Comp Negotiation Tool — Eccles School of Business

Guide — What This Tool Does & Why

Six steps, in order: benchmark sources build a defensible target, you compare a real offer against your counter, you practice the conversation, then you score and track it. Everything below is reference material — skim it now, or come back to a specific term whenever a step doesn't make sense. Nothing here is required reading before you start.

Why 88% and 110%?

These bands aren't a market statistic — they're a negotiation heuristic.

The 88% floor approximates a walk-away sanity check: a number meaningfully below target that should prompt you to pause and reassess, not an offer to reject automatically.
The 110% stretch is a deliberately ambitious opening ask — high enough to leave room to move down to your actual target without looking unreasonable, a standard anchoring tactic in negotiation.

Treat both as reference points to reason from, not thresholds the tool is telling you to enforce.

No external citation backs the specific 88%/110% figures — they're this tool's own instructional design choice, not a number pulled from a study.

Why are sources weighted differently? (Step 1)

Each benchmark source you can pick from carries a preset weight from 1.00 (Eccles Employment Report) down to 0.50 ("Other," unverified). The ranking reflects a widely-accepted general principle in compensation research: government and institutional data (BLS, DOL, your program's own employment report) is harder to game and easier to verify than crowdsourced, self-reported platforms, so it counts for more in the weighted average.

The ranking order is a defensible general principle. The exact numbers (why Levels.fyi is 0.70 and not 0.65 or 0.75, for instance) are this tool's own editorial judgment, not values drawn from a specific cited study.

Experience premium & geographic differential

Experience (years): years of direct, differentiating experience above what this role already expects — not your total years worked.

Experience premium rate: how much each qualifying year adds to your baseline, before the cap is applied. Default 1.5%/year.

Experience premium cap: the maximum total premium your experience can add, no matter how many years you claim. Default 7.5%.

Geographic differential: adjusts your baseline for cost-of-labor differences in your target city. Default 5% — for your own city's actual figure, use BEA Regional Price Parities or BLS Occupational Employment & Wage Statistics — not a local CPI index.

The experience premium is a teaching assumption, not a labor-market rule. Apply it only to direct, differentiating experience above what the role already expects.

The 1.5%/year rate and 7.5% cap are this tool's own instructional defaults, not figures drawn from a cited study — BEA and BLS above are real external sources for the geographic differential itself, but the experience-premium numbers aren't sourced the same way.

Target, Floor & Stretch — how they're calculated

Three reference points derived from one calculation, not three independent numbers.

Formula: Target = Baseline × (1 + capped experience premium) × (1 + geographic differential); Floor = Target × 0.88; Stretch = Target × 1.10.

Why it matters: Target is your primary ask, grounded in the baseline and adjustments above. Floor is a walk-away sanity check, not a number to volunteer. Stretch is your opening ask, leaving room to negotiate down to Target without looking unreasonable.

How to use it: open near Stretch, defend Target with the reasoning above, and treat Floor as a private signal to pause and reassess — never state it out loud to a recruiter.

Risk weight — what it actually is

Risk weight is not calculated from anything you enter — it's a fixed assumption, the same for every student: base salary 1.00, signing bonus 0.95, bonus 0.70, equity 0.50. It represents how likely that dollar is to actually show up: base is nearly certain, a bonus assumes roughly 70% attainment, and equity is discounted hardest because of vesting risk and uncertain future value. The risk-adjusted figures in Step 3 apply these weights so an offer heavy on equity or bonus isn't mistaken for an offer of equal guaranteed cash.

These four numbers are this tool's own editorial judgment, not coefficients pulled from a named compensation study. Real research on bonus attainment exists (compensation-consulting surveys of large companies), but none of it validates 0.70 specifically over a nearby number like 0.65 or 0.75 — treat these as reasonable, not authoritative.

Why risk-adjusted value? (Step 3)

A dollar of unvested equity is not a dollar of base salary — it can be reduced, delayed, or lost if you leave early or the company underperforms. Risk weights (shown per row in the Offer Terms table) discount less-certain components so the comparison reflects that, rather than treating every dollar in an offer as equally guaranteed.

Why practice cold first? (Step 4)

Attempt your response before seeing any AI coaching. That's what makes the improvement after feedback real, not rehearsed — if you see the "right answer" first, you're just copying it, not building the skill.

Pacing and length — practice tips (not scored)

Pacing: research led by MIT Sloan's Jared Curhan, published in the Journal of Applied Psychology, found that pausing 3–9 seconds before responding tends to precede negotiation breakthroughs — rushing to fill silence, or answering instantly, is associated with less reflective, weaker outcomes for both sides. Practice sitting with a pause before you respond, especially after a lowball or a deadline threat, instead of rushing to fill the silence.

Length: there's no ideal word count, and this isn't graded — but over-explaining (justifying your ask at length, oversharing personal financial reasons) can reveal leverage you didn't mean to give away, while an answer that's too terse can read as cold or dismissive. Aim for a considered sentence or two, not a monologue or a one-line demand.

Neither of these is something this tool actually measures — there's no live conversation or timer here. Treat this as coaching to apply in real practice, not a standard the rubric checks.

The AI scorer — how it actually works

The heuristic scorer is a keyword scan, not an AI judgment. It reviews your logged response and pasted transcript for expected words and phrasing patterns — it does not evaluate your actual negotiation skill the way a coach or a live AI judge would. Treat every auto-filled score as a starting estimate, review and adjust it yourself.

Rubric criteria, level thresholds & the Fundamentals Check

Four criteria, 0–5 each, 20 points total: Evidence, Clear Ask, Professional Tone, Close.

Level thresholds: 18–20 = Strong · 15–17 = Proficient · 11–14 = Developing · below 11 = Needs Rework.

The Fundamentals Check: scoring 1 or below on Evidence or Clear Ask limits your level to "Developing," regardless of total — these two are treated as non-negotiable fundamentals, since a well-delivered ask with no evidence or no actual number behind it isn't yet a complete negotiation.

Step 1 — Benchmark Sources (optional)

You can add up to three benchmark sources to strengthen your target. Sources are optional — this tool works with zero sources entered — but you are responsible for verifying the accuracy of any figures you rely on in a real negotiation.

Comp Sources

Each source carries a preset weight based on how verifiable it is — pick a source and its weight shows automatically underneath. Institutional and government data count more toward your baseline than crowdsourced or self-reported figures, because they're harder to game and easier to check.

Baseline

Sources entered	0 of 3
Weighted average of entered sources	—

This is where the weights actually matter: a source weighted 1.00 pulls the average toward its number twice as hard as one weighted 0.50. Two sources of equal dollar value but different weights don't count equally.

Baseline market salary — used in Step 2 (editable)

Pre-fill this from your sources, or type a number directly if you have zero sources.

Before You Start

Why 88% / 110%? These are negotiation heuristics, not market statistics — full explanation in the Guide tab.

Accountability: This tool does not verify that a source, URL, or dollar figure you enter is accurate. That responsibility is yours. Double-check anything you plan to cite in an actual negotiation.

Step 2 — Base Salary Triangulation

Anchors your ask in verified data before you speak to a recruiter, rather than a number based on personal need.

Your Info

Relevant pre-MBA experience (years)

Geographic differential (%)

Advanced: experience premium assumptions

The default below is a reasonable teaching assumption — adjust only if you have a specific reason to.

Experience premium rate (% per year)

Experience premium cap (%)

What these inputs mean

Definitions for experience premium, its cap, and the geographic differential are in the Guide tab.

Your negotiation range

Reference floor (88%)

—

Triangulated target base

—

Stretch base (110%)

—

Open near Stretch, defend Target, treat Floor as private — never state it out loud. Full formula and reasoning in the Guide tab.

Show your work — intermediate calculation

Baseline market salary (Step 1)	—
Uncapped experience premium	—
Applied experience premium (capped)	—

Step 3 — Offer Comparison & Risk-Adjusted Value

A dollar of unvested equity is not a dollar of base salary. Risk weights make that visible before you decide what to trade for what.

Offer Terms

Fill in the initial offer column with what's on your actual offer letter. The counter column's base salary auto-fills from your Step 2 target; everything else you set yourself for what you plan to ask for.

Metric	Risk weight	Initial offer	Your Counter	Variance

Signing bonuses often carry clawback terms — model the actual dollar exposure in the Offer Compensation Planner.

Bring In an Offer from the Offer Compensation Planner

Already have a real offer modeled in the Offer Compensation Planner? Paste the transfer text copied from that tool's own export box (its Corporate mode) to fill the Offer Terms above instead of typing them. If you haven't used that tool yet, that's the normal case at this stage — just fill in Offer Terms manually above.

Send This Offer to the Offer Compensation Planner

Do this once your Offer Terms above reflect what you want to send — whether you filled them manually or imported them above. Copy this transfer text, then open the Offer Compensation Planner and paste it into its "Bring In an Offer from the Comp Negotiation Tool" box (visible in that tool's Corporate mode) — it auto-fills the matching fields there, initial and counter. Corporate mode only: the Offer Compensation Planner's Startup mode (SAR stake, phantom equity, KPI-weighted bonus) has no equivalent here and can't be exchanged either direction.

Calculated outputs — Year 1

Use this one: "Risk-adjusted economic value." It already accounts for how likely each part of the offer is to actually show up, so it's the number to lead with when deciding what to say. The other two rows are supporting context, not competing answers.

Output	Initial offer	Your Counter	Net gain

3-year projection (3% annual merit assumption)

Output	Initial offer	Your Counter	Net gain

Vesting is modeled straight-line from Year 1. If your equity has a cliff longer than one year, the Year 1–2 vested-equity figures above will overstate what actually vests — adjust manually or treat Year 1 equity value as $0 until your cliff passes.

Step 4 — Practice With AI

No API key is used by this tool. Four steps, in order: get the prompt, attempt your response before seeing any AI coaching, have the real AI conversation and paste it back, then score it. Attempting it yourself first is what makes the improvement after feedback real, not rehearsed.

A. Copy this prompt

Paste this into Claude, ChatGPT, or any AI assistant. Have the full conversation there — then come back for Step B.

B. Practice your response

Before you go have the AI conversation, write or say what you would actually respond in this scenario. Type it, or use the microphone (Chrome desktop works best). This is your unaided attempt — nothing to look up, nothing to get right, just what you'd really say.

C. Paste your AI conversation transcript

Now go have the conversation using the prompt from Step A, then paste the full transcript here. This is the main input Step D uses to score your practice.

D. Score this practice

Click below for a rough, keyword-based score. This is not a real AI judgment — review and adjust every result yourself. Your score lands in Step 5.

Roleplay Evaluation Rubric

Score your own transcript, or ask a friend, mentor, or coach to score it with you. Scored 0–5 per criterion.

How scores get here: click Score my response in Step 4 and Evidence, Clear Ask, and Close fill in automatically from a keyword scan (a rough estimate, not a judgment call). Professional Tone always stays blank — that one takes actual judgment, so you (or a peer/coach) read the response and score it yourself. Auto-filled scores are meant to be reviewed and changed too, not accepted as-is.

These scores are not saved automatically. Go to Step 6 (Tracker) and click "Save current session" before you leave this page, or your scores are lost.

Criterion	0 — Missing	2–3 — Developing	4–5 — Proficient	Score (0–5)

Total score

0 / 20

Performance level

—

Fundamentals Check

—

Why this exists: see the Guide tab.

Progress Tracker

Save a snapshot after each practice session to track improvement over time. Stored in this browser only — not shared or backed up elsewhere.

Saved	Evidence	Ask	Tone	Close	Total	Level

No sessions saved yet.