The Screen-Anchored Pay Number, Traced to the Offer

You gave the recruiter a number in the first ten minutes of the screen. Now it is somewhere in a system, and you can feel it moving ahead of you through the loop. This guide is for the candidate who wants to see exactly where that number goes, which round it hardens in, and whether there is still a soft point left to move it before the offer lands. I carry one illustrative candidate all the way through, with real benchmark figures at each fork, including the misstep.

The candidate: mid-level, states "$115k" on the recruiter screen. The posted band is $110k to $140k. Midpoint is $125k. Watch that "$115k" travel.

Where the number is first requested and how it travels

The expectation question is dropped early, usually in the first ten minutes of the recruiter phone screen, and sometimes earlier still on the application form itself. From that moment it is a record that follows you, not a passing remark.

A recruiter asks it almost offhand, and a lot of good candidates blurt a figure they picked in the shower and lock themselves in for the rest of the process. On application forms the constraint is mechanical: modern Applicant Tracking Systems scan salary fields for valid numerical values, and an entry of "0", "Negotiable", or any non-numeric text in a number-only field often triggers automatic rejection before a recruiter ever sees your name. Then the number is reused. Recruiters refer back to the range you shared in that first screener to craft the offer.

So the answer to "why does my screen number matter so much" is that it is not a conversation, it is a data capture. The figure you say is logged and becomes the reference point for a document written weeks later.

For our candidate, "$115k" is now the seed. It sits $10k under midpoint. Everything downstream grows from it.

Why the number hardens between the screen and the offer

The number hardens in the ATS at capture, not at the offer call. By the time you are on the offer call you are countering a figure that was derived from your own screen number, which is the hardest kind of number to move.

This is the single insight most published advice misses. Every list of deflection scripts treats the offer as the negotiation event. But the mechanism in the sourcing is clear: the figure is captured at the screen and reused to draft the offer. The cheap edit - changing a note in a recruiter's file before anyone commits it to a letter - is only available in the middle of the loop. After the offer is written, you are asking someone to unwind work they have already done and, often, already circulated for approval.

That ratio is the shape of the table. The person capturing your number does this every week; the average candidate does it a few times a decade. In Refolk's index of professional profiles, US profiles listing Salary Negotiation as a skill number 1,930 against 110,470 with a recruiter or technical recruiter title. The counterparty is far better staffed for the conversation, which is exactly why you have to control the number at the point it is cheap to control.

How much your stated number actually constrains the offer

The low number you name caps your ceiling even when the budget was higher, and initial offers already cluster below midpoint. Combine the two and a below-midpoint statement lands you near your own floor.

There are two documented anchors working against you at once. First, offers start low by default: it is not uncommon for an employer to begin a new employee 10 to 15 percent below the salary midpoint. A worked benchmark puts the initial offer around $63,750, which is 85 percent of a $75,000 midpoint. Second, if you name a number that is too low, you cap your offer at that number even if the budget was higher.

Here is the compression on our candidate. Midpoint $125k. A default offer at 85 percent of midpoint would be roughly $106k. But the candidate said "$115k", so the draft is built off that, likely landing between $106k and $115k - well under the $140k band top and under the $125k midpoint. Two forces squeezed the same person from both sides.

The last row is the slack you are leaving on the table. A CareerBuilder survey found 52 percent of employers typically offer less than they are willing to pay, and 26 percent said their opening number is $5,000 or more below what they would actually approve. That gap is real money that a below-midpoint screen number quietly forfeits.

A below-midpoint number said in the shower compresses both ends of the offer at once.

Does negotiating move the number, and by how much

Negotiating moves the number for most people who try, and the premium is well documented. The problem is not whether to push; it is that few candidates do, and they leave measurable money behind.

Set that against behavior. Roughly 55 percent of candidates accept the first offer without negotiating, and Pew's survey of 5,188 workers found only about 30 percent asked for higher pay when they were last hired. The UCLA Anderson field experiment with about 3,858 tech candidates who countered found an average win of 12.45 percent, roughly $27,000 a year. The upside is documented and large; the take-up is low.

That is the asymmetry to hold in your head: the odds favor the person who pushes, but the counterparty is better staffed and most peers do not push, which is why a clean, evidenced reset lands so much better than the deflection scripts everyone has already read.

The procedure: trace and reset one number

This is the full thread, from the screen where the number is set to the ceiling where base stops moving. Lay your own loop against it and mark which step you are standing on right now.

Step 1 in practice: the deflect-and-flip

You are not refusing to answer, you are redirecting to information that helps both parties, and many recruiters will share the range when asked this way. If they hold firm, give a range - and make the bottom of that range your target, not your floor.

Deflect: "Happy to talk numbers. Since you know the budget for this role better than I do, what range have you got approved for it? I'll tell you honestly whether that works."

If pushed for a number: "Based on the scope as I understand it and market data for this role, I'm targeting the [target]-to-[anchor] range, where anchor is 10-15% above target. I'd want to confirm scope before locking that in."

Our candidate should have said the second line with a range like "$128k to $145k", not the single "$115k". That one substitution changes the seed the whole offer grows from.

Step 3 in practice: the mid-loop re-open

This is the reset window, and it is the whole point of the guide. If asked again at the hiring-manager stage, it is reasonable to say: "I provided a range earlier in the process - has that changed anything on your end?" That reopens the conversation without sounding like you are reneging.

"Thanks for the [round] - it clarified a lot. Now that I understand the [specific scope: ownership of X, the on-call load, the team size], the role is broader than what I calibrated my earlier number against. Based on that scope and current market data for the role, I'd update my target to [new number]. Wanted to flag it now rather than at the offer stage so we're aligned."

The phrase "so we're aligned" matters. You are framing the update as coordination before a document exists, which is exactly the cheap edit the mechanism allows.

What to cite, and where citing stops working

Cite public wage sources to justify a number, but know the hard stop: external data stops working the moment it would break the employer's internal consistency. After that, only total comp moves.

The named public sources career centers point to are Glassdoor, Payscale, Salary.com, and the Bureau of Labor Statistics wage data. Pay-transparency laws add posted ranges to that arsenal: as of 2026, 18 US states plus Washington, D.C. have enacted pay-transparency laws requiring employers to disclose salary ranges, with penalties from $100 to $250,000 per violation. New effective dates include Virginia on July 1, Maine on July 29, and Connecticut on October 1, so re-check your state before you cite a range as binding.

But do not treat the posted band as the true ceiling. Some published ranges cover only 25 to 75 percent of actual salaries, so corroborate the posted number against BLS, Glassdoor, or Payscale before you build a case on it.

The wall is internal equity. Paying a candidate much more than the listed range can create legal exposure if it is inconsistent, and disclosure requirements cover existing employees too, so employers must stay consistent internally rather than bring in an exception. Once base hits that hard ceiling, pivot immediately to total compensation and future-raise mechanics.

Refolk tailors your resume to every posting and scores how well you actually fit, which matters here because a fit-and-scope argument is far more credible when your materials already show the scope you are claiming. Refolk builds that fit case from your own history so the reset conversation has evidence behind it.

How this goes wrong: the failure modes

Most resets fail in one of eight predictable ways. Each has a tell for when the signal is lying to you, and each has a check you can run before you act.

• Deflection that reads as evasion. The recruiter marks you "guarded" and ends the screen. Give one clean deflection, then a range if pushed - do not stonewall past the first ask.
• The range trap. State target-to-floor and they hear the floor; they tend to hear the low end. Always state anchor-to-target, never target-to-floor.
• Reset attempted too late. Once you are on the offer call the number has hardened. The check: did you re-open before an offer was drafted? If not, the soft window is gone.
• A reset that sounds greedy, not evidenced. The ask sounds entitled instead of evidence-based, which makes the employer defend the midpoint or lower end more aggressively. Tie every new number to scope, market, or leverage.
• Citing market data past the internal-equity wall. External data stops working when it breaks internal consistency. If base will not move, pivot to total comp instead of repeating the comp.
• ATS field self-elimination. "Negotiable" or non-numeric text auto-rejects before a human sees you. Enter a numeric minimum-acceptable anchor.
• Trusting the posted band as the real ceiling. Ranges may reflect only part of actual pay. Corroborate against BLS, Glassdoor, or Payscale.
• Panic-accepting the first offer. About 55 percent do, leaving documented money behind. Counter once before you sign.

The most expensive of these is #3, reset attempted too late. It is expensive because it is invisible - nothing goes wrong on the call, you simply never opened the cheap window, and you spend the offer call fighting a number you seeded yourself.

Read that funnel as an instruction: the biggest leak is not losing a negotiation, it is never entering one.

Geography and the state of the field

Your leverage depends on where you are. US candidates have both more peers modeling negotiation and more public anchors to cite than candidates in markets with less disclosure and thinner negotiation-skill supply.

In Refolk's index, the US shows roughly 12.6 times the UK's supply of candidates listing salary negotiation as a skill. Pair that with 18 US states now mandating range disclosure and US candidates have more public anchors to point at and more peers modeling the behavior. If you are outside those markets, lean harder on BLS-equivalent public data and on the scope argument, because the posted-range lever may not exist for you.

Before you call the reset done

Run this list before you accept anything. It is the difference between tracing the number and being dragged by it.

What to do next

If you are still on the screen: use the deflect-and-flip, and if pushed, give an anchor-to-target range. That is the one moment where the number is fully in your hands.

If you already gave a low number and are mid-loop: this is your window. Send the reset message tied to the scope you have now learned, and get the updated number acknowledged in writing before an offer is drafted. Do not wait for the offer call to fix it.

If you are already on the offer call: the screen anchor has hardened, but you are not out of moves. Counter once with a defensible number, and when base stops moving, pivot to signing bonus, equity, and review timing - and capture every concession in writing. The evidence says countering works for most people who try it; the only unforced error left is not trying.

Keep this document open against your own loop and mark the step you are on. The number is already moving. The question is whether you are moving it or watching it.