# Decoding One Posting's Pay Range, From Wide Band to Target Ask

*You will take one posted salary range, classify it into one of five types, find the real target-level band inside it, and name a number you can defend.*

- Canonical URL: https://www.refolk.ai/candidates/guides/decoding-one-posting-pay-range
- Pillar: Reading the market
- Format: Teardown
- Published: 2026-08-11
- Last reviewed: 2026-08-11
- Reading time: 16 min

You are looking at one posting. It says $130,000 to $500,000, and before you name a number or even decide to apply, you need to know what that band actually pays and where inside it to aim. This guide carries that single listing all the way through: the arithmetic, the classification, the public-data benchmark, and two wrong turns I want you to see me take and back out of. By the end you can run the same passes on your own posting and land on a number you can say out loud without flinching.

This is the Tuesday task pay-transparency laws created. The laws made employers post a range; they did not make the range honest, and most did not even define how wide it can be. So the work fell to you.

## Why a posted range needs decoding at all

A posted range is a legal artifact, not a price. Most transparency statutes require only a "good faith" range the employer honestly expects to pay, and outside New Jersey's proposed rule, none caps how wide it can be. That gap is the whole problem: the law that was supposed to make pay legible often produces a band so broad it hides the number it was meant to reveal.

The scale of the shift is real. As of recent Indeed data, roughly 60% of postings included salary information, up from 18% a few years earlier. So you will see ranges constantly now, and the median one is fine. The trouble is the tail: the $50,000-to-$550,000 listings that technically comply while telling you nothing.

**60% - Share of Indeed postings that included salary info, up from 18% a few years prior**

Ranges are now the norm, which is exactly why decoding the bad ones is a repeatable task.

Our worked example, $130,000 to $500,000, is one of those. I will treat it as a real listing for a senior engineering role and take it through every pass. Follow along on your own.

## Reading one range: capture the width first

Start with two numbers you can compute in under two minutes: the spread and the multiple. The spread is (max - min) / min. The multiple is max / min. These two figures decide almost everything that follows, because width is your first and cheapest signal.

For $130K to $500K:

- Spread: (500 - 130) / 130 = 285%
- Multiple: 500 / 130 = 3.8x

Now compare against what a normal range looks like. The compensation platform Pave, measuring width as (max - min) / midpoint, finds the modal posted band is 30% wide, meaning roughly plus or minus 15% around the midpoint. Glassdoor data reported by CNBC shows most posted ranges are actually narrower than plus or minus 20%. And Glassdoor economist Daniel Zhao offers a rule of thumb any practitioner can carry: a range where the top is about 1.5x the bottom is reasonable.

| Benchmark | Value |
|---|---|
| Modal posted width (max-min)/mid | 30% |
| Companies at 20% / 30% / 40% width | 23% / 38% / 27% |
| Zhao "reasonable" max multiple | 1.5x min |
| NJ proposed statutory cap | 60% of min |
| Worked example $130K-$500K | 285% spread, 3.8x |

Note how close the two independent ceilings sit. Zhao's 1.5x multiple maps to a 50% spread; New Jersey's proposed cap is 60% of the minimum. Two unrelated benchmarks land near the same wall. Our example, at 285% and 3.8x, is not near that wall. It is in a different building.

> **Rule:** The 1.5x screen
>
> If the top of a posted range exceeds about 1.5x the bottom, stop treating the midpoint as a target. A range past that ceiling is either multi-level or gamed, and either way the midpoint is meaningless for you.

## Why the width is legal, and why that matters to you

The width is legal because most statutes never defined width. That is not a loophole employers found; it is the loophole the laws left open. Understanding this stops you from reading a wide range as a mistake you can call out. It is a design choice, and your job is to see through it, not to protest it.

The one exception worth knowing is New Jersey's proposed rule, which would cap the spread at 60% of the minimum. Under it, $100,000 to $160,000 complies and $100,000 to $200,000 does not. It is the first state to prescribe width, and it is still proposed, not final. Colorado, separately, bans open-ended ranges like "$60,000 and up." Everywhere else, "good faith" is the only constraint, and good faith is not measurable from the outside.

| State | Employer threshold | Model | Width rule |
|---|---|---|---|
| Colorado | 1 employee | Range in posting | No open-ended ranges |
| New York | 4 employees | Range in posting | Good-faith only |
| California | 15 employees | Range in posting | Good-faith only |
| New Jersey | 10 employees | Range in posting | 60% cap (proposed) |

The practical read: in most of the country, a 3.8x range breaks no rule. So do not wait for the employer to explain it. You decode it yourself, then you decide.

> Transparency laws made employers post a range, not tell the truth, and the width is where the difference hides.

## The five range types, and the signal that names each

There are five types a posted range can be, and you classify by combining width with the posting's features. No single source enumerates all five; this taxonomy is synthesized from several, and the distinctions matter because each type demands a different next move.

- **Target-level band.** A 20% to 40% spread around a single job level. This is the honest, common case. The midpoint is meaningful. Aim near it, adjusted for your experience.
- **Multi-level / broadband.** Spans junior to senior, often with explicit "L4 to L6" or level cues. A tech listing that covers three levels can be genuinely wide and genuinely honest. The top is reachable if you are targeting the top level.
- **Compliance-cynicism.** Absurdly wide, like $50,000 to $550,000, with no explanation. Built to satisfy the law while hiding real numbers. The width is the tell, and the absence of any stated reason confirms it.
- **Truncated lowball.** The posted band sits below the true internal band, with the upper end hidden. Here the danger is anchoring low, because even the posted top may undershoot what the role actually pays.
- **Genuine flexibility.** Moderate width explained by geography or "depends on experience," where multiple locations or seniority reasonably widen the band.

The classifier that beats width is the stated reason. Research found that attributing a wide range to cost of living raised applicant attraction, while attributing the same width to candidate qualifications lowered it. So the exact same numeric spread can be genuine-flexibility or compliance-cynicism depending on what the posting says about it.

#### Classifying a posted range

Horizontal axis runs from Width within norm (under 1.5x) to Width absurd (over 1.5x). Vertical axis runs from Reason stated (geo, levels, DOE) to No reason given.

| Quadrant | What it means |
| --- | --- |
| Target-level band | Trust the midpoint, aim by experience |
| Genuine flexibility / multi-level | Read the reason, benchmark the level you want |
| Truncated lowball risk | Cross-check external data, the top may be low |
| Compliance-cynicism | Discard the midpoint, benchmark from scratch |

*Width tells you how far apart the numbers are; the stated reason tells you why, and the why is the better classifier.*

For our $130K-to-$500K example, the tie is between multi-level broadbanding and compliance-cynicism. Some sources would call a very wide tech band honest broadbanding; others call it malicious compliance. The level cues in the posting break the tie. If the listing names levels or says "Senior to Staff," lean multi-level. If it is a single title with no level language and no geographic note, lean compliance-cynicism.

## The procedure, start to finish

Here is the full pass, the same one I ran on the example. It takes about 90 minutes the first time and half that once the moves are familiar.

#### From posted range to defensible ask

1. **Capture the range and compute width** - Write down min and max, then compute (max-min)/min and max/min. For $130K to $500K that is 285% spread and 3.8x.
2. **Screen against reasonableness thresholds** - Compare to the 30% modal width, the 60% cap, and Zhao's 1.5x rule. The example fails all three.
3. **Read the posting features** - Note DOE language, single vs multiple ranges, level or L-number cues, geographic notes, and tiered structures.
4. **Classify into one of five types** - Combine width and features into one label; the level and geography cues break any multi-level versus compliance tie.
5. **Pull public benchmarks for the target level** - Match your SOC occupation, pull metro OEWS 25/50/75/90 percentiles, cross-check a level reference for your specific level.
6. **Locate the target-level band inside the posting** - Overlay your external band on the posted range; the overlap is your realistic zone.
7. **Map yourself to a percentile** - Entry near 10th to 25th, deep experience near 75th to 90th, with a documented rationale.
8. **Set the defensible ask** - Anchor on the benchmark, not the midpoint; apply the +25% floor bump only if width already failed the screen.

## Benchmarking the true target level against public data

Benchmark to one occupation, one metro, and the median, or your number will drift. The BLS Occupational Employment and Wage Statistics program (OEWS) publishes annual wage percentiles at the 10th, 25th, 50th, 75th, and 90th for roughly 830 occupations by state and metro. That is your spine.

The documented procedure is precise about what not to do:

1. Match your role to the single closest SOC occupation. Do not combine codes.
2. Pull metro-level percentiles, not national, so cost of living is baked in.
3. Read the 50th percentile as "typical," not the mean.
4. Cross-check the specific level on a level reference such as Levels.fyi for tech, plus Glassdoor or LinkedIn.

Two of those steps prevent the most common self-inflicted errors. BLS explicitly warns against averaging percentiles across combined occupations, because the arithmetic is invalid. And the mean is skewed upward by top earners, which is the quiet reason self-built benchmarks run high. The median is the honest anchor. Employers themselves commonly build their ranges off the 25th-to-75th percentile band, so if you know your metro's 25th and 75th, you roughly know the shape of an honest employer band for that level.

For the example, suppose your metro OEWS puts the target level's 50th percentile at $190K and the 75th at $230K, and a level reference for the same senior level clusters around $195K to $240K in total cash. That external band, roughly $190K to $240K, is your reality. Overlay it on the posted $130K to $500K. The overlap starts at $190K, not $130K, and does not extend to $500K. Your realistic zone just went from a $370K span to a $50K one.

> **Watch out:** Do not average across SOC codes
>
> If your role straddles two occupations, pick the closer one and use it alone. Averaging their percentiles produces a benchmark BLS itself says is invalid, and you will carry the error into your ask.

Refolk does this cross-check for you against its own index rather than a static table. If you have a resume in it, Refolk scores how well you fit a given posting and where your history places you against comparable profiles, which turns "roughly the 75th percentile" into a number tied to your actual record. Try that when the manual overlay leaves you unsure which level you are really benchmarking. You can start at [Refolk](/candidates).

## The two wrong turns I want you to watch

Two mistakes are so common they deserve to be walked through, not just listed.

**Wrong turn one: anchoring on the midpoint.** My first instinct on the example was to split the difference. $130K plus $500K over two is $315K, so aim there, right? No. A midpoint is only meaningful when the range is tight enough to have a real center. At 3.8x there is no center, only two numbers far apart. Anchoring on $315K would have had me either overshoot a role that tops out near $240K in reality, or, worse, let the employer talk me down from a fantasy figure I could not defend. The fix is the 1.5x screen: because the spread failed it, I discarded the midpoint entirely and benchmarked from scratch.

**Wrong turn two: reading the width as personal upside.** My second instinct was that $500K top meant real headroom, so I should aim high inside the posting. But the width here was not headroom; if it was multi-level, the $500K belonged to a Staff or Principal level I was not targeting, and if it was geographic, the top belonged to a high-cost metro I was not in. Either way the top was not mine to claim. The fix was to benchmark my specific level and metro, which is why the external band mattered more than the posted ceiling.

Both wrong turns share a root: letting the posting's numbers set the frame. The discipline is to build your frame from public data first, then use the posting only to check where your number lands.

## How this goes wrong: the failure modes

Most bad asks come from one of these eight errors. Each has a false positive and a specific check.

| Failure mode | What it looks like | The check |
|---|---|---|
| Anchoring on the midpoint | Treating $315K as the target on a 3.8x range | Does the spread exceed 1.5x? If yes, discard it |
| Mistaking multi-level for lowball | Assuming a wide tech top is unreachable | Look for explicit level or "L#" cues |
| Averaging BLS percentiles | Combining SOC codes into one benchmark | Use one detailed occupation, one metro |
| Using the mean, not the median | Benchmarking to an inflated "typical" | Use the 50th percentile as typical |
| Treating "DOE" as a number | Inferring a floor the posting never stated | Pull external data; in transparency states, request the range |
| Reading geographic width as upside | Expecting the top in a low-cost metro | Check for listed locations, use metro OEWS |
| Assuming the posted top is the ceiling | Believing the range caps what the role pays | Cross-reference a level source for the same level |
| Bumping an honest range | Applying +25% to a tight, good-faith band | Only bump if width already failed the screen |

Two of these deserve emphasis because they push in opposite directions. The +25% floor bump comes from one named practitioner's rule: when a range is deliberately too broad to be meaningful, raise your minimum ask by 25%. It is a single practitioner's rule, not an established or peer-reviewed figure, so treat it as a lever, not a law. And it is only for gamed ranges. Apply it to a tight, honest band and you price yourself out of a role you wanted. The screen you ran in step two is exactly what tells you which case you are in.

## When the market moves your leverage without moving the range

The posted range is only half the negotiation; the depth of the talent pool is the other half, and it is invisible in the listing. A scarce market lets you push toward the upper band a posting hides, even when the numbers on the page never change. This is where Refolk's index says something no posting will.

| Market | Current "Software Engineer" profiles | Top metro share | Named top employer |
|---|---|---|---|
| United States | 348,552 | San Francisco (dispersed) | Google |
| United Kingdom | 42,944 | London ~40% of sample | Deloitte Digital / JPMorgan |
| US-to-UK ratio (derived) | 8.1x | - | - |

In Refolk's index the US Software Engineer pool is 8.1x the UK's. In the scarcer market, a candidate can credibly push toward the top of a band the posting keeps vague, because the employer has fewer alternatives. Geographic concentration compounds it: with roughly 40% of UK Software Engineer profiles in London, a London posting's "wide" range often reflects one expensive metro rather than multi-level breadth, so the top is more reachable there than the width alone implies. That is a classification clue and a leverage clue at once. Read a wide London range as likely geographic-then-scarce before you read it as gamed.

**8.1x - How much larger the US Software Engineer pool is than the UK's in Refolk's index**

348,552 US profiles against 42,944 UK; a thinner pool moves your leverage even when the range does not.

To see the pool for a specific market and level yourself, run the search against real profiles rather than guessing at scarcity.

Ask me this: `UK-based backend engineers in London with 8+ years experience at scale-ups` - [run the search](https://www.refolk.ai/start?q=UK-based%20backend%20engineers%20in%20London%20with%208%2B%20years%20experience%20at%20scale-ups).

*Returns the current pool of senior London backend engineers, so you can judge how thin the market is before you push toward the top of a posted band.*

## The final ask, and the one line that defends it

Land on one number and one sentence. For the example, my external band was roughly $190K to $240K, I placed myself at the 75th percentile on deep experience, and I judged the range gamed under the 1.5x screen. That put my ask near $235K, with the +25% bump considered but not fully applied because my benchmark already sat high inside the posting.

**The one-line ask justification**

```
For a senior [role] in [metro], the market median runs about [50th pct] and the 75th about [75th pct]. Given my [years] and [specific evidence], I'm targeting [number], which sits inside your posted range.
```

*Fill in your occupation, metro, percentile, and number, then say it aloud until it sounds like a fact, not a hope.*

The sentence works because every clause is checkable. It names a median, a percentile, and a rationale, and it points at the posted range rather than fighting it. That is the difference between an ask and a wish.

## Keeping this current

OEWS updates annually with a May reference period, so pull fresh percentiles each year rather than trusting last cycle's numbers. Transparency law is moving too: New Jersey's 60% cap was proposed, not final, and state counts vary from 14 to 18 plus DC depending on the source, so re-check your state before you assume a range is out of bounds. And re-run the pool search when your target metro or level changes, because scarcity is the input that moves fastest and shows up nowhere in the posting.

#### Before you name a number

- [ ] You computed the spread and multiple and screened them against 1.5x
- [ ] You classified the range into one of the five types with a stated reason
- [ ] You benchmarked one SOC occupation, one metro, at the 50th and 75th percentiles
- [ ] You overlaid the external band on the posting and found the overlap zone
- [ ] You placed yourself at a percentile with a documented rationale
- [ ] You applied the +25% bump only if the range already failed the width screen
- [ ] You can state your ask in one checkable sentence out loud

Run this once and it feels slow. Run it three times and it becomes the fast, quiet thing you do before every application that posts a range, which is now most of them.

## Frequently asked questions

### What does a wide salary range mean on a job posting?

A wide range means the employer either genuinely spans several job levels, is complying with a transparency law without revealing real numbers, or is hiding the true target band. Width alone does not tell you which. Screen the spread against the 1.5x rule of thumb, then read the posting for level cues, geographic notes, and DOE language. The stated reason for the width classifies it better than the number itself.

### Is the posted salary range real or just for compliance?

Most transparency statutes require only a good-faith range the employer honestly expects to pay, and none except New Jersey's proposed rule caps width, so absurdly broad ranges can be technically legal while hiding the real numbers. A spread near the 30% modal width or under about 1.5x reads as genuine. A 3.8x multiple like $130K to $500K is outside every documented norm and should be treated as gamed until the posting explains the width.

### Where should I aim inside a posted pay range?

Aim at your externally benchmarked band for the specific level you target, not the posted midpoint. Pull metro-level OEWS percentiles for your occupation, cross-check a level reference, and overlay that band on the posting. The overlap is your realistic zone. Then place yourself by experience, near the 25th percentile as a new entrant or the 75th to 90th if deeply experienced, and set your ask there.

### Should I use the posted midpoint as my target number?

No, not when the range is wide. A $130K to $500K midpoint of $315K is meaningless because the spread exceeds every reasonable threshold. Anchoring on it either overshoots or lets the employer pull you down. Build your target from public market data first, benchmark to the median rather than the mean, and use the posted range only to check where your defensible number lands inside it.

### How do I benchmark a posted range against public data without getting it wrong?

Match your role to one detailed SOC occupation, pull OEWS percentiles for one metro, and use the 50th percentile as typical. Do not average percentiles across combined occupations, which BLS explicitly warns against, and do not use the mean, which is skewed upward by top earners. Cross-check the specific level with a tech-level reference for roles that use levels.

---

*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/decoding-one-posting-pay-range*