The Offer Letter Has Six Numbers. You're Evaluating One.
- Base salary has the narrowest negotiation band (5-10%); equity grants and signing bonuses are where real movement happens.
- Annualize all equity and compare cash flow by year: Amazon's 5/15/40/40 vesting schedule makes a $200K grant worth $10K in year one, not $50K.
- Startup options are a different instrument: a 1x liquidation preference means investors recover their capital before common shareholders see a dollar, as in the Trados acquisition where common received zero.
- Benefits add $15K-$25K/year in real value: 401K match, health insurance premium share, and ESPP lookback provisions are routinely under-weighted in offer comparisons.
- Your manager's track record is the most important variable and appears nowhere on the offer letter; ask where former reports went.
- Career trajectory compounds faster than year-one comp: a role with visibility, honest feedback, and strong peers will outperform a higher-salary role on legacy systems within two to three years.
You pull up Levels.fyi, find your target role, and sort by base salary. Company A is $185K, Company B is $170K. You text a friend: "I think I'm going with A."
That decision just cost you $40,000 a year.
Not because Company B was necessarily better. Maybe it wasn't. But salary is the one number in an offer that's hardest to negotiate, loudest in your head, and least connected to what you'll actually take home in year three. Engineers who know how to evaluate a software engineer job offer routinely find that the "lower" offer is worth more. And they find the landmines before signing.
Build the Actual Job Offer Spreadsheet
Total compensation has five or six components depending on the company, and most of them have more flex than base salary.
Base salary gets all the attention but has the narrowest negotiation band, typically 5-10% above the initial offer. Equity grants, signing bonuses, and equity refresh schedules are where real movement happens. Recruiters usually have more budget authority over those items because they're one-time decisions or off-cycle budget lines, not permanent payroll costs. Which is why, when you push on base and get stonewalled, you should pivot to the signing bonus.
The components worth measuring:
| Component | Notes |
|---|---|
| Base salary | Baseline, hard to move much |
| Annual bonus | Formula-driven; ask for historical payout % |
| Equity (initial grant) | Annualize over vesting period |
| Signing bonus | Negotiable; watch for clawback clauses |
| Benefits (health, 401K match, ESPP) | Often $15K-$25K/year in real value |
| Equity refreshes | Critical for year 3+ comp; ask about criteria |
When you receive an offer, ask for each component broken out. Annualize the equity: a $200K RSU grant over four years is $50K per year in vested shares assuming a flat stock price. Add the bonus target and multiply by the historical payout rate, which the recruiter will share if you ask directly. Then add the benefits value.
One important trap: vesting schedules differ dramatically between companies. Amazon's standard 5/15/40/40 schedule means an initial $200K grant vests $10K in year one and $30K in year two. The same grant at a company with a standard 25/25/25/25 quarterly vest is $50K each year. Compare the actual cash flow across four years, not the headline number. Amazon's schedule is not a secret, but a surprising number of people sign and then do the math at the end of year one.
Public RSUs and Startup Options Are Two Different Instruments
A restricted stock unit at a public company is a straightforward thing. On vest day, you receive shares worth whatever the stock is trading at. It's ordinary income, you can sell immediately, and the paper value has a direct relationship to the cash in your account.
At a startup, an option grant is a different animal. The kind that bites.
You receive options at a strike price set to the 409A valuation, the IRS-mandated fair market value of common stock. The 409A is always lower than what investors paid, because preferred shareholders have liquidation preferences, board control, and anti-dilution protections. Common stock gets paid last. A company announcing a "$500M valuation" might have a $4.00 common stock 409A, because investors bought preferred at a 3-4x premium to common.
Standard startup vesting has a one-year cliff: you own zero shares if you leave before month twelve. After leaving, you typically have 90 days to pay the strike price plus taxes on the spread out of pocket before the options expire. Most people don't write that check, walking away from years of vested equity for shares they can't sell.
Private company options should be valued at a significant discount to their paper value, weighted by your probability estimate of a liquid exit happening within your planning window. Model multiple scenarios. "The company is worth $2B on paper" is not a scenario. It's a feeling.
The Liquidation Preference Trap Nobody Explains in the Offer
Most engineers don't learn this part until it's too late. Then they learn it the hard way, at a company all-hands where the founder says "exciting news" and the subtlety arrives 48 hours later in your bank statement.
Common stockholders can receive nothing in an acquisition, even when the acquisition price sounds large.
The canonical example is Trados Inc., acquired by SDL plc in 2005 for $60 million. The investors held preferred stock with a $57.9 million liquidation preference (their original $28.2M investment plus a cumulative 8% dividend). A management incentive plan took another $7.8 million. Fees consumed the rest. Common shareholders received exactly zero dollars. Not a rounding error. An actual legal zero. The Delaware court ruled the common stock was worth $0 at the time of sale.
This wasn't illegal. It was the math working exactly as written.
Your startup valuation is $1B. The preferred liquidation stack is also $1B. Funny how that works.
A 1x non-participating liquidation preference means investors get their investment back before a single dollar reaches common shareholders. In a modest exit, say $60M for a company that raised $40M across multiple rounds, the preferred shareholders recover most of their capital while employees get a fraction of what the math suggested. Participating preferred structures are even more aggressive, letting investors take their liquidation preference and then participate pro-rata in remaining proceeds alongside common holders.
Before you value startup equity, ask what the total raised capital is and what the liquidation preference terms look like. A recruiter may not know this, but a quick AngelList or Crunchbase search will show funding rounds. Model a conservative exit at 1.5x to 2x total raised, apply the preference stack, and see what common holders actually receive. That's your realistic baseline, not the last-round valuation times your ownership percentage.
Benefits Worth $15,000 to $25,000 a Year
Benefits feel like afterthoughts on an offer sheet. In dollar terms, they're not. You just can't sort by them on Levels.fyi, so your brain discounts them.
Health insurance is the most quantifiable. According to the 2025 KFF Employer Health Benefits Survey, average annual premiums for employer-sponsored family coverage are $26,993. Employers cover roughly 75% of family premiums. A company that fully covers single coverage is providing roughly $9,000 in untaxed compensation that doesn't appear on the offer letter. Deductible, out-of-pocket maximum, and premium share all matter. A $3,000 annual difference in out-of-pocket maximums is real money, especially if your health situation ever gets interesting.
401K matching is often the most straightforward underpriced benefit. A 6% match on a $180K salary is $10,800 per year, pretax. A company offering 3% match is structurally paying you $5,400 less per year compared to a 6% match competitor. That difference compounds over a career.
Employee stock purchase plans with a lookback provision are better than they look. The best structures offer a 15% discount on the lower of the stock price at the start or end of the offering period. If the stock rises 20% during a 24-month period, the effective discount on purchase date market price is roughly 32%. Contributing 10-15% of salary to an ESPP at a company with this structure generates $5,000 to $15,000 in additional returns annually, with near-guaranteed floor returns from the discount alone. This number shows up nowhere in the offer headline.
Your Manager Is the Career Investment Nobody Quantifies
No line in an offer letter tells you what your manager is actually like. That's the most important variable in the entire decision, and it's the one people research least because it's awkward to find out.
A manager who advocates for your promotion, routes you toward high-visibility projects, delivers real feedback, and protects your time will do more for your trajectory than a $15K base salary difference. The inverse is also true. A manager who treats 1-on-1s as status updates, assigns you to unmaintained systems nobody cares about, and competes with you for credit will actively damage your career at any compensation level. You won't notice it immediately. You'll notice it two years later when you're trying to explain what you shipped.
The questions worth asking directly: where have members of your team gone after leaving? What does your 90-day onboarding plan look like? How are promotion decisions made, and what's the typical timeline?
A manager who can name three former reports who got promoted or moved into better opportunities externally is a fundamentally different situation than one who deflects the question. One says "Sarah is at Stripe now, Marcus is eng manager at Figma." The other says "we really try to promote from within." Those are not equivalent answers.
Ask to speak with current team engineers. Not the hiring manager. Engineers will tell you the truth about on-call rotation, technical debt load, and meeting overhead. Probe specifically: Is there a real runbook for incidents, or do pages always go to the same two people? What percentage of sprint work is unplanned? Has the manager pushed for headcount in the last year, and what happened? At SpaceComplexity, voice-based mock interviews surface the "why are you looking" question constantly, and engineers who regret past offer decisions almost always point to a misread of the team situation, not the comp.
See also the questions worth asking before you accept an offer in our guide on what to ask your interviewers at the end of the technical round.
Work-Life Balance Is a Metric, Not a Feeling
"We have a great work-life balance" is something every recruiter says. It is meaningless. Ignore the claim and ask for data.
The questions that return actual information: What does on-call rotation look like, and how are 2 AM incidents handled? What time do engineers typically start and finish their days? What happened the last time someone needed an extended leave? Is there an expectation to respond to messages outside business hours, and in what timeframe?
Glassdoor and Blind are useful when you read for patterns, not individual reviews. A complaint about one manager in 2023 is noise. "Chronic overwork," "no work-life balance," or "high turnover" appearing consistently across hundreds of reviews, different years, and different departments is a durable signal about organizational culture, not one bad team.
Remote and hybrid policies deserve a careful read before signing. "Hybrid" can mean two days in the office or five. A 90-minute daily commute adds roughly 40 hours of unpaid time per month relative to a fully remote role. That's a real hourly rate difference. Do that math once and you'll never skim the "work location" line again.
Career Trajectory Compounds Faster Than Salary
Once you've built the table, you have four dimensions to weigh:
- Annualized total comp (base + target bonus + annual equity + benefits)
- Equity quality (liquid RSUs versus illiquid options; liquidation preference stack for startups)
- Career trajectory (manager quality, team visibility, promotion history)
- Lifestyle cost (commute, work hours, on-call load, culture fit)
The factor that compounds most over five years is career trajectory, not year-one total comp. A role that pays $15K less but puts you on a team where you'll ship genuinely hard problems, receive honest feedback, and work beside people who will recruit or recommend you later will outperform the higher-salary role on a legacy system within two or three years.
The salary number on the offer letter is where negotiations start. It shouldn't be where decisions end. Build the table. Do the math. And for the love of all that is good, read the vesting schedule.