# The Two-Offer Decision, Scored to One Comparable Number Each

*You will turn two offers into one comparable after-tax number each, weight the non-money factors on a fixed scale, and apply a rule that resolves even the hard case.*

- Canonical URL: https://www.refolk.ai/candidates/guides/two-offer-decision-scored
- Pillar: Offers and negotiation
- Format: Framework
- Published: 2026-09-24
- Last reviewed: 2026-09-24
- Reading time: 17 min

You have two offers in hand and one question that will not resolve itself: which is actually worth more, and how do you take it without second-guessing for a year. This guide is for a candidate holding or expecting two competing offers who wants a repeatable way to turn each into a single comparable number, score the parts that are not money on a fixed scale, and apply a rule that decides even when the money and the gut disagree. It delivers one after-tax, location-adjusted, risk-discounted figure per offer, a weighted rubric two people would grade the same way, an explicit tie-break, and a negotiation step most comparison pages skip.

Most public offer-comparison pages stop at a total-comp table and a generic weighted worksheet. That leaves you with two piles of numbers and no rule for the hard case where the higher-paying offer is not the better job. This framework carries each offer all the way to one figure, then hands you a rule for the case where the numbers are close.

## Why the base salary comparison lies

Base salary is the worst single number to compare two offers on, because it is only a fraction of what the job is actually worth. In the BLS Employer Costs for Employee Compensation data for March 2026, wages and salaries accounted for 69.9% of private-industry employer compensation cost, with benefits making up the remaining 30.1%. A $500 base edge is noise once retirement match, bonus, and health coverage enter the math.

The practical consequence is that the gap between two offers usually shrinks once you total everything, and sometimes it flips entirely. Practitioner framing puts the off-base share even higher than the BLS figure, because it includes bonus and equity: in a typical professional role, non-salary components add 20 to 50% to the value of a package. That is enough to reverse a ranking. In one worked case, a $150,000 package beat a $165,000 one by $33,000 a year once benefits and equity were counted.

**30.1% - Share of private-industry employer compensation cost that is benefits, not wages**

BLS Employer Costs for Employee Compensation, March 2026. A base-only comparison ignores roughly a third of the package.

Benefits share also moves by sector, which matters if your two offers straddle private industry and government. The table below is worth holding open while you assign dollar values to each offer's benefits.

| Sector | Wages share | Benefits share |
|---|---|---|
| Civilian (all) | 61.5% | 38.5% |
| Private industry | 69.9% | 30.1% |
| State and local govt | 61.5% | 38.5% |

If one of your offers is public-sector, its lower base may carry a materially larger benefits load. Do not compare its base to a private offer's base and call it a day.

## The components you annualize, per offer

For each offer, total these into one annual number: base, bonus or commission, retirement match, the dollar value of the health plan and PTO, any signing bonus, and stipends. That single figure, not the base, is the object you compare.

The two moving parts that trip people up are the bonus and the signing bonus. Score the bonus at a realistic payout percentage, not the target. A 20% target that pays out at 60% of target is a very different number, so ask the recruiter for last year's actual payout before you plug in anything. Annualize the signing bonus by spreading it over two years, which is the standard calculator convention, and read the separate signing agreement for a clawback trigger before you count a dollar of it.

> **Rule:** Total both offers before you judge either
>
> No component gets to sit outside the annual number. Base, realistic bonus, match, benefit dollar values, signing spread over two years, stipends, minus commute. If it is not in the total, it is not in the decision.

Commute is the one deduction people forget. Subtract the annual dollar cost of the commute from each total before you compare, because a shorter commute is real money and real time that the offer letter never mentions.

## How to translate equity into today's dollars

Equity is where a comparison most often lies, because the headline number and the cash value are not the same thing. A recruiter may say "total comp is $300,000" when a third of that is illiquid paper equity. Value it honestly before it enters the total.

Public-company RSUs are valued near face because they are liquid: take the grant divided by the vesting years at the current price. Private-company options and RSUs get a liquidity haircut, typically 30 to 60% against equivalent public RSUs, because you cannot sell them until an exit. Options carry a second problem on top of illiquidity: they only pay above strike. If the price stays flat or falls, options may expire worthless, so check the strike against the current 409A price before you assign any value.

#### What kind of equity you are actually holding

Horizontal axis runs from Illiquid (private) to Liquid (public). Vertical axis runs from Pays regardless of price (RSU) to Pays only above a strike (option).

| Quadrant | What it means |
| --- | --- |
| Private option | Deepest discount; check strike vs 409A, model time-to-liquidity |
| Public option | Discount for strike risk; value only the in-the-money portion |
| Private RSU | Apply the 30-60% liquidity haircut; taxed at vesting |
| Public RSU | Value near face; grant divided by vesting years at current price |

*Liquidity and payoff structure decide whether equity is cash-like or a bet with a timer.*

Two mechanical facts to keep in the model. RSUs are taxed as ordinary income at vesting, so the pre-tax grant overstates take-home. Standard vesting is four years with a one-year cliff, so first-year realized equity is lower than an even annualization implies. Treat private equity as a bet with a timer, not as cash, and the 30 to 60% haircut stops feeling arbitrary.

> Private equity is a bet with a timer, not cash, which is exactly why the haircut exists.

## How to normalize two offers across metros and tax

Two offers in different cities are not comparable until you convert both to the same purchasing power and then compute after-tax. The published formula is straightforward: equivalent salary equals current salary times the target city index divided by the current city index. The index standard is the BEA Regional Price Parities, where 100 equals the US national average.

The worked example makes the size of the effect obvious. Consider $140,000 in Seattle and $95,000 in Columbus. Seattle's cost of living is roughly 70% higher, so after adjustment the Seattle salary is worth about $82,000 in Columbus purchasing power, substantially less than the $95,000 Columbus offer. Then add tax: Washington has no state income tax and Ohio's rate is 3.99%, which widens the gap further.

| Offer | Nominal | Columbus-equivalent purchasing power |
|---|---|---|
| Seattle | $140,000 | ~$82,000 |
| Columbus | $95,000 | $95,000 |

State tax must be handled as a separate step, because most cost-of-living indexes exclude it. The swing is real: moving from California, with a roughly 9.3% effective rate, to Texas at 0% saves about $9,300 a year on a $100,000 salary before any cost-of-living adjustment. Compute after-tax with federal plus state plus local rates as its own line, then apply the cost index.

> **Watch out:** A blended cost index hides the housing gap
>
> Groceries differ roughly 15% between cities, but housing differs 400 to 600%, and housing is 30 to 50% of a worker's budget. A single blended index averages these and understates the real gap. Pull the housing sub-index alone before you trust any one multiplier.

Housing almost single-handedly determines the value of a relocation salary change, so the blended number is the one that will mislead you. When two offers are in the same metro, this whole section collapses to a single after-tax calculation and you can skip the index ratio.

## Scoring the non-money dimensions on a fixed rubric

Once you have two normalized after-tax numbers, score the parts that are not money on a weighted decision matrix so that two reasonable people would grade the same case the same way. The method is importance weight times score, summed per offer. Assign the weights first, then rate each offer 1 to 5 on each dimension, where 5 is excellent and 1 is poor.

The discipline that makes this trustworthy is the order of operations. Set and lock the weights before you look at either offer's scores. If you tune the weights after scoring, you have built a machine that confirms whatever you already wanted.

**Non-money weighted rubric (lock weights before scoring)**

```
Dimension          Weight   Offer A (1-5)   Offer B (1-5)
Manager quality      0.25          _              _
Scope of role        0.25          _              _
Growth / learning    0.20          _              _
Stability            0.15          _              _
Flexibility          0.15          _              _
-----------------------------------------------------------
Weighted total = sum of (weight x score) for each offer
```

*Adjust the five dimensions and the weights to your own priorities, then do not touch them once you start scoring. Score 1 to 5 each.*

Weights should sum to 1. Salary does not appear in this rubric on purpose, because you already turned money into one comparable number in the earlier steps. The rubric is only for what money cannot capture. When you finish, you have two weighted totals sitting next to your two normalized after-tax figures, which is everything the decision rule needs.

## The tie-break rule that resolves the hard case

Here is the rule the ranking pages omit. If the two normalized after-tax totals differ by more than 10 to 15%, money leads and you take the higher one. If they land inside that band, the weighted rubric decides. This is the whole point of doing the arithmetic: it tells you when to stop trusting the arithmetic.

The threshold comes from a practitioner framework: when total compensation is within 10 to 15% between two offers, non-financial factors typically drive long-term satisfaction more than the salary delta. No regulator or peer-reviewed source sets this number. It is a guide-level convention, so treat the band as a default and widen or narrow it to your own risk tolerance and financial runway.

#### The decision path from two offers to one choice

1. **Two normalized numbers** - After-tax, location-adjusted, equity-discounted total per offer
2. **Compute the gap** - Percentage difference between the two normalized totals
3. **Gap over 15%** - Money leads; take the higher total, sanity-check against timing
4. **Gap under 10-15%** - Weighted rubric decides; the higher weighted total wins
5. **Negotiate the trailing offer** - Anchor and counter before you commit
6. **Re-run and commit** - Recompute with final numbers, then sign

*Money resolves the case only when the gap is wide; inside the band, the rubric decides.*

Whichever side wins, sanity-test it against timing and family constraints before you act. The matrix is a decision aid, not an instruction. A "rigorous" score that violates a hard constraint you already knew about is not a verdict, it is a mistake with a number attached.

## The procedure, end to end

Run these eight steps in order. The first six get you to a provisional winner; steps seven and eight capture the upside most candidates leave on the table.

#### From two offer letters to a signed acceptance

1. **List every component of both offers** - Pull both letters and write down each component per job: base, bonus target versus last year's actual, signing, equity, match, benefit dollar values, PTO, stipends. Done: two column-aligned component lists.
2. **Translate equity to today's dollars** - Public RSUs as grant divided by vesting years at current price; private options or RSUs get a 30-60% liquidity haircut and a strike-versus-409A check. Done: one annual equity figure per offer.
3. **Add to one annual total-comp number each** - Base plus base times realistic bonus percent, plus annual equity, plus signing divided by two, plus benefit dollar values, minus commute. Done: two nominal annual totals.
4. **Normalize for metro and state tax** - Multiply by target-over-current cost index ratio, then compute after-tax with state tax handled separately. Done: two purchasing-power-equivalent after-tax numbers.
5. **Score the non-money dimensions** - Lock weights first, then rate each offer 1 to 5 on manager, scope, growth, stability, flexibility. Multiply and sum. Done: two weighted totals.
6. **Apply the tie-break rule** - If normalized totals differ by more than 10-15%, money leads; inside that band, the rubric decides. Done: a provisional winner with a written reason.
7. **Negotiate the trailing offer before deciding** - Use the leading offer as an anchor and send one counter. Most employers give 5-7 business days; ask once for a one-week extension if needed. Done: a refreshed offer or a firm best-and-final.
8. **Re-run with final numbers and commit** - Repeat the total, normalize, and tie-break steps with post-negotiation figures. If the winner holds, sign. Done: a signed acceptance.

Reconstructing a competing offer's real total from a headline number is tedious, and it is where the comparison usually goes wrong. When you are tailoring applications to the roles that produce these offers in the first place, [Refolk](/candidates) writes your resume from your own history and tailors it to each posting, so more of your energy goes into evaluating offers rather than chasing them.

## Why you negotiate the trailing offer before you decide

Deciding before you negotiate the trailing offer leaves an average of 18.83% on the table. This is the step comparison pages skip, and it is close to free expected value, because a polite counter rarely gets an offer pulled.

The numbers make the case. More than half of candidates, 55%, do not try to negotiate their salary. Of those who do, 66% secure a higher offer, and one separate compilation reports 85% succeed. The average increase among negotiators was 18.83% over their original offers. Negotiation is asymmetric and underused: the downside is a "no," the upside is thousands of dollars, and the base rate of success is high.

**66% - Share of candidates who negotiate and secure a higher offer**

Pew Research Center, as cited across salary-negotiation compilations. Average increase among negotiators: 18.83%.

Use the leading offer as an explicit anchor for the trailing one, and buy yourself the time to do it. Most employers give 5 to 7 business days to respond, and you can ask for a one-week extension once if you need it. The goal is either a refreshed offer that changes the tie-break or a firm best-and-final that confirms it. Either outcome makes your final decision more defensible.

#### The negotiation opportunity most candidates skip

| Stage | Figure | Note |
| --- | --- | --- |
| Candidates who could negotiate | 100 | Everyone holding an offer |
| Candidates who actually negotiate | 45 | 55% never ask (Pew) |
| Negotiators who win a higher offer | 30 | 66% of askers succeed |

*Attrition from "could negotiate" to "did" is where the average 18.83% upside is lost.*

## How this goes wrong: failure modes and false positives

Most bad two-offer decisions come from the same short list of errors. Each has a tell and a check. Work this list before you trust any total.

- **Counting the bonus at target, not actual.** A 20% target that pays out at 60% inflates the total. Check: ask for last year's actual payout percentage and use that.
- **Valuing private equity at headline.** The recruiter's "$300k" may be a third illiquid paper equity. Check: apply the 30-60% liquidity haircut and confirm strike against the 409A price.
- **Ignoring state tax inside the cost index.** Most indexes exclude tax, so a no-income-tax state looks worse than it is. Check: compute after-tax as a separate step.
- **Counting the full signing bonus under a clawback.** Simple clawbacks require full repayment if you leave within a year, and periods run 6 months to 2 years, often pro-rata. Check: read the separate signing agreement for the trigger and term.
- **Reverse-engineering the rubric weights to justify the gut pick.** A "rigorous" score that only confirms a prior. Check: set weights before scoring and do not touch them after.
- **Treating the matrix output as a verdict.** It is a decision aid, not an instruction. Check: sanity-test the winner against timing and family constraints.
- **Deciding before negotiating the trailing offer.** Leaves the average 18.83% upside on the table. Check: send one anchored counter before you commit.
- **Trusting a blended cost index over the housing sub-index.** A blended number hides a 400-600% housing gap. Check: pull the housing sub-index alone.

> **Watch out:** The clawback is a separate document
>
> Signing bonuses commonly carry a clawback that requires repayment if you leave within the covered period. Never count a signing bonus at full value without reading the signing agreement for the trigger and term. A pro-rata one-year clawback changes the annualized number.

The subtle failure is the sixth one. A candidate who has already decided will quietly adjust weights until the matrix agrees, then present the score as objective proof. The fix is procedural, not moral: lock the weights, then score, and let the number land where it lands.

## Keeping the decision honest and current

Before you sign, verify the work rather than trust that you did it right. A good two-offer decision survives someone else re-running it from your notes.

#### Verify before you sign

- [ ] Every offer component is inside the annual total, including benefit dollar values and commute.
- [ ] The bonus is scored at last year's actual payout percentage, not the target.
- [ ] Private equity carries the 30-60% liquidity haircut and options were checked against strike.
- [ ] The signing bonus is annualized over two years and the clawback trigger and term are read.
- [ ] State tax is computed separately from the cost-of-living index.
- [ ] The housing sub-index was pulled on its own, not left inside a blended number.
- [ ] Rubric weights were locked before scoring and never adjusted after.
- [ ] The trailing offer received one anchored counter before the final decision.
- [ ] The final numbers were re-run and the winner still holds against timing and family constraints.

Two inputs go stale and are worth re-checking rather than memorizing. Benefit and compensation-cost shares move: re-pull the BLS ECEC release for the current split rather than reusing the 30.1% figure indefinitely. Cost indexes and state tax rates also change, so recompute both against current data at the moment you decide, not from a saved calculation.

One caution on leverage. Scarcity of your skill moves your negotiating position far more than geography does. In Refolk's index of professional profiles, 580,819 US profiles list Python against 73,991 in the UK, making US supply 7.85 times the UK's, while only 3,377 US profiles list Rust, roughly 0.58% of Python's US supply. If your skill is scarce, you have more room to negotiate the trailing offer; if it is abundant, anchor politely and expect a smaller move.

| Segment | Profiles | Derived ratio |
|---|---|---|
| Python, United States | 580,819 | 1.0x (base) |
| Python, United Kingdom | 73,991 | US is 7.85x UK |
| Rust, United States | 3,377 | Python is 172x Rust |

If you want a read on how contested your position is, look at where people with your exact background actually landed. Refolk turns that into a plain search.

Ask me this: `People who moved from a San Francisco tech job to a Columbus or Dallas employer in the last 18 months.` - [run the search](https://www.refolk.ai/start?q=People%20who%20moved%20from%20a%20San%20Francisco%20tech%20job%20to%20a%20Columbus%20or%20Dallas%20employer%20in%20the%20last%2018%20months.).

*Returns real profiles of people who made the relocation trade you are weighing, so you can see who accepted the lower nominal number and what they did next.*

Run the eight steps once and you will never again stare at two offer letters wondering which is worth more. You will have two numbers, a rubric, a rule, and a counter already sent.

## Frequently asked questions

### How do I compare two job offers when one has a lower base but higher total comp?

Total both offers fully before you judge, because base is only about 70% of what an employer spends on you. Add benefit dollar values, retirement match, realistic bonus, and annualized equity, then normalize for cost of living and tax. A lower-base offer can win by tens of thousands: in one worked case a $150k package beat a $165k one by $33,000 a year once benefits and equity entered. The base line alone routinely misleads.

### What discount should I apply to private-company equity?

Apply a 30-60% liquidity haircut against equivalent public RSUs, because private options and RSUs cannot be sold until an exit event. Sources disagree on the exact figure inside that band, and some also discount further for expected time-to-liquidity. For options specifically, check the strike against the current 409A price, since options pay nothing at or below strike and can expire worthless. Treat private equity as a bet with a timer, not cash.

### How close do two offers have to be before soft factors decide?

The practitioner convention is a 10-15% band on normalized total comp. Inside that band, non-money factors like manager, scope, and growth typically drive long-term satisfaction more than the salary delta, so let your weighted rubric pick. Outside it, money leads and the rubric only sanity-checks. This threshold is a guide-level convention, not a regulated or peer-reviewed number, so treat it as a default to adjust for your own risk tolerance.

### Should I negotiate before choosing, or just take the higher offer?

Negotiate the trailing offer first. Only 55% of candidates ask, yet 66% of those who do secure a higher offer, at an average increase of 18.83%. Offers are rarely pulled for a polite counter, so buying a few days to anchor the trailing employer against your leading offer is close to free expected value. Most employers give 5-7 business days and will grant a one-week extension once if you ask.

### Why does cost of living matter more for relocation than I expect?

Because housing, not the blended index, drives the math. Groceries differ around 15% between cities while housing can differ 400-600%, and housing is typically 30-50% of a worker's budget. A blended cost index averages these and hides the real gap. Pull the housing sub-index alone before you trust any single cost-of-living multiplier, then apply state tax separately since most indexes exclude it.

---

*From the Refolk guide library. I revise these guides rather than replacing them, so the current version is always at https://www.refolk.ai/candidates/guides/two-offer-decision-scored*