoring someone you hired. If you leave every year, you’re mostly writing CRUD apps and doing onboarding, indefinitely. You get breadth without depth, which is fine early on but starts to calcify your growth at mid-career.

There’s also the interview tax. A serious job search costs 2 to 4 months of mental overhead: prep, take-homes, system design rounds, offer negotiations. Do that every 12 to 14 months and you’re spending a significant share of your career looking for the next thing instead of building anything.

A Rough Framework, Not a Formula

I’m wary of giving hard tenure targets because they depend heavily on the company, the role, and what you’re trying to get out of your career. That said, some loose patterns hold up:

• In your first three years, switching every 1.5 to 2 years is generally fine. You’re gathering signal about what environments suit you. Depth matters less than exposure at this stage.
• Between three and seven years, try to have at least one stint that runs 2.5 years or longer. This is where you demonstrate you can take real ownership and see projects through.
• Past seven years, the pattern matters more. A mix of 3- and 4-year stints with one shorter role looks deliberate. Continuous 1-year hops at this point will cost you senior and staff-level opportunities at companies that care about retention signals.

These aren’t rules. If the job is genuinely bad, leave regardless of where you are on this timeline. A toxic environment compounds in ways that damage you in ways a resume gap doesn’t. And if an offer comes along that’s legitimately life-changing, the “right” tenure number on your CV doesn’t matter much.

How to Tell the Story in Interviews

If you’ve already got a resume that looks choppy, the interview is where you recover it. Framing matters more than most people think, and most people practice it far too late.

Each move needs a clear “toward” story rather than an “away from” story. “I wanted to learn distributed systems and this was the right environment for it” lands differently than “the previous place was frustrating.” Recruiters hear the latter version constantly. The former suggests you have a plan.

When asked about short stints directly, being specific defuses most concern. “The startup ran out of funding at month 11” is not a pattern. “I took a contract while looking for the right full-time fit” makes sense. What doesn’t work is vague answers like “I was ready for a new challenge” for every single move.

Practice those answers out loud before interviews. Reading them on a page is different from saying them under mild pressure. Tools like LastRoundAI’s mock interview platform let you run through behavioral questions in a low-stakes setting so you’re not finding your words for the first time when it counts. The behavioral interview questions guide covers the specific framing techniques that hold up under follow-up questions.

One honest read on all of this: the optimal switching cadence is probably faster than most large companies would prefer and slower than most ambitious engineers actually practice. Where exactly it lands depends on your specific situation in ways that no blog post can fully resolve.