For Compensation and Benefits Analysts ·
What you'll accomplish
This guide gives you a structured process for using Claude to research market compensation for job roles that don't appear in your standard survey databases: the AI/ML roles, hybrid titles, and new job families that arrive before Mercer and Radford catch up.
What you'll need
Start by clearly defining what the role does. Often the challenge with emerging roles is that the title means different things at different companies.
Write a 2–3 sentence role description:
What you should see: A clear role description that distinguishes it from adjacent established roles.
I'm a compensation analyst trying to price a job role that doesn't have a direct survey benchmark. I need help:
1. Identifying the closest established benchmark roles I could use
2. Understanding what premium (if any) these skills command over the benchmark
3. Finding public data sources that might have relevant data
4. Building a reasonable salary range recommendation
Company context: We're a [size]-person [industry] company. Our general pay philosophy is [e.g., 50th percentile / 65th percentile for tech roles]. Our location is primarily [city/state or remote].
Role to price: [paste your role description]
Claude will suggest comparable established roles. Follow up to validate:
For the benchmark roles you suggested, what survey families in Mercer, Radford, or WTW would these typically fall under? What are the common job titles used in surveys for this type of work?
Also: what skills or requirements in this role might command a premium above the standard benchmark? (For example: specific programming languages, security clearances, industry-specific domain knowledge)
What you should see: A list of 2–4 benchmark roles Claude recommends, with rationale for why each is a reasonable match or partial match.
Now validate Claude's analysis against live market data:
What are the current salary ranges being advertised for [role title] on job postings? What factors most influence where in the range a candidate would fall?
Also, what companies are the primary employers competing for this talent? Understanding who's competing for these candidates helps me set the right target percentile.
Claude will synthesize publicly available information. Note: this data has a knowledge cutoff, so cross-reference with LinkedIn Salary, Indeed Salary, and Levels.fyi (for tech roles) for real-time validation.
Ask Claude to help structure your recommendation:
Based on this research, help me build a compensation recommendation memo for this role. I need:
1. The benchmark role(s) we're using and why
2. The estimated market median for this role in [location] at [company size/type]
3. A recommended salary range (minimum, midpoint, maximum) at our [X]th percentile target
4. Rationale for any premium above standard benchmark (if applicable)
5. Confidence level in this estimate given data limitations
Format as a 1-page compensation recommendation memo that I can save in the role's file.
Emerging roles often need to be repriced once survey data catches up. Ask Claude to help you set a review flag:
What's the typical timeline for new job families like this to appear in major compensation surveys? What should I watch for to know when better benchmark data is available?
For a new tech role:
Help me price a [title] role at a [size] [industry] company targeting the [X]th percentile. The role does: [brief description]. Required skills: [list]. Closest established benchmark I could find: [title]. Is this the right benchmark, and is there a premium for the specific skills this role requires?
For a hybrid role:
This role combines [function 1] and [function 2] responsibilities. I can't find a survey benchmark for the combination. How do I price a blended role? Should I weight toward one function? What are the risks of pricing it too high vs. too low for retention?
For a role with evolving scope:
We have an existing [title] role (currently grade 5, midpoint $85,000). The scope is expanding to include [new responsibilities]. Does this warrant reclassification to a higher grade? What benchmarks should I use to evaluate whether the current grade is still appropriate?