Salary Negotiation Is Mostly Decided Before You Open Your Mouth

Before you say a word, the employer has already decided roughly 80% of your number: the pay band, the level they slotted you into, the midpoint they're targeting, and how far above it they'll stretch for a new hire. Negotiation tactics work on what's left. That last slice is real money. It's also small, bounded, and mostly set by choices you made weeks earlier. The 80% is an estimate, derived from how band width and leveling structure box in the final number, not a single measured statistic. The research below is what makes it defensible.

The negotiation advice industry sells you the opposite story. It says the number is up for grabs the moment you're on the offer call, and that the right script, the right pause, the right counter, will move it. The data says the number was mostly fixed before that call started. So the negotiation that moves the most money isn't the one where you counter. It's the one where you pick which company, which level, and which role you walk into.

Does salary negotiation actually work, or is the number already set?

Both, in that order. The number is mostly set, and then negotiation moves a real but limited piece of it.

Here's the structure most people never see. A large employer doesn't price you from scratch. They price the role into a band: a range with a floor, a midpoint, and a ceiling, typically spanning a fixed percentage spread above and below the midpoint, as compensation-benchmarking practitioners like Pave describe. For a new hire, most employers anchor toward the midpoint, not the ceiling. The ceiling usually needs special approval to breach.

So when a recruiter "has room," they mean a few points inside that band. They do not mean the top of the posted range, and they almost never mean above it. Your counter is moving you from, say, the 45th percentile of the band to the 55th. That can be thousands of rupees or dollars. It is not the difference between a junior band and a senior one.

How much does negotiating actually move the offer?

For the people who try, more than you'd think, and less than the books promise.

In a study of roughly 3,858 tech job seekers, those who countered their initial offer got an average increase of 12.45%, about $27,000 a year, per NBER research summarized by UCLA Anderson. That's not nothing. That's a holiday, a wedding, a year of rent.

But notice who that 12.45% is measured on: people who already cleared the loop, already got an offer inside a band, and then pushed within it. The counter moved them up inside the box. It did not change which box they were in.

And most people don't even do that. Only about 30% of U.S. workers asked for higher pay the last time they were hired, per a Pew Research Center survey of 5,775 adults. Of the ones who asked, 28% got the full amount, 38% got something less, and 35% got only the original offer. So two-thirds of askers improved their position. That's a strong case for asking. It's also a strong case for noticing the ceiling on what asking does.

Why does the gap exist before you've even negotiated?

Because the opening offer is already shaped by your level, your market slot, and patterns the employer applies before you say anything.

A Columbia Business School analysis of more than 700,000 initial salary offers found a 5.5% gender gap in those first offers, net of detailed controls for job, employer, occupation, industry, location, and human capital, published in Organization Science. Read that again. The gap shows up in the opening number, before a single round of back-and-forth, and it's wider in male-typed occupations.

That's the tell. If outcomes diverged purely on negotiation skill, the offers would start equal and split during the conversation. They don't. They start unequal. The structure does the work upfront.

It compounds at the firm level too. In German employer-employee data covering 772 employers and about 10,000 workers, firms that allow individual wage negotiation show a 5% gender pay gap, while firms that don't allow individual negotiation show none, per Upjohn Institute research. Same study: employers vary pay 6 to 12% between similar workers. The room to negotiate is the same room where gaps grow.

What does the "ask gap" really tell us about tactics?

That what you ask for matters enormously, and that your ask is itself downstream of information you usually don't have.

Economist Nina Roussille studied Hired.com and found women's ask salaries ran 2.9% below comparable men's after controlling for resume characteristics. That single ask gap explained the entire residual gender gap in final offers, in her Quarterly Journal of Economics paper. The number you put in the box becomes the number you get.

Then comes the part that should change how you think about all of this. When Hired.com stopped showing a blank ask field and instead prefilled it with the median bid for comparable candidates, the ask gap, the bid gap, and the final offer gap all went to zero. Women received no fewer bids and no fewer offers as a result, same paper.

Nobody learned to negotiate better. Nobody memorized a script. The platform just told people what comparable candidates were getting. The variable wasn't tactics. It was information. People who knew the market priced themselves at the market. People who didn't, didn't.

That's the whole game in one experiment. The asker who knows the band asks for the band. The one who's guessing leaves money on the table before the conversation starts.

Does pay transparency give you more room to negotiate?

Usually less. This is the irony nobody warns you about.

You'd assume a posted range hands you ammunition. Sometimes the opposite happens. When employers must disclose ranges, many respond by anchoring publicly and holding firmer, because paying one worker high now means renegotiating with everyone else later. Across markets studied, wages declined about 2% under transparency, per Cullen and Pakzad-Hurson in Econometrica. The posted range becomes a wall, not a doorway.

It's not all downside. After Colorado's transparency law, disclosed salaries rose 3.6% and actual earned salaries rose 1.3% across the board, and Colorado women's relative wages climbed from 80 cents to 85 cents per male dollar, per a UC San Diego and USC study reported by Colorado Newsline. Transparency lifts the floor for people who had the weakest hand under secrecy. It just doesn't widen the room for the individual negotiator.

And the ranges aren't always honest. About 24% of postings covered by transparency laws still don't include salary info at all, even where disclosure jumped 20 percentage points after the laws passed, per the New York Fed's Liberty Street Economics.

The transparency trap, concretely. A job seeker in New York City sees a posted range of $90K to $140K and prepares to push toward $135K. What they don't know: the employer posted the top to comply with the law, is targeting $100K, and reads the range as a demand anchor, not an invitation. Many such employers won't go above roughly $105K for any one person, because doing so can trigger renegotiation pressure across the team. The $140K was never real money. The reader spends two weeks preparing for a fight over a number that doesn't exist.

What actually moves the number, then?

Three things, and all of them happen before or around the offer, not in the counter.

1. The level you enter at. This is the biggest lever and the most ignored.

Move	Where it happens	Typical impact
Counter the offer	Offer call	Up within the band, often single-digit %
Get leveled up (e.g. L4 to L5)	During the interview loop	Whole new band, often a much higher midpoint
Enter a higher-tier company	Application stage	Different band architecture entirely
Bring a competing offer	Just before the deadline	Grounds for a band exception above midpoint

Weak vs strong. Weak: a candidate spends two weeks memorizing negotiation scripts before a Google offer call, counters by $10K on a role the recruiter already budgeted to the L4 midpoint, and gets $5K. Strong: the same candidate uses leveling data and a Praxy read of the role to argue for L5 during the loop. L5 carries a band midpoint tens of thousands higher. The negotiation call barely matters, because the level change already did the work.

2. A competing offer. A mid-level engineer gets $210K from Company A at the band midpoint, then discloses a $235K offer from Company B. Company A comes back at $228K. Not because the recruiter admired the tactic. Because the competing offer gave the hiring manager grounds to request an exception above the standard midpoint. The competing offer bought a band override. That's the mechanism the NBER counters were riding on too.

3. What anchors your band on the way in. In a market without a salary history ban, a candidate who discloses $95K gets slotted into the lower third of a $95K to $130K band, because their history anchors the number. The same candidate in a jurisdiction with a history ban gets offered more for the same job, because the employer has to price on market data instead of the candidate's past. What you reveal, and what you're forced not to reveal, sets the band before any counter.

What's the honest counterpoint to all this?

That negotiation does work at the margin, and the margin is worth real money to the minority who use it.

The 12.45% from the NBER study is genuine. The 66% of Pew askers who improved their outcome is genuine. The downside of asking, politely and once, is close to zero, and the upside is a meaningfully bigger number. If you take nothing else from this: when you get the offer, counter. Two-thirds of people who ask get more.

The structural argument is also strongest for big, formal employers. At a Fortune 500 or large tech firm, the band is rigid and you're a slot. At an early-stage startup, comp is often improvised, and a single conversation can move the number 20 to 30%. Senior and staff roles, principal engineers, VPs, also carry wider bands and more exception authority. The box is tightest in the middle of the ladder at large companies, and loosest at the extremes.

So negotiate. Just don't believe the counter is where your salary is decided. It's the polish, not the sculpture.

What to do now

Spend your energy where it actually moves the number, in this order:

Before you apply, learn the band. Find out what the level you're targeting actually pays at the companies you're targeting. The Hired experiment proved it: people who know the market price themselves at the market. Guessing is the expensive part. If you don't have a competing offer to anchor against, there's still a way to ask for the right number.
During the loop, fight for the level, not the offer. Argue scope and impact while leveling is still open. One level up beats any counter you'll make later, and the level you settle into early is the one that quietly compounds for years.
Pick the company whose band fits you. A higher-tier employer's floor can sit above another's ceiling. The company choice is a salary decision, and the same role can pay multiples more depending on where you sit.
Generate a competing offer before the deadline. It's the one thing that reliably buys a band exception. Line up your timelines so two processes finish close together.
Then counter, once, politely. Two-thirds of askers improve their number. Just go in knowing the size of the box you're in.

The point isn't to stop negotiating. It's to stop treating the offer call as the whole fight when most of it was decided weeks before. Understand the box before you're inside it.

Want to know what band you're actually walking into before the offer call? Tell me the role, the level, and the companies you're chasing on WhatsApp, and I'll help you read the box, push for the level, and time a competing offer. That's the negotiation that moves the number.