Salary Benchmarking: Market‑rate compensation

Audience: founders and early HR/ops leaders who need a practical, defensible way to set cash and equity pay. Goal: design a simple, auditable system you can maintain in hours per quarter—not weeks.

What “market rate” really means​

“Market rate” isn’t a single number. It’s a policy choice: a percentile of a reference market (e.g., 50th or 75th) for a clearly defined job at a clearly defined level and location, plus your equity and benefits philosophy. Public datasets (BLS), private surveys (Radford/Aon), startup panels (Carta, Pave), and crowdsourced sources (Levels.fyi) each see a different slice of that market—and each has noise and lag. Use more than one lens. (Bureau of Labor Statistics, Aon, Carta, pave.com)

What you’ll produce​

• A one‑page Compensation Philosophy (how you pay, where you pay, your target percentiles, and equity stance).
• A lean job architecture (titles/levels) and pay bands (min–mid–max) per level.
• A location policy (location‑agnostic vs. location‑based factors).
• An equity rubric (new‑hire grants, refresh cadence, promotion top‑ups).
• A lightweight governance model (approval matrix + exception log).
• A review cadence (at least annually) tied to cash runway and market checks.

Step 1 — Write your compensation philosophy​

Keep it short and explicit:

• Markets & percentiles: e.g., “We pay P50 cash / P75 total comp in our reference markets.”
• Cash vs equity: early stage = lower cash, higher equity; later stage = closer to cash market. YC’s rule of thumb: the first ten employees typically share ~10% of the company—useful as a sanity check, not a mandate. (Y Combinator)
• Geography: location‑agnostic (one band) or location‑based (factors by market). GitLab’s public handbook is a good model for transparent, market‑based location factors. (The GitLab Handbook)
• Transparency: commit to showing ranges internally; if you hire in states with pay‑range laws, you’ll be posting them anyway (see Compliance below). (Washington State Legislative Information, Findlaw)

Step 2 — Choose your data sources (triangulate)​

Use at least one authoritative, one startup‑specific, and one crowdsourced/real‑time source:

• Authoritative (government): BLS OEWS for wages by occupation, state, and metro; great for anchors and audits. Example: Software Developer median = $133,080 (May 2024). (Bureau of Labor Statistics)
• Enterprise surveys: Radford (Aon) and similar subscription surveys offer deep, audited tech job coverage across 1,400+ roles; gold standard if you can afford it. (Aon)
• Startup‑specific panels: Carta Total Compensation (large real‑time startup dataset integrated with cap tables) and Pave (includes the Option Impact/Advanced‑HR survey) are widely used for venture‑backed companies. (Carta, pave.com, Built In San Francisco)
• Crowdsourced/real‑time: Levels.fyi provides fresh compensation snapshots and role‑level comparables—most useful for tech roles and negotiations; treat as directional. (Levels.fyi)
• Open company policies: Buffer, PostHog, and (historically) GitLab publish compensation formulas and/or calculators—useful exemplars for location factors and transparency. (Buffer, PostHog, The GitLab Handbook)

Tip: build a simple “triangulation sheet”—one row per role/level, columns for each source, plus notes on sample size, geography, and refresh date.

Step 3 — Define roles and levels (job architecture)​

• Keep levels coarse (e.g., IC1–IC6, M1–M3) with crisp scope/impact definitions.
• Map titles to survey job codes carefully; the biggest benchmarking error is title inflation (e.g., calling an IC3 a “Senior” when the market doesn’t).
• If you’re small, prioritize bands for your top 8–12 roles you’ll hire this year.

Step 4 — Pick percentiles and your cash/equity mix​

• Choose one reference market per role (e.g., “SF Bay Area P50 base” or “US remote national P50”); for must‑win roles, consider P75.
• Convert that benchmark into your midpoint.
• Decide on equity philosophy by stage; for earlier hires, target ownership ranges instead of dollars (see Equity below).

Step 5 — Build ranges (bands) the simple way​

• Set min–mid–max around the midpoint to allow growth without immediate promotion.
• A practical default width: ±15% around mid (range spread ~30%).
• Track compa‑ratio (employee pay ÷ midpoint) and range penetration to manage fairness over time.

Example: If P50 base for Senior SWE in your reference market is $180k, set the band at $153k–$180k–$207k. (Numbers shown are illustrative; use your triangulation sheet for the true midpoints.)

Step 6 — Decide your location policy​

Common models:

• Location‑agnostic: one national band per level; simple and transparent; costs more if you hire outside top markets but avoids internal friction.
• Location‑based factors: define a market factor (e.g., SF = 1.00, Austin = 0.90, Warsaw = 0.75) applied to midpoints; base factors on labor markets, not cost of living. GitLab’s approach is a good reference. (The GitLab Handbook)
• Anchored multi‑market: keep a “HQ anchor” (e.g., SF midpoints) and publish a short list of supported markets with fixed factors. Buffer’s historical formula (e.g., 85% and 75% of SF for certain locations) illustrates the idea. (Buffer)

Whichever you choose, publish it, apply it consistently, and set transfer rules for employees who move.

Step 7 — Benchmark equity as rigorously as cash​

What to standardize:

• Grant sizes by level and stage: tie to ownership ranges (e.g., Staff Engineer 0.25–0.5% at pre‑Series A; adjust down as you scale). YC’s “~10% across first ten hires” helps calibrate your early bands. (Y Combinator)
• Cap table math: use fully diluted shares when translating percent to option count (outstanding + all convertibles + option pool). (Carta, Cooley GO)
• 409A valuation: sets the FMV for common stock and thus the option strike price; refresh at least annually or after material events. (Carta, JPMorgan Chase)
• Accounting: equity is expensed under ASC 718; finance will care about P&L impact of new grants and refreshes. (EY, Grant Thornton)

Quick formula Option count ≈ percent × fully‑diluted shares Always show candidates both ownership % and resulting options, plus the latest 409A and last preferred price for context. (Carta)

Step 8 — Plan for total cost, not just salary​

Benefits, payroll taxes, equipment, and allowances add up. As a national average, benefits run ~31% of total compensation cost (wages $32.92 + benefits $15.00 on total $47.92 per hour, March 2025). Budget offers using total employer cost, not just base pay. (Bureau of Labor Statistics)

Step 9 — Compliance & pay transparency (US, UK/EU)​

If you hire where salary range postings are required, you must disclose min–max and often a general description of benefits/other comp:

• Washington: job postings must include salary range and a general description of benefits and other compensation. (Washington State Legislative Information, Washington Labor & Industries, Washington Labor & Industries)
• California: covered employers must provide pay scales and include them in job postings (Labor Code §432.3); the state publishes complaint procedures. (Findlaw, CalDIR)
• New York State/City and others have similar requirements; monitor local rules. (Rippling)
• EU (for teams in Europe): the EU Pay Transparency Directive (EU) 2023/970 requires member states to implement strong transparency rules by June 2026; prepare now. (EUR-Lex, Ogletree)
• UK: employers with 250+ employees must report gender pay gaps annually. (GOV.UK, gender-pay-gap.service.gov.uk)

This page is not legal advice. Work with counsel on where you’re hiring.

Step 10 — Operationalize (so it actually works)​

• Offer rubric: map candidates to level → market → percentile, then apply any location factor and your equity rubric.
• Negotiation guardrails: allow limited movement (e.g., ±5% within band + equity within published range).
• Approvals: define who can approve out‑of‑band offers; keep an exceptions log (with expiration dates) to avoid silent drift.
• Review cadence: re‑bench annually (or twice a year in fast markets) and adjust bands; Carta’s and other startup reports help you spot movement. (Carta)
• Refreshes & promotions: schedule equity refreshes (e.g., annual 10–25% of initial grant for strong performers) and define promotion increments in both cash and equity.

A minimal template you can adapt​

Comp Philosophy (one page):

• We target P50 cash, P75 total comp for priority roles.
• We use location‑based market factors (published by city clusters). (The GitLab Handbook)
• Equity is benchmarked to ownership ranges by level and stage. (Y Combinator)
• We publish pay bands internally and post ranges externally where required. (Washington State Legislative Information, Findlaw)
• We review bands annually and adjust for material market changes. (Carta)

Band build (per role & level):

• Pick reference market and percentile → midpoint.
• Set min = mid × 0.85, max = mid × 1.15 (adjust spread to your needs).
• Add equity range (ownership %), translate to options using fully diluted shares. (Carta)

Equity notes:

• Grants are ISOs for employees where eligible; NSOs otherwise.
• 409A updated at least annually and after material events. (Carta)
• Refreshes annually; promotion grants top‑up to new range.

Common pitfalls (and how to avoid them)​

• Using posting ranges as data: ranges are often wide and compliance‑driven—use them as signals, not anchors. (Rely on OEWS/paid surveys/startup panels for anchors.) (Bureau of Labor Statistics, Aon, Carta)
• Title mismatch: “Senior” in one company = “Mid” elsewhere. Align to survey job codes before benchmarking.
• Ignoring equity math: quoting options without ownership % (and fully diluted denominator) confuses candidates; always show both. (Carta)
• Stale 409A / equity accounting surprises: refresh valuations and forecast ASC 718 expense before making big waves of grants. (Carta, EY)
• Budgeting off base pay only: plan the total employer cost; benefits alone average ~31% of comp. (Bureau of Labor Statistics)

Starter toolset​

• Data: BLS OEWS (free), Radford (Aon) (paid), Carta Total Comp (SaaS + reports), Pave with Option Impact (SaaS). (Bureau of Labor Statistics, Aon, Carta, pave.com)
• Transparency exemplars: Buffer salaries, PostHog compensation, GitLab’s handbook on remote compensation. (Buffer, PostHog, The GitLab Handbook)
• Sanity‑check comps: Levels.fyi real‑time charts for tech roles. (Levels.fyi)

Quick checklist (print this)​

• Define markets, percentiles, geography model.
• Build bands for this year’s critical roles.
• Publish your equity rubric and translate to % + options. (Carta)
• Align recruiters/hiring managers on the offer rubric and approval flow.
• Post ranges to comply where required (WA, CA, NY, etc.). (Washington State Legislative Information, Findlaw)
• Forecast total employer cost (wages + benefits + equity expense). (Bureau of Labor Statistics, EY)
• Re‑benchmark annually; update philosophy and bands as needed. (Carta)

Sources worth bookmarking​

• BLS: OEWS by occupation/state/metro; ECEC for employer cost structure. (Bureau of Labor Statistics)
• Radford (Aon) enterprise survey. (Aon)
• Carta compensation reports & platform. (Carta)
• Pave benchmarking data (inc. Option Impact). (pave.com, Built In San Francisco)
• Levels.fyi comp charts and role‑level data. (Levels.fyi)
• Pay‑transparency rules: WA RCW 49.58.110; CA Labor Code §432.3; EU Directive 2023/970; UK gender pay gap reporting guidance. (Washington State Legislative Information, Findlaw, EUR-Lex, GOV.UK)

This guide is educational, not legal or tax advice. Work with counsel and your accountant to tailor it to your company’s stage, locations, and financing.