A retention bonus — also called a stay bonus or stay-pay — is a lump sum paid to an employee on the condition that they remain with the company through a specific date. It is not a performance bonus, and it is not recognition. It is a contract to buy time.
That framing matters, because retention bonuses work well for exactly the thing they're built for and poorly for everything else. This guide covers the structures that hold up, sample terms, the tax and clawback details people get wrong, and the honest case for when a retention bonus is the wrong tool entirely.
A quick disambiguation, since search sends these to the same place: this post is about retention bonuses paid to employees. It has nothing to do with the Employee Retention Credit (ERC), which is a US payroll tax credit from pandemic-era legislation. Different thing entirely — if that's what you came for, you want a tax advisor, not us.
When a Retention Bonus Is Genuinely the Right Tool
There's a narrow set of situations where stay-pay is not just defensible but clearly correct. All of them share one feature: a known end date, and a cost of early departure that dwarfs the bonus.
- Acquisition or merger. The deal requires specific people through close and integration. Their leaving mid-process doesn't cost you a replacement — it costs you the deal terms.
- Announced wind-down or site closure. You've told people their roles end in nine months. Absent stay-pay, the best people leave first, and the ones who keep the lights on are the ones with the fewest options. Retention bonuses here are as much an ethics measure as a business one.
- A critical migration or launch with a hard date. The person holding the institutional knowledge for a system nobody else fully understands, six months from cutover.
- Post-acquisition founder or key-person lock-in. Standard practice, usually structured as equity rather than cash.
Notice what all four have in common: the risk is time-bounded and identifiable. You know who, you know until when, and you can price the alternative.
When It Backfires
The failure mode is using stay-pay to solve ordinary attrition — someone is unhappy, you sense they're looking, you offer money to make them stay.
This almost never works, for a reason worth stating plainly: a retention bonus buys a notice period, not commitment. Someone who was leaving for reasons of manager, growth, or workload is still leaving for those reasons; you've simply scheduled it. The industry rule of thumb — that a large share of employees who accept a counteroffer leave within a year anyway — exists because the underlying cause was never compensation.
Worse, it's contagious in a way that's difficult to contain. Retention bonuses do not stay secret. Once one person's payout is known, you have taught the organization that the reliable route to a raise is to credibly threaten to quit. That's an expensive lesson, and the people who learn it fastest are rarely the ones you most wanted to keep.
If you're offering a retention bonus because someone might leave, you're probably too late. If you're offering one because you know exactly what date you need them through, you're using it correctly.
Structuring One That Holds
Amount
Typical ranges, for planning:
| Situation | Typical bonus |
|---|---|
| Standard stay-through-date, individual contributor | 10–15% of base salary |
| Critical role, M&A or migration | 15–25% of base salary |
| Executive / key person, deal-contingent | 25–100%+, frequently in equity |
| Site closure / wind-down | Often expressed as weeks of pay, stacked on severance |
The sizing question isn't "what will they accept" — it's what does their early departure cost? If a migration slips two quarters because the one person who understands the legacy schema left, price against that, not against a percentage convention.
Timing and structure
Single lump sum at the end is the simplest and the most common. Two structures work better in longer engagements:
- Split payment — e.g. 40% at the midpoint, 60% at the end date. Reduces the "I'll just leave now, the payout is a year away" problem.
- Tranches with milestones — payments tied to both a date and a deliverable. Better alignment; more administrative overhead and more room for dispute about whether a milestone was met. Only worth it for large amounts.
The agreement itself
A retention agreement should be short and unambiguous. At minimum it needs:
- The retention date, stated exactly.
- The amount and payment timing, gross, with the tax treatment noted.
- What counts as qualifying employment — active, in good standing, not under a performance plan.
- Termination handling. This is the clause people miss: if you terminate them without cause before the date, the bonus should still pay out. Otherwise you've written an agreement that lets you take the retention benefit and then avoid the cost, and word of that gets around fast.
- Clawback terms, if any — see below.
- An explicit statement that it doesn't alter at-will status or constitute an employment-term guarantee.
Have counsel review it. Retention agreements are enforceable contracts and the clauses that matter are the edge cases.
Tax treatment
In the US, retention bonuses are supplemental wages: taxable income, subject to withholding, commonly at the 22% federal supplemental rate (37% above $1M), plus FICA and state. The practical consequence is that a "$10,000 retention bonus" lands as roughly $6,500–7,000 in someone's account, and if you haven't said so, the disappointment eats a meaningful chunk of the goodwill you just paid for. Communicate the gross and the approximate net.
Clawbacks
Clawbacks — repayment if the employee leaves within some window after the payout — are common in signing bonuses and much rarer in retention bonuses, for good reason. You paid for a date; they hit the date. Extending the obligation past the payout tends to read as a bait-and-switch and can be difficult to enforce depending on jurisdiction. If you need commitment past the retention date, that's a second agreement, priced separately.
The Cost Comparison Nobody Runs
Before committing to stay-pay, it's worth pricing the alternative properly, because the two numbers are usually closer than people assume.
Replacing an employee runs roughly 33% of salary at the low end (Work Institute), 50–60% as a typical figure (SHRM), and up to 150%+ for specialized or senior roles (Gallup). On a $120,000 engineer, that's $40,000 to $180,000 in recruiting, ramp time, lost productivity, and the drag on everyone who covers the gap. A 15% retention bonus on that salary is $18,000. If the bonus genuinely secures a needed nine months, the math is not close.
But run the same comparison for ordinary attrition and it inverts, because the bonus doesn't prevent the departure — it defers it. You pay $18,000 and the $40,000–180,000, just in a different quarter. Our turnover cost calculator will do this arithmetic on your actual headcount and salary bands.
What Actually Retains People
The uncomfortable part. Retention bonuses are a tool for known, dated, specific risks. They are not a retention strategy, and organizations that reach for them repeatedly usually have a problem further upstream.
The research on why people leave is consistent and mostly not about money. The Work Institute finds year after year that roughly three in four voluntary departures were preventable — driven by career growth, manager relationship, and feeling valued, all of which the employer controls. Gallup attributes roughly 70% of the variance in team engagement to the manager. And employees who don't feel adequately recognized are about twice as likely to say they'll quit within a year (Gallup / Workhuman), while organizations with strong recognition cultures report up to 31% lower voluntary turnover (Deloitte).
None of those are fixed by a lump sum. They're fixed by things that are cheaper and slower: managers who notice work, growth paths that are real, and a culture where contribution gets named in public rather than absorbed silently.
Our bias, stated plainly: we make Propsly, a Slack peer recognition tool, so we have an obvious interest in the "feeling valued" line item. What we'd argue anyway is the sequencing: stay-pay is emergency medicine, and if you're prescribing it routinely, the thing to examine is why. If you want that argument without the product attached, the retention math and why recognition programs fail both make it.
A Decision Checklist
Before you write the agreement, four questions:
- Is there a specific date? If you can't name it, this isn't a retention bonus — it's a raise with extra steps, and you should give a raise instead.
- Have you priced early departure? If the bonus exceeds the cost of them leaving, don't do it.
- Do you know why they'd leave? If the reason is money, stay-pay may work. If it's manager, growth, or burnout, it will buy a delay at full price.
- Can you defend it publicly? Assume the amount becomes known. If you can't explain to the rest of the team why this person and not them, you have a fairness problem that will cost more than the bonus.
The Short Version
Retention bonuses are a precision instrument for a specific job: getting a named person through a named date when their early exit would cost you far more than the payment. Structured properly — clear date, clear amount, tax communicated, termination-without-cause still paying out — they're clean and effective.
Used as a response to someone handing in notice, they're an expensive way to reschedule a resignation and teach everyone watching how to get a raise. The organizations that rarely need them aren't the ones with bigger budgets. They're the ones where people didn't get to the point of leaving.