The Counteroffer Risk Read: Close Clean, Pre-Empt, Price In, or Run a Backup
Score any finalist on 11 sourced dimensions and land on one of four counteroffer verdicts, each with a set action and deadline, before the offer goes out.
Key takeaways
- Score counteroffer risk on both sides: six candidate dimensions decide whether a counter would be enough, and five employer dimensions decide whether a counter will come and how hard.
- The most common counteroffer is a 10% raise, while job seekers say they need about 33% to stay, so a typical counter covers only about 30% of what it would take.
- Departures make counters more likely, not less: Myers Industries gave its remaining eligible executives retention awards worth 75% of base salary on the day its CEO departed.
- In Refolk's index only 19.5% of US Site Reliability Engineers list both Kubernetes and Go, and a backfill pool that thin should raise the employer-side counter score.
- Posted salary ranges contain actual pay only about 67% of the time, so treat the incumbent's range as a ceiling and score any range wider than 60% of its minimum as unknown.
- The 80% counter-acceptor attrition figure has never been traced to a study; the best-sourced figure is CEB's 50% leaving within 12 months, with undisclosed methodology.
I run this read before an offer goes to a finalist: how likely is their current employer to counter and keep them, and what should the offer and the slate look like given that risk. It is for in-house recruiters, sourcers, talent leaders, and founders who close their own hires. You score six candidate-side and five employer-side dimensions from named evidence and land on one of four verdicts, each with an action and a deadline.
What does a counteroffer risk read decide?
A counteroffer risk read decides how to structure the offer and the slate before the candidate resigns, by landing on one of four verdicts: Close Clean, Pre-Empt, Price In, or Run a Backup. A counteroffer is any retention offer the current employer makes once it learns the person is leaving, usually after the resignation is handed in.
Most advice on how to predict if a candidate will accept a counteroffer reads only the candidate, and mostly their interview demeanor. That is half the case. The employer decides whether to counter and how hard; the candidate decides whether that counter is enough. This framework scores the two sides separately, then combines them.
The four verdicts:
- Close Clean. Low risk on both sides. Extend the offer on the normal timeline.
- Pre-Empt. The candidate is anchored, but the employer has strong reasons to counter. Do the resignation work before the offer goes out.
- Price In. The candidate is movable on money, but the employer is unlikely to spend big. Build the offer so a typical counter cannot match it.
- Run a Backup. Both sides are high. Make the offer, and take the runner-up to offer-ready in parallel.
The stakes are measurable. In Robert Half's survey, 28% of professionals said they had backed out of a job after accepting it, and 27% of those cited an acceptable counteroffer. A renege is a candidate withdrawing after accepting. For entry-level hiring, NACE puts reneges at about 7% of accepted offers. Act on the risk before the offer goes out, not after the counter lands.
What do the published counteroffer numbers actually say?
In published surveys, roughly half to two-thirds of countered employees accept the counter. Between about 40% and 74% of those who accept leave within a year, and most of that spread comes from who answered the survey. Use these figures as context for the read, never as a probability for the person in front of you.
| Source and market | Who answered | Accepted counter | Left within 12 months |
|---|---|---|---|
| Robert Half, UK | Employees | 53% | 74% of acceptors |
| Robert Half, Australia | Employers | 78% (derived) | 41% of acceptors (derived) |
| The Creative Group, US | Workers | 66% | Not reported |
| CEB via HBR, cross-industry | Undisclosed | Not reported | 50% |
Three things in that table matter for the read.
- The respondent type moves the number by about 30 points. Robert Half's employer-reported Australian figures and its employee-reported UK figures sit roughly 30 points apart on 12-month departure. Employers have a motive to believe their counters work. Write the respondent type next to any benchmark you quote.
- Counters are common. Hudson found 57% of professionals had received a counteroffer, but only 27% found it enough to stay. In Robert Half's Australian survey, 85% of employers said they had extended a counter in the previous 12 months. The Creative Group found 44% of workers were presented with one after giving notice.
- The best-sourced 12-month figure is 50%. CEB data, cited by HBR, says 50% of employees who accept a counteroffer leave within 12 months. Its methodology was never published. SHRM reports that roughly 57% change companies within 24 months, and credits that figure to LiveCareer.
How to score the six candidate-side dimensions
The candidate-side score measures whether a counter of typical size would be enough to keep this person. Score each dimension 0 (low risk), 1 (mixed), or 2 (high risk) from its named evidence source, so the side totals 0 to 12.
| Dimension | Evidence source | 0 (low risk) | 2 (high risk) |
|---|---|---|---|
| C1 Real reason for leaving | First screen, own words | Boss, scope, or challenge, with an example | Pay only, or vague |
| C2 Internal leverage | Direct question | Not raising pay internally | Raised pay this cycle, or used an offer before |
| C3 Comp pre-close | Notes before final round | Number stated and agreed | No number, or it keeps moving |
| C4 Offer uplift on stated base | Stated base vs your offer | 30% or more | Under 10% |
| C5 Counter-question answer | Pre-offer call | Specific, ties back to C1 | "Depends on the number" |
| C6 Resignation readiness | Rehearsal call, notice terms | Rehearsed; start inside notice plus 3 to 4 weeks | Not rehearsed; long or open start |
Anything between the two columns scores 1. If the evidence source has produced no answer, score 1 and mark it U for unknown.
C4 carries weight because of the gap between what employers offer and what candidates say they need:
| Source | Figure |
|---|---|
| Survey of hiring managers | 10% is the most common counter (36% of respondents) |
| ZipRecruiter, employers | Average counter is 10% |
| ZipRecruiter, job seekers | 33% average raise needed to stay |
| Hudson | 30% minimum for most; 32% would take no counter |
So the typical counter covers about 30% of the raise job seekers say they need (10 divided by 33). A cash-only counter usually fails unless pay is the candidate's only reason for leaving and your uplift is under about 10%.
What each candidate dimension proves, and when it lies
- C1 Real reason for leaving. This proves whether money can fix the problem. In Hudson's data the top reasons for leaving were a bad boss (35%), salary and benefits (32%), and needing more challenge (17%). A counter is mostly money, and money does not replace a manager. It lies when the first reason the candidate gives is the polite one, so ask twice and score the second answer.
- C2 Internal leverage. This proves whether the candidate may be using your process as a bargaining chip. Robert Half found 18% of people have used a job offer that way. It lies when the candidate asked for a raise a year ago and was refused. That employer has already said no, so score 1, not 2.
- C3 Comp pre-close. This proves the final number will not be a surprise. It lies when the number was agreed before the candidate saw the market, so reconfirm it at the final round.
- C4 Offer uplift on stated base. This proves how much room a counter has to close the gap. It lies when the stated base leaves out bonus or equity the candidate would walk away from, or when the base is a guess. If the candidate will not state a base, mark it U.
- C5 Counter-question answer. This proves the candidate has thought about the moment the counter arrives. It lies when the answer is fluent and enthusiastic but C2 says otherwise. Demeanor is the weakest evidence in the model, so never let C5 outvote C2.
- C6 Resignation readiness. This proves the resignation is a plan with a date, not just an intention. It lies when a long notice period is set by contract rather than chosen by the candidate. Score the start date you control, not the notice clause.
How to score the five employer-side dimensions
The employer-side score measures whether the current employer will counter and how hard. Score each dimension 0, 1, or 2, so the side totals 0 to 10. Anything you cannot back with evidence scores 1 and gets a U flag.
The motive comes from employers themselves. In survey data on why employers counter, 58% of managers said they did not want to lose institutional knowledge, 42% did not want to spend time or money replacing the person, and 39% of executives named keeping hard-to-find skills as the main reason. Smaller companies were more likely to counter than large ones.
| Dimension | Evidence source | 0 (low pressure) | 2 (high pressure) |
|---|---|---|---|
| E1 Pay gap vs posted range | Employer's posting, same title and level | Offer clears band top by over 10% | Offer sits inside their band |
| E2 Backfill difficulty | Pool count for the skill intersection | Deep, common pool | Narrow intersection or thin market |
| E3 Key-person exposure | Scope, tenure, named ownership | One of several peers | Sole owner or longest tenure |
| E4 Recent departure cluster | Team exits, 8-K Item 5.02 | No recent exits | Two or more exits, or a retention award |
| E5 Counter culture and calendar | Candidate reports, company size | No documented counters, large firm | Documented counters, smaller firm |
E1: Pay gap against the incumbent's posted range
Pull the employer's own posting for the same title and level, then compare your offer to the top of that range. If your offer clears the top by more than 10%, the employer has to break its own band to match, so score 0. If it lands within about 10% of the top, score 1. If it sits inside their band, they can match without an exception, so score 2.
A posted range is not the incumbent's pay. Glassdoor's analysis found posted ranges contain actual pay about 67% of the time. In 22% of listings actual pay was below the range and in 11% above it, and over 60% of salaries fall below the range median. Ranges in states with pay-transparency laws are about as accurate as ranges elsewhere. Treat the top of the band as an estimate of the employer's ceiling. When you need the candidate's likely current pay, use the lower half of the range and confirm it against their stated base. No one has published a validated method for estimating one named person's pay gap from their employer's range, so E1 is a bounded estimate, not a measurement.
E2: Backfill difficulty
Backfill difficulty is how hard it would be for the employer to replace the person from the open market. Count the pool for the candidate's title in their market, then count the pool for the skill intersection that makes them hard to replace. Score 2 when the intersection is under about a fifth of the title pool, or the market is several times thinner than a comparable one. Score 0 when the role draws on a deep, common pool.
| Segment (Refolk's index) | Count | Derived |
|---|---|---|
| SRE, United States | 6,632 | Baseline |
| SRE, United Kingdom | 1,435 | US pool is 4.6x the UK pool |
| SRE, US, Kubernetes + Go | 1,294 | 19.5% of US SREs |
A London SRE with Kubernetes and Go has to be replaced from a far thinner bench than a generalist SRE in the US, and the employer knows it. In a 25-profile sample, Google, Apple, and JPMorganChase lead the US SRE employers and Apple leads the UK sample. That sample is too small to rank employers, so treat it as directional only.
Counting the intersection is where most reads stall, because you have to size a specific skill mix in a specific market. In Refolk I describe the pool in plain English and get back the people who match, so E2 rests on a count instead of a guess.
E3: Key-person exposure
Key-person exposure is how much institutional knowledge would leave with the candidate. The evidence is their scope, their tenure compared with the rest of the team, and whether they are the named owner of a system in public repositories, docs, or talks. Score 2 for a sole owner or the longest-tenured person on the team, 1 for an owner who has a backup, and 0 for one of several interchangeable peers. This dimension lies when titles inflate scope, so confirm ownership with a concrete example in the screen.
E4: Recent departure cluster
A departure cluster is two or more recent exits from the candidate's team or leadership in the public record. Score 2 for a cluster or a disclosed retention award, 1 for a single exit, and 0 for none. If the employer is public, check SEC Form 8-K Item 5.02 filings, which disclose officer departures and related compensation. A Refolk search for people who left a named team at that employer in the last six months will show you any cluster.
Departures make counters more likely, not less. Myers Industries' CEO departed on the same day its remaining eligible executives received cash retention awards equal to 75% of base salary, payable on the first anniversary. Five Below valued each executive's special retention award at $1.5 million. Boards spend money to stop a second exit.
E5: Counter culture and the pay calendar
Counter culture is an employer's documented habit of countering. The evidence is the candidate's account of colleagues who were countered, any counters they have had themselves, and company size. Score 2 for documented counters at a smaller company and 0 for no documented counters at a large one.
This dimension has two limits:
- No study links refresh-grant or bonus calendars to counter likelihood. An upcoming bonus or refresh can add at most 1 to this dimension. CEB did find that job-hunting activity rises 6 to 9% after work anniversaries, but that is a timing signal about the candidate, not the employer.
- Who gets countered is unequal. Men receive counters more often than women (61% against 52% in Robert Half UK's data). Score the employer's documented behavior. Never score your guess about whether "people like this candidate" get countered, or the score will carry that bias.
How to turn the two scores into a verdict
Plot candidate risk against employer pressure. Low on both is Close Clean, a low candidate score with high employer pressure is Pre-Empt, a high candidate score with low employer pressure is Price In, and high on both is Run a Backup. A side counts as high at 5 or more (out of 12 for the candidate side, 10 for the employer side).
The four counteroffer verdicts
Apply these overrides after plotting:
- Pay-only reason and thin uplift. If C1 and C4 both score 2, the verdict is at least Price In, because a 10% counter will match you.
- Known leverage. If C2 scores 2, the verdict is at least Price In. If the employer side is also high, it is Run a Backup.
- Retention money on the record. If E4 found a disclosed retention award near the candidate's team, the employer side counts as high whatever its total.
- Too many unknowns. Two or more U flags on one side make that side high. You cannot call Close Clean on a read you could not back with evidence.
| Verdict | Trigger | Action | Timing |
|---|---|---|---|
| Close Clean | Both sides under 5 | Standard offer and close | Normal timeline; start inside notice plus 3 to 4 weeks |
| Pre-Empt | Candidate low, employer high | Rehearse resignation, hiring manager call, tight start | All before the offer goes out |
| Price In | Candidate high, employer low | Uplift above counter band; fix non-pay reason in writing | Number agreed before the final round |
| Run a Backup | Both high | Offer plus runner-up taken to offer-ready | Runner-up live until day one |
For Price In, set the uplift so that the candidate is still clearly ahead even after a 10% counter on their stated base, and write the scope or manager change that answers C1 into the offer. Some practitioners size sign-on pay against the pay the candidate gives up by leaving. Its value as a defense against counters is unverified, so use it because it is fair, not because it guarantees a close.
To get two recruiters to the same verdict, have them score independently, then compare dimension by dimension. A disagreement almost always sits on one dimension, and you settle it by checking that dimension's evidence source.
Finalist: Role: Scored by: Date: CANDIDATE SIDE (0 = low risk, 2 = high risk) C1 Real reason for leaving ........ [ ] Own words: C2 Internal leverage .............. [ ] Raised pay internally? When? Outcome? C3 Comp pre-close ................. [ ] Agreed number: C4 Offer uplift on stated base .... [ ] Stated base: ___ Uplift: ___% C5 Counter-question answer ........ [ ] Verbatim: C6 Resignation readiness .......... [ ] Notice: ___ Proposed start: ___ Candidate total (0 to 12): ___ High if 5+ U flags: ___ EMPLOYER SIDE (0 = low pressure, 2 = high pressure) E1 Pay gap vs posted range ........ [ ] Range: ___ to ___ Width over 60% of minimum? Y/N E2 Backfill difficulty ............ [ ] Title pool: ___ Intersection: ___ E3 Key-person exposure ............ [ ] Ownership evidence: E4 Recent departure cluster ....... [ ] Exits: ___ 8-K 5.02 retention award? Y/N E5 Counter culture and calendar ... [ ] Calendar adds at most 1 Employer total (0 to 10): ___ High if 5+ U flags: ___ Overrides: none / pay-only + thin uplift / known leverage / retention money / 2+ U flags VERDICT: Close Clean / Pre-Empt / Price In / Run a Backup Action: Deadline: Runner-up and owner:
Copy one per finalist; score 0, 1, or 2 per line, write U beside any unknown, and name the source.
The step-by-step read, from first screen to day one
The read runs in eight steps alongside your normal process, owned by the recruiter, sourcer, hiring manager, and TA ops. Steps 1 to 5 produce the verdict before the offer goes out, and steps 6 to 8 protect the window between acceptance and start.
The counteroffer risk read
- Find the real reason for leavingRecruiter, first screen, about 20 minutes. Ask why they are looking, then ask again, and write the reason in their own words.
- Pull the employer evidenceSourcer, 30 to 60 minutes. Collect the posted range for the same title and level, backfill or same-team postings, recent team exits, and 8-K Item 5.02 filings if public.
- Pre-close on comp and scopeRecruiter, before the final round. Get a stated number the candidate agrees to, so the final offer is not a surprise.
- Ask the counter question and rehearse the resignationRecruiter, 15 to 30 minutes, before the offer. Ask what they would do if their employer countered, then walk through the resignation itself.
- Score and record a verdictRecruiter plus hiring manager, 15 minutes. Score independently, settle differences by evidence, plot, apply overrides, and record the verdict with an action.
- Set the start dateHiring manager. Agree a start inside the notice period plus 3 to 4 weeks, in writing.
- Keep the runner-up warm until day oneRecruiter. Schedule a runner-up touchpoint after the finalist's resignation date and keep the slate open until they start.
- Log the outcomeTA ops. Record verdict, scores, acceptance-to-start days, and the withdrawal reason for every renege.
Sources disagree on the order of step 4. Ambassador Group says to rehearse the resignation before the offer goes out. Older guides from Capstone and Arlington put it after acceptance. I put it before. A candidate who has rehearsed the resignation has already decided what a counter would need to look like, and you want to know that while you can still change the offer.
Where counters land
- Verdict recordedOffer structure and runner-up plan set before the offer
- Offer acceptedStart date set inside notice plus 3 to 4 weeks
- Resignation handed inThe employer learns, and counters typically land here
- Notice periodRecruiter and hiring manager stay in contact; runner-up stays warm
- Day oneLog acceptance-to-start days and the outcome
1. "Before we put anything in writing: what would you do if your current employer made a counter-offer?" 2. "If they matched the number, what would still be true about the job that made you start looking?" 3. "Have you raised pay or title with your manager in the last year? What did they say?" 4. "Walk me through the resignation. Who do you tell first, when, and what will you say?" 5. "If they ask what it would take to keep you, what is your answer?" 6. "Your notice period is ___. If you start on ___, does anything stand between those two dates?"
Run it after comp is pre-closed and before the offer goes out; record answers verbatim against C2, C5, and C6.
How does the counteroffer risk read go wrong?
Most failed reads come from treating one weak signal as proof, whether that is demeanor, a posted range, an employer's own survey, or the 80% myth. Each failure below comes with a check you can run on the case in front of you.
- Candidate demeanor read as low risk. The false positive is the enthusiastic finalist who stays put. Robert Half found 18% of people have used a job offer as a bargaining chip, and they rarely look nervous doing it. Check: ask directly whether they have already raised pay internally, and let C2 outvote C5.
- Posted range read as incumbent pay. Only 67% of ranges contain actual pay, and pay skews below the midpoint. Check: estimate current pay from the lower half of the range, and confirm against the candidate's stated base.
- A very wide range scored as a large gap. A $50,000 to $180,000 range carries almost no information. Check: if the width is over 60% of the minimum, score E1 as unknown, not large.
- Employer-reported retention treated as fact. Australian employers report that 46% of countered staff were still there after 12 months, while UK employees report that 74% of acceptors left. Check: record which kind of respondent each benchmark comes from before you use it.
- A departure cluster scored as low counter risk. The reasoning goes "they're losing people, they can't afford a counter." Myers Industries shows the opposite: the CEO's exit triggered retention money for everyone who stayed. Check: search 8-K Item 5.02 filings and score a cluster as higher pressure.
- A long acceptance-to-start window treated as harmless. Counteroffers land in that window, usually once the resignation is handed in. Check: keep the start inside notice plus 3 to 4 weeks, and log the days for every renege.
- The 80% statistic used in a close. The candidate looks it up, finds no source, and stops trusting you. Check: use CEB's 50% and say its methodology is undisclosed.
- A counter-culture score that encodes gender. Men are countered more often than women (61% against 52%), so "people like this get countered" is a biased input. Check: score E5 only from the employer's documented behavior.
A departure on the candidate's team makes a counter more likely, because boards pay to stop the second exit.
How to keep the read calibrated after every close
To keep the read calibrated, log every outcome against the verdict you gave, so your thresholds come from your own closes rather than from published surveys. The published numbers differ by 30 points depending on who answered, and your own log does not have that problem.
For every offer, record:
- The verdict, both side totals, all 11 dimension scores, and any U flags.
- Days from acceptance to start.
- Whether a counter arrived, and its size if the candidate shares it.
- For every renege, the stated withdrawal reason, in the candidate's words.
When you have enough closes to compare, ask one question per verdict. Did any Close Clean case renege? If so, find the dimension that missed and adjust its scoring or the threshold of 5. Do not add new dimensions to cover one bad case. Did Run a Backup cases close at the same rate as Close Clean? If so, you are over-scoring, and the runner-up work is costing time you could save.
Re-pull time-sensitive inputs for each case instead of reusing an old sheet. Employers change their posted ranges, pay-transparency rules change, and pool counts move as people change jobs. When the runner-up becomes the hire, Refolk can recount the pool for the next search in the same plain-English query.
Before the offer goes out
- All 11 dimensions have a score and a named evidence source.
- Every unknown is marked U, and no side with two or more U flags was called low.
- E1 used the employer's own posting for the same title and level, and ranges wider than 60% of their minimum were scored U.
- E4 includes an 8-K Item 5.02 check if the employer is public.
- Two scorers reached the same verdict, or settled their disagreement by evidence source.
- The verdict, action, and deadline are written down before the offer is sent.
- The start date sits inside the notice period plus 3 to 4 weeks.
- The runner-up has a named owner and a touchpoint date after the finalist's resignation.
- No unsourced statistic appears in close notes or candidate conversations.
Questions practitioners ask
How do you predict if a candidate will accept a counteroffer?
Score both sides instead of reading demeanor. The candidate side asks whether a typical counter would be enough: real reason for leaving, internal leverage, pre-closed comp, uplift over base, their answer to the counter question, and resignation readiness. The employer side asks whether a counter will come: pay gap against its posted range, backfill difficulty, key-person exposure, recent departures, and counter culture. High on both sides means you make the offer and run a backup in parallel.
Do 80% of people who accept a counteroffer really leave within a year?
Nobody has traced that figure to a study. It is attributed to SHRM, the WSJ, and a "National Employment Association", and none of those trails holds up. The best-sourced figure is CEB's: 50% of counter-acceptors leave within 12 months, though its methodology was never published. Robert Half UK's employee survey found 74% of acceptors left within a year, while its Australian employer survey implies far fewer. Always quote the source and who answered.
How do I prevent a candidate from accepting a counteroffer?
Do the work before the offer, not after the counter. Pre-close the number early, ask what they would do if their employer countered, rehearse the resignation, set a start date inside notice plus three to four weeks, and keep the runner-up warm until day one. Practitioners document all of these steps, but no measured effect on counter-driven reneges has been published, so log your own outcomes and check what works.
Does an employer that just lost people counter less often?
No, the evidence points the other way. Departures raise the cost of losing one more person, so boards and managers spend money to stop a second exit. Myers Industries gave its remaining eligible executives cash retention awards worth 75% of base salary on the day its CEO departed. For public employers, check SEC Form 8-K Item 5.02 filings, and score a departure cluster as higher counter pressure, not lower.
What should I do when a candidate reneges after accepting because of a counteroffer?
Go back to the runner-up first, because reconnecting with a strong runner-up often closes faster than restarting the search. Then log the acceptance-to-start days, the verdict you gave, and the stated withdrawal reason. If the case was scored Close Clean, find the dimension that missed. Robert Half found 27% of people who backed out after accepting cited an acceptable counteroffer, so this is a common loss you can track.
How big is a typical counteroffer?
About 10%. In one survey of hiring managers a 10% raise was the most common counter, cited by 36% of respondents, and ZipRecruiter found the average counter was also 10%. Job seekers told ZipRecruiter they would need about 33% to stay, and Hudson found most professionals saw 30% as the minimum acceptable counter. A typical counter therefore covers about 30% of what candidates say it would take.
Try it on the search you came here for
Stop building boolean strings. Just describe the person.
Type one sentence. I plan the search, read GitHub, public LinkedIn and Crunchbase records, and the open web as it is right now, and hand back a ranked list with the reason next to every name.
01Describe them
One plain sentence. Role, city, stack, stage, whatever matters to you.
02I read the web live
GitHub, public LinkedIn and Crunchbase records, the open web. Not a database that went stale last quarter.
03You read the shortlist
Ranked, with the reasoning under every name. Open a profile, ask a follow-up, narrow it down.
- Staff backend engineers in NYC who shipped Rust in production
- Series A fintechs in SF under 50 people, growing headcount this year
- Maintainers of fast-growing Rust web frameworks on GitHub
- No boolean, no filters, no seat to buy. One box.
- Read at search time, so a profile updated yesterday counts today.
- Every step visible as it runs, every name with its reason.
500 free credits on sign-up. No card, no demo call. See real searches.