Validate ranges against live market benchmarks

Map your range min, mid, and max against live market percentiles in a single prompt - so you know exactly where every cut stands before the next offer exception forces the question.

"Our midpoint is at p75 - but is it, really?"

The case

Why this comes up

In comp, you're frequently asked how your ranges compare to the market. Validating a range used to mean pulling survey cuts, reconciling sources that don't agree, and building a piecemeal spreadsheet. With Analyst Agent you get market percentiles mapped against your range min/mid/max in a single prompt.

Business problem

Most teams find out their ranges are out of date when an offer exception appears or a candidate says no. Without a fast validation loop, range refreshes are slow, inconsistent, and often incomplete - resulting in avoidable offer exceptions and internal equity exposure.

What you get
  • Clear market position for each range point – dollar values and percentile positioning.
  • Implied positioning for every cut (e.g., “midpoint is at p63,” “max falls below p75”).
  • A full population view for multi-level or multi-geo refreshes.
  • Ranges with data to support them, so you're confident in leadership discussions.
Sample Prompt

Validate our Software Engineering IC4 base salary range in the San Francisco Bay Area against market. Show market p50 and p75, and where our min, mid, and max land relative to each

Prompt variations

For Software Engineering IC3-IC6 in the US, compare our range midpoints to market p50 and flag every level where we're more than 5% off.

What this replaces
  • Manual survey downloads and spreadsheet lookups.
  • Reconciling sources that pull from different cuts and don't quite agree.
  • Spot-checking a handful of high-visibility roles while the rest go unreviewed.
  • Repeated back-and-torth interpreting what a range means relative to market.
Comp workflow

Validate market position of ranges vs Compa market data (or survey!)

Use case

Stay Ahead

Time saved

Hours per role → seconds per population

Get a demo
Sign in