01 · Scope first
Verify that you have two live, written offers to compare.
This guide owns a narrow moment in the career journey: you have at least two current offers with written terms, and you have not accepted either one. It is not a general career-choice exercise, a salary benchmark, a negotiation script, or a stay-versus-leave framework.
- Record the response deadline exactly. Include the date, local time, time zone, source, and whether an extension was actually confirmed. Recording a deadline does not imply a right to more time.
- Confirm the offer stage. Distinguish a formal written offer from an interview discussion, recruiter estimate, contingent indication, or generic benefits page.
- Pause for authenticity concerns. The worksheet cannot prove an offer is legitimate. Independently contact the employer through a trusted channel before sharing identity, banking, tax, or immigration documents. The FTC’s current U.S. warning about fake job-offer texts describes common referral signals.
Do not paste an offer letter here. Enter short, pseudonymous claims only. Exclude employer and person names, home address, account information, identification numbers, medical details, immigration documents, passwords, and full emails or plan documents.
02 · Two independent axes
A claim can be known poorly, or unknown for a precise reason.
The Offer Evidence Record separates knowledge state from evidence state. “Recorded” answers whether a claim has been entered. “Written term,” “direct written confirmation,” “verbal statement,” “inferred,” or “no evidence” describes provenance—not truth, enforceability, or probability.
| Claim | Knowledge state | Evidence state | What the record does |
|---|---|---|---|
| Start date appears in the offer letter | Recorded | Current written term | Shows the date and source without calling it guaranteed. |
| Manager says remote work is “usually fine” | Recorded | Direct verbal statement | Keeps the words and date separate from a written commitment. |
| Health-plan document was requested | Awaiting response | No evidence yet | Surfaces the request instead of filling the gap with an estimate. |
| Two documents give different vesting dates | Conflicting | Retain both sources | Preserves both versions and prompts a clarifying request. |
| Relocation does not apply | Not applicable | No evidence needed | Keeps inapplicability distinct from missing information. |
Blank means unset. It never becomes zero, “no,” or not applicable. An explicit zero remains zero. This distinction is essential when one offer has a zero-dollar item and the other simply has no recorded answer.
03 · Original utility
Build a private, claim-centered Offer Evidence Record.
Use aliases such as Offer A and Offer B. For each lane, enter the shortest useful claim, choose its knowledge and evidence states, and add the source date, applicability, or conditions. The enhanced worksheet runs in your browser memory. It does not upload, save, email, or send field contents, and it has no score, weight, completion percentage, rank, or recommendation.
Offer Evidence Record
Private by default: use aliases and short notes. Nothing in this worksheet should require an employer name or sensitive document. Closing or refreshing the page clears the in-memory record unless you separately make your own notes.
Evidence summary
The enhanced worksheet will place like-for-like differences, preserved unknowns, evidence asymmetries, conflicts, and next questions here. It will not produce a winner.
04 · Do not flatten the terms
Ranges, conditions, benefits, and equity stay in separate lanes.
A range is not its midpoint. A target bonus is not base pay. An equity grant is not cash. A generic benefits summary is not necessarily the plan that applies to this offer. Keeping these claims separate prevents a tidy-looking total from hiding assumptions.
| Lane | Preserve | Do not infer |
|---|---|---|
| Base pay | Currency, unit, period, basis, exact value or range | Net pay, purchasing power, inflation, or currency conversion |
| Bonus or commission | Target or range, eligibility, performance basis, payment date, forfeiture | Probability, expected value, or guaranteed payout |
| Equity | Instrument, quantity if documented, vesting, exercise or forfeiture terms, source | Cash equivalent, future price, liquidity, tax result, or investment advice |
| Health and retirement | Current applicable documents, eligibility, waiting period, match and vesting terms | Personal suitability, future cost, adequacy, or plan selection |
| Work reality | Who said what, when, and whether it applies to this role | Company-wide fact, culture score, or outcome forecast |
For U.S.-specific document checks, the Centers for Medicare & Medicaid Services explains Summary of Benefits and Coverage materials, while the U.S. Department of Labor describes plan information and Summary Plan Descriptions. Those sources help identify documents; they do not decide which plan is right for you.
05 · Narrow arithmetic
Subtract only exact terms with the same comparison basis.
A literal difference is allowed only when both values are exact and share the same currency, unit, period, basis, and conditions. The formula must remain visible, and everything outside that narrow pair remains excluded.
Synthetic example—no real employers: Offer A records USD 112,000 per year in fixed base salary from a current written offer. Offer B records USD 105,000 per year on the same basis. The record may show 112,000 − 105,000 = 7,000 USD/year. That is one recorded difference. It excludes taxes, benefits, bonus, equity, commute, and every other claim; it is not total compensation and does not make Offer A the winner.
Now add a bonus range of 0–15% to Offer A and a conditional USD 10,000 bonus to Offer B. The record shows both terms and their conditions but performs no midpoint, probability, or expected-value calculation. If Offer A has only a recruiter statement about remote work while Offer B has a current written schedule, the record surfaces an evidence asymmetry rather than awarding points.
- Different currencies, units, periods, pay bases, or conditions stop arithmetic.
- Ranges remain ranges even when both ranges overlap.
- Qualitative statements remain attributed statements.
- Offer display order stays the order you entered; unrelated unknowns never reorder it.
06 · Questions are output
Make missing, conflicting, and time-sensitive evidence impossible to overlook.
A useful comparison does not hide unfinished work. It should surface what was not asked, what is awaiting a response, what was not provided, what is not understood, and which sources conflict. Every open item becomes a bounded next request rather than a guessed value.
Turn a gap into a specific request
- Plan applicability
- “Could you send the current plan document that would apply to this role and confirm the eligibility date?”
- Conditional pay
- “Could you confirm in writing the target, measurement basis, payment timing, and circumstances that would make this amount unavailable?”
- Work arrangement
- “Which location and schedule terms are written for this role, who may change them, and when are they reviewed?”
- Conflicting dates
- “These two documents show different dates. Which version currently applies to this offer?”
The U.S. Department of Labor’s retirement-plan guide illustrates why eligibility and vesting need document-bound dates. The IRS discussion of stock options in Publication 525 also shows why equity and tax treatment depend on facts and timing; this page therefore leaves equity unvalued and routes individual questions to qualified professionals.
07 · Keep judgment human
Record your constraints and what evidence could change your view.
Personal constraints belong in the record, but the page does not set priorities or weights. You decide whether a claim supports, conflicts with, or leaves a constraint unresolved. Examples might include a latest feasible start date, a schedule you can actually maintain, a travel limit, or a required accessibility conversation.
Then write one reversal prompt for each offer: “What credible new evidence would change my current view?” The answer might be an applicable plan document, written schedule terms, clarification of a repayment clause from a qualified professional, or a direct conversation with the person who would manage the role.
Decision research shows that framing, elicitation, and the act of listing reasons can influence expressed preferences; those findings do not validate this tool or determine an employment choice. They are why the record exposes sources and assumptions instead of turning them into hidden weights. See the experiments by Tversky and Kahneman (1981), Wilson and Schooler (1991), Payne, Bettman, and Johnson (1988), and Lichtenstein and Slovic (1971).
08 · Boundaries
Know when the record stops and qualified help begins.
The Offer Evidence Record is educational. Labels such as “written” and “confirmed” describe provenance; they do not establish accuracy, enforceability, entitlement, or a guarantee. The page cannot interpret a contract, certify an employer, value equity, choose insurance, estimate taxes or net pay, or predict an employment outcome.
- Legal rights, contract language, noncompetes, classification, notice, at-will terms, clawbacks, and enforceability
- Tax treatment, net pay, equity exercise, investment risk, liquidity, or present value
- Insurance and benefits selection, coverage suitability, medical claims, or individualized retirement adequacy
- Immigration status, sponsorship, work authorization, or jurisdiction-specific employment rules
- Fraud investigation or a declaration that an offer is legitimate
- Any decision to accept, decline, negotiate, or request an extension
The Bureau of Labor Statistics Employer Costs for Employee Compensation release is aggregate context showing that employer compensation has multiple components; it does not price an individual offer. For suspected fraud in the United States, verify independently and use the FTC’s ReportFraud portal. Official U.S. links on this page are jurisdiction-specific and do not establish rules elsewhere.
Sources and their roles
- Centers for Medicare & Medicaid Services — Summary of Benefits and Coverage (reviewed August 23, 2026). Supports asking for current applicable plan materials; not insurance interpretation.
- U.S. Department of Labor — Plan Information (reviewed August 26, 2026). Supports provenance fields for Summary Plan Descriptions; not an individualized benefits decision.
- U.S. Department of Labor — What You Should Know About Your Retirement Plan (reviewed August 23, 2026). Supports eligibility and vesting fields; not a retirement-value estimate.
- Internal Revenue Service — Publication 525 (reviewed August 23, 2026). Supports the separate, conditional equity lane; not tax or investment advice.
- Bureau of Labor Statistics — Employer Costs for Employee Compensation (reviewed August 23, 2026). Supports component separation only; aggregate data is not imputed to either offer.
- Federal Trade Commission — “That job offer text is probably a scam” (reviewed August 23, 2026). Supports independent verification and referral; the record does not classify legitimacy.
- Tversky & Kahneman (1981), “The Framing of Decisions and the Psychology of Choice”. Primary experimental background on framing; not validation of the Offer Evidence Record.
- Payne, Bettman & Johnson (1988), “Adaptive Strategy Selection in Decision Making”. Primary experimental background on strategy use; not a prescribed job-choice method.
- Wilson & Schooler (1991), “Thinking Too Much”. Primary experimental background on reasons and preference reports; not evidence that analysis is universally harmful.
- Lichtenstein & Slovic (1971), “Reversals of Preference Between Bids and Choices in Gambling Decisions”. Primary experimental background on elicitation effects; not an employment-outcome study.
Editorial boundary: the Offer Evidence Record is an owner-original, unvalidated framework. Its synthetic conformance cases test only whether programmed states remain distinct and whether arithmetic stays narrow; they do not establish reliability, validity, or better career outcomes.
Common questions
Questions people ask while comparing offers
Should I choose the offer with the highest salary?
No universal rule follows from one pay figure. First verify that the figures share a basis, then keep every other claim, condition, unknown, and personal constraint visible. This page does not make the choice.
Can the worksheet calculate total compensation?
No. It never adds base pay, conditional bonus, equity, benefits, time, commute, flexibility, or qualitative claims into one total. That would require assumptions the record is designed to expose.
How should I treat a verbal promise?
Record the speaker’s role, date, exact scope, and the fact that it was verbal. Ask whether the employer can confirm the term in writing. Do not silently upgrade it to a written commitment.
What if I think an offer may be fraudulent?
Stop the comparison and verify the employer independently. Do not send money or sensitive identity or banking information. The worksheet cannot certify legitimacy; use authoritative reporting and local support channels.
Can I add my current job as a third offer?
This record is intentionally limited to two or more live written offers. If your real question is whether to remain in your current role, use the stay-versus-leave decision framework so the status quo is examined on its own terms.