Compensation and leveling
SkillMediaBuilds and maintains the leveling framework and pay structure — level definitions, salary bands, benchmarking, pay equity, and how raises and promotions are decided. Use this to design or revise leveling, set or adjust salary bands, benchmark against market, handle a compensation request or counteroffer, run a review cycle, or diagnose pay compression and equity issues.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Compensation and leveling skill
What this skill tells your AI
The instructions your AI receives, as published by cbrock84/headcount in plugins/people/skills/compensation-and-leveling/SKILL.md and read by ahel’s review.
Compensation touches employment law, pay transparency requirements, and equal pay obligations that vary by jurisdiction. Structural work here is fine; specific decisions about individuals should be reviewed by qualified counsel or an HR professional.
Leveling first
Pay structure without a leveling framework produces negotiated salaries, and negotiated salaries produce inequity that correlates with who negotiates hardest.
Define each level by scope and impact, not tenure or task list:
- What ambiguity can they handle — a defined task, a defined problem, an undefined problem, a problem nobody has identified?
- What is the blast radius of their decisions — their work, their team, the function, the company?
- What do they do for others: execute, contribute, guide, or set direction?
Levels must be distinguishable in a sentence. If two adjacent levels cannot be told apart by someone who does not know the people in them, they are one level.
Bands
For each level, benchmark against a market defined by the roles you actually compete with for candidates — not the whole industry, and not aspirational peers.
- Set a target position (at market, above, or below) and state it as policy rather than deciding case by case.
- Bands wide enough to allow growth within a level, narrow enough to mean something.
- Re-benchmark on a schedule. Markets move, and bands that do not move create compression that eventually costs more to fix than to prevent.
Compression and equity
Compression — new hires paid near or above tenured staff — is the predictable result of moving markets and static internal pay. It is corrosive because it is discovered, and it is always discovered.
Run a pay equity analysis on a schedule: pay by level, controlling for level and location, disaggregated by demographic. Where a gap exists, fix it directly rather than waiting for the next cycle. Findings here need qualified review before action.
Decisions
- Raises for sustained performance at level, promotions for sustained performance at the next level. A promotion is recognition that someone is already operating there, not a bet that they will.
- Counteroffers rarely work and reset expectations for everyone who observes them. Where a counteroffer is right, it should reflect a correction you should have already made.
- Every exception is a precedent. Document the reasoning, because you will be asked to repeat it.
Tooling
Benchmark data is the purchase that matters, and it is where the scale tier is real: free and crowd-sourced ranges are directional at best; Pave, Carta Total Comp or Option Impact suit venture -backed companies; Radford, Mercer or Willis Towers Watson are the surveys large employers price against, and similar.
Administration: the compensation module of the HRIS, or CompTrack, Assemble, and similar.
Whatever the source, know its effective date, its scope, and how it defines the level you are matching to. Two surveys disagreeing usually means they are describing different jobs.
Never
- Set an individual's pay before their level is settled.
- Use a candidate's prior salary as an input. It is prohibited in a number of jurisdictions and it imports someone else's inequity.
- Discuss an individual's compensation in a forum where their level has not been calibrated.
Signals
- GitHub stars
- 1k
- Forks
- 209
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
compensation-and-leveling- Source
- github.com/cbrock84/headcount