Compensation Philosophy & Pay Bands

⬅️ Back to Day 4: Convert

If you're hiring past your first 5-10 employees, you need a compensation philosophy — not "we pay market" but a documented framework that says how you pay (cash + equity mix), where you anchor (50th vs 75th percentile), how leveling works (IC1 → IC6), and what you'll do when someone gets a competing offer. The naive approach: negotiate every offer separately and hope it's fair. The structured approach: define philosophy + bands + leveling on paper before hire #10, then update annually. This is one of those things that's easy to defer and painful to retrofit (revealed pay disparities cause attrition + lawsuits).

What Done Looks Like

A compensation system in place:

1. Decide compensation philosophy first

Before pay bands, decide how you pay. The four-axis framework:

Define compensation philosophy.

Axis 1: Cash vs equity mix
- High cash / low equity (mid-market SaaS, agencies)
- Balanced (most B2B SaaS, $50-70k salary cushion + meaningful equity)
- Low cash / high equity (early-stage; below-market salary, big equity)

Pick one and apply consistently. Mixing creates resentment.

Axis 2: Where you anchor (percentile)
- Below market (50th percentile) — hard to recruit; cash-conservative
- At market (50-65th percentile) — standard for early-stage
- Top of market (75th-90th percentile) — premium players (FAANG-style)
- Combined: 50th cash + 90th equity (early-stage common)

Axis 3: Geographic strategy
- Single rate (everyone paid same regardless of location)
  - Pro: simple, fair-feeling
  - Con: SF/NYC under-paid; rural over-paid
- Tiered (zones — SF/NYC > LA/Seattle > Other US > Outside US)
  - Pro: market-aligned
  - Con: feels unfair when teammate makes more for same work
- Hybrid (single rate within tier; tier varies)

Axis 4: Transparency
- Open (all salaries public — Buffer, Whole Foods historically)
- Bands public, individuals private
- Closed (not discussed)

Pick philosophy and document. Each axis is a values statement.

For [COMPANY], output:
1. Cash vs equity recommendation
2. Anchor percentile rationale
3. Geographic strategy
4. Transparency level
5. 1-page philosophy doc

The default for B2B SaaS at $1M-30M ARR in 2026: 50-65th percentile cash, top-of-market equity, US-tiered geo, bands public.

2. Build a leveling system

Before pay bands, define what each level means.

Build a leveling system.

Engineering ladder example (most B2B SaaS):
- IC1 / Junior — entry-level, learning, 0-2 yrs experience
- IC2 / Engineer — solid contributor, owns features, 2-4 yrs
- IC3 / Senior — independent owner, mentors juniors, 4-7 yrs
- IC4 / Staff — drives cross-team work, technical leader, 7-10+ yrs
- IC5 / Principal — senior staff equivalent, sets technical direction, 10+ yrs
- IC6 / Distinguished — rare; sets company-wide direction

Manager ladder:
- M1 / EM — manages 4-8 ICs
- M2 / Senior EM — manages 8-15 or multiple teams
- M3 / Director — manages managers
- M4 / Senior Director / VP

Each level has:
- Scope (project / feature / system / org-wide)
- Autonomy (close supervision → independent → setting direction)
- Mentorship (mentee → peer → mentor → multiplier)
- Communication (within team → cross-team → cross-org)
- Influence (small → medium → large → company)

Other functions adapt similar ladders:
- Design: D1 → D5 + Design Manager → Director
- Product: PM → Senior PM → Group PM → Director / VP
- Sales: AE1 → AE5 + Sales Manager → Director
- Customer Success: CSM1 → CSM5 + CS Manager → Director
- Marketing, Finance, Ops: similar

Anti-patterns:
- Different ladders per role (engineering 8 levels, design 4) — feels unfair
- Levels with no clear scope ("Senior" means nothing without definition)
- Levels too narrow (every 2 years promotion expected) — burns out promotion review

For [COMPANY], output:
1. Recommended level count (5 IC + 4 manager is typical)
2. Scope definition per level
3. Promotion criteria (what does going from IC2 to IC3 require?)
4. Adjacent ladders (sales / CS / product)
5. Calibration process (multiple managers review levels together)

The reason leveling matters first: pay bands without leveling = chaos. "What's a Senior worth?" depends on what Senior means.

3. Build pay bands per level

Now anchor numbers to levels.

Build pay bands.

Sources for market data (in order of reliability):
- Levels.fyi (B2B SaaS / tech; great for engineering)
- Pave (subscription; deepest mid-market data)
- Radford (compensation surveys; enterprise-quality)
- Carta Compensation (subscription; equity-focused)
- Glassdoor / Comparably (free but noisy)
- Founders Network (peer comparisons)

Band structure:
- Each level has a min/mid/max
- 50% spread typical (e.g., $120k-$180k for IC3)
- Mid is target for "performing at level"
- Min is for "new to level / not yet up to speed"
- Max is for "ready for promotion but not yet there"

Example bands (B2B SaaS Series A in SF, 2026):
- IC1 / Junior Engineer: $110k-$140k base, 0.05-0.1% equity
- IC2 / Engineer: $130k-$170k base, 0.1-0.2%
- IC3 / Senior: $170k-$220k base, 0.15-0.3%
- IC4 / Staff: $220k-$300k base, 0.3-0.6%
- IC5 / Principal: $290k-$400k base, 0.5-1.0%

Note: numbers shift dramatically by stage, geo, and time. Pull current data.

Build for each role family:
- Engineering
- Product
- Design
- Sales (different — heavy variable comp; see sales-compensation-plans)
- Customer Success
- Marketing
- Finance / Ops / People

Anti-patterns:
- Single band across functions (engineering ≠ ops at same level)
- Bands too narrow (no negotiation room)
- Bands too wide (creates confusion about what a level is)
- Bands stale (don't update annually)

Output:
1. Band per level per function
2. Mid-point as target
3. Geographic adjustments
4. Equity grant ranges
5. Annual update cadence

The market-data trap: don't use Glassdoor as primary source. Levels.fyi for engineering; Pave / Radford for cross-function. Numbers vary 30%+ between sources.

4. Equity grant guidelines

Equity is harder to standardize but more important than cash for early-stage retention.

Build equity grant guidelines.

Stage-appropriate ranges:
- Pre-seed: founder-only typically; first hires get 1-5%
- Seed: senior hires 0.5-2%; engineers 0.1-0.5%
- Series A: senior hires 0.3-1%; engineers 0.1-0.4%
- Series B: senior hires 0.1-0.3%; engineers 0.05-0.2%
- Series C+: most hires <0.1%

By level (Series A as anchor):
- IC1: 0.05-0.10%
- IC2: 0.10-0.20%
- IC3: 0.15-0.30%
- IC4: 0.30-0.60%
- IC5: 0.50-1.00%
- VP: 0.5-2.0%
- C-level: 1-3%

Vesting:
- 4-year vest (1-year cliff) is standard
- Some companies: 1-year cliff + monthly after (most common)
- Others: quarterly vesting after cliff

Refresh grants:
- After year 2-3, give "refresh grants" to retain
- Typical: 25-50% of original grant, vesting 4 years from refresh date
- Triggers: promotion, exceptional performance, retention need

Equity types:
- ISOs (US tax-advantaged for employees) — preferred
- NSOs (non-qualified) — for non-US or post-cap
- RSUs (later-stage; pre-IPO common) — different tax

Repricing (rare; controversial):
- If valuation drops, options under-water
- Repricing options to current FMV is shareholder-controversial but sometimes done
- Discuss with board / lawyer

Output:
1. Grant ranges per level for current stage
2. Vesting schedule
3. Refresh grant policy
4. Communication template (explaining equity to candidates)
5. 409A valuation cadence (annual minimum)

The 2026 reality: equity grants compressed since 2022. Series A engineers commonly get 0.05-0.15% now (vs 0.2-0.5% in 2020). Adjust expectations.

5. Geographic compensation strategy

Remote-first companies face this hardest.

Pick geographic compensation strategy.

Strategy 1: Single global rate
- Companies: Buffer, GitLab (historically; revised), some indie startups
- Pro: simple, feels fair, attracts talent from non-tech-hub cities
- Con: pays SF/NYC below-market; pays rural above-market
- Outcome: hires from low-cost areas; loses to tech-hub competitors

Strategy 2: Geographic tiers (most common in 2026)
- Companies: most YC / VC-backed B2B SaaS
- Tiers (US example):
  - Tier 1: SF, NYC (100% benchmark)
  - Tier 2: LA, Seattle, Boston, DC (90-95%)
  - Tier 3: Other US metros (80-85%)
  - Tier 4: Outside US (varies wildly)
- Pro: market-aligned; cost-aware
- Con: same-role-same-pay-different-pay feels unfair; hard for moves

Strategy 3: Hybrid
- Single rate within US (or within country)
- Different rates outside
- Common at mid-market

Strategy 4: Cost-of-labor (not cost-of-living)
- Levels.fyi-style market data per metro
- Adjust based on what local market pays for the role
- Pro: market-rational
- Con: complex to communicate

Decision tree:
- Single rate works only if you commit to recruiting from anywhere + accept losing top-tier-hub talent
- Tiered works for most VC-backed; needs published tier policy
- Cost-of-labor needs sophisticated comp team

For [COMPANY], output:
1. Recommended strategy
2. Tier definitions (if tiered)
3. Adjustment math
4. Move policy (employee changes location)
5. Communication template

The move-policy rule: if employee moves from Tier 1 → Tier 3, do their pay drop? Most companies say yes (with notice + grace period). Some say no for existing employees. Decide before someone moves.

6. Transparency: open / partial / closed

Pay transparency is increasingly required by law (CA, NY, CO, WA all require salary ranges in job postings as of 2026).

Decide transparency level.

Level 1: Closed (legacy default)
- Salaries not discussed
- Bands not published
- Companies: most pre-2020 startups
- Increasingly illegal (CA / NY / CO / WA require ranges in postings)
- Verdict: don't choose this in 2026

Level 2: Bands public, individuals private
- Compensation policy + band ranges published internally
- Individual salaries not shared
- Most B2B SaaS in 2026 default
- Pro: meets legal requirements; gives clarity
- Con: requires trust that bands are followed

Level 3: Full transparency
- Every salary public internally (Buffer model)
- Sometimes externally too
- Pro: maximally fair-feeling; recruiting differentiator
- Con: high coordination cost; benchmark drift; competitive intel leakage

Legal requirements (2026):
- US: CA, CO, NY, WA, IL, HI, NV require salary ranges in job postings
- EU: Pay Transparency Directive (effective 2026) requires similar
- Best practice: include ranges in all postings regardless

For [COMPANY], output:
1. Recommended transparency level
2. What's published (bands / ranges / individual / equity)
3. Communication channel (internal docs / Notion / All-hands)
4. Legal compliance checklist by jurisdiction
5. Audit cadence (verify actual salaries match bands)

The "bands public, individuals private" pattern is the 2026 default. Meets legal requirements, provides clarity, manageable coordination cost.

7. Annual review cycle

Run an annual compensation review.

Cadence:
- Once per year (not multiple)
- Tied to fiscal year or calendar year
- Bigger companies do twice (Q2 + Q4) — overkill for small

Process:
- Month -3: Manager submits proposed changes (raises, promotions, refresh)
- Month -2: HR/Comp team calibrates across teams (avoid bias)
- Month -1: CEO/exec review + budget approval
- Month 0: Decisions communicated; raises effective

Components:
- COL (cost-of-living) raises (typically 2-5%)
- Merit raises (performance-based; 5-15% for high performers)
- Promotion raises (move to new band; 10-20%)
- Equity refresh (every 2-3 years for non-promoting)

Calibration:
- Compare similar levels across teams
- Identify outliers
- Address pay equity gaps
- Discuss promotion candidates with multiple managers

Communication:
- Manager 1:1 to deliver news
- Written letter (raise amount, equity refresh, etc.)
- New offer letter if level changing

Anti-patterns:
- Skipping years (compounds dissatisfaction)
- No COL raise even when inflation 5%+ (effectively pay cut)
- Promotion-only raise philosophy (creates "must promote to get raise" bias)

Output:
1. Annual review timeline
2. Calibration process (cross-team)
3. Budget allocation (% of payroll for raises)
4. Equity refresh policy
5. Communication template

The "no COL raise" trap: founders sometimes skip annual raises during cash-tight periods. If inflation is 5%+, that's a real pay cut. Plan for at least COL even when limiting merit.

8. Counter-offer policy

When an employee receives an outside offer, what do you do?

Define counter-offer policy.

Three approaches:

Approach 1: Never counter
- Companies: Netflix (formerly), Patagonia
- Logic: if they wanted to leave, can't keep them
- Pro: clean; no precedent set
- Con: lose great employees over fixable cash gaps

Approach 2: Match market always
- Logic: if outside offer reveals you're under-paying, fix it
- Pro: keeps top performers
- Con: incentivizes employees to "shop" for raises

Approach 3: Selective counter (most common)
- Counter if: top performer + below band + retention urgent
- Don't counter if: average performer / already at top of band / strategic concern
- Counter is a one-time event (don't expect to do it twice)

Counter-offer rules (if you counter):
- Match base, not necessarily everything
- Add equity refresh (matters more)
- Add commitment from manager (1:1 weekly, growth plan)
- Be honest about why this employee is worth it

Pitfalls:
- Counter creates ongoing resentment (employee feels "had to threaten")
- Other employees notice; expect similar treatment
- Most counter'd employees still leave within 12 months

Communication:
- Have the conversation in 24h, not 2 weeks
- Be direct: "We value you; here's what we can do"
- Don't compete on cash; compete on equity + growth + role

Output:
1. Recommended approach
2. Counter triggers (when to counter)
3. Counter mechanics (what to offer)
4. Manager + HR coordination
5. Track outcomes (counter retention rate after 12 months)

The data on counter-offers: ~50-80% of employees who accept counters leave within 12 months anyway. Counter is a buy-time-not-buy-loyalty move.

9. Pay equity audit

Required by law (US: state-level), required ethically (always).

Run a pay equity audit.

Goal: identify pay gaps by gender, race, age, tenure that aren't explained by performance/level.

Process:
- Pull all employee data: role, level, location, pay, demographics, hire date, performance rating
- Group by similar role + level + location
- Statistical analysis: median + mean + standard deviation per group
- Flag outliers: someone 20%+ below comparable peers
- Investigate: is there an explanation? (years at company, experience, performance)
- If unexplained gap: remediate

Tools:
- Pave (built-in equity audit)
- Carta Compensation
- Salesforce CRM data (if HR data is there)
- DIY in spreadsheet for <100 employees

Common findings:
- Gender pay gap (men paid more for same role)
- Internal-promote vs external-hire gap (external hires paid more)
- Tenure-based gaps (long-tenure under-paid relative to recent hires)
- Negotiation-aggressiveness gap (people who negotiated harder paid more)

Remediation:
- One-time raises for under-paid identified
- Communicate clearly without exposing individuals
- Track over time

Cadence:
- Annual minimum (US states require disclosure in some cases)
- Document methodology + findings
- Board / legal review

Output:
1. Audit methodology
2. Statistical analysis approach
3. Remediation budget
4. Reporting format
5. Continuous monitoring (don't let gaps re-form)

The negotiation-aggressiveness gap: women / minorities historically negotiate less aggressively. Result: same-role pay gap. Pay-equity audit catches this; offer-letter-band-anchoring prevents it (refuse to go above band regardless of negotiation skill).

10. Communicate the system

Transparency about the system matters more than transparency about individuals.

Communicate compensation philosophy.

Audience-specific:

Candidates (during recruiting):
- Salary range in job posting (legally required in many jurisdictions)
- Equity range disclosed at offer stage
- Philosophy explained: "We're 50th percentile cash, top-of-market equity"

New hires (during onboarding):
- Comp policy doc (1-pager)
- Band overview for their role family
- Annual review process timeline
- Refresh grant policy

Existing employees (annual cadence):
- Annual all-hands or town hall on comp philosophy
- Updated bands published
- Process changes explained

Managers (training):
- How to advocate for raises / promotions
- How to communicate decisions
- Calibration norms

External (recruiting marketing):
- Public job postings with ranges
- Glassdoor reviews show transparent comp
- Employer brand differentiator

Anti-patterns:
- "Don't discuss salary" policy (illegal in US under NLRA)
- Different stories to different employees (causes distrust)
- Bands published but actual practice different (worst-case)

Output:
1. Comp policy doc (1-page summary)
2. Detailed bands document (internal Notion / handbook)
3. Annual all-hands deck on comp
4. Manager training module
5. Job-posting template with range

The "don't discuss salary" gag: US National Labor Relations Act protects employees discussing pay. "Confidentiality clauses" prohibiting salary discussion are illegal. Don't write them.

What Done Looks Like

A working compensation system:

The mistakes to avoid:

Negotiating each offer separately. Creates pay gaps; legally risky; costs money long-term.
No equity refresh policy. Year-2-3 employees feel unappreciated; voluntary attrition spikes.
Stale bands. Update annually; market shifts 5-15% per year.
Geographic strategy that won't survive scaling. "Single rate" sounds great until you can't compete for SF talent.
Hidden salaries with verbal "we're competitive." Employees discover gaps; trust collapses.
No pay-equity audit. Discovered later as lawsuit; costs 10-100x preventive cost.
Counter-offer reflex without policy. Sets precedent that all retention requires resignation threat.

Compensation Philosophy & Pay Bands

What Done Looks Like

1. Decide compensation philosophy first

2. Build a leveling system

3. Build pay bands per level

4. Equity grant guidelines

5. Geographic compensation strategy

6. Transparency: open / partial / closed

7. Annual review cycle

8. Counter-offer policy

9. Pay equity audit

10. Communicate the system

What Done Looks Like

See Also