There's a specific moment where most recognition pitches die: the CFO asks "how do you know?" and the answer is a vibe. Somebody cites a giant headline number — turnover costs two times salary! recognition cuts attrition by a third! — the CFO mentally files it next to "our TAM is $40 billion," and the budget request quietly evaporates.
That's a shame, because the recognition vs. turnover case doesn't need the giant numbers. The retention math works fine — better, actually — when you build it the way a CFO would: low-end cost assumptions, discounted effect sizes, full program costs counted, and a break-even so small it's almost embarrassing. This post builds that model step by step. Steal all of it.
Why CFOs Reject Most Retention Math
Finance leaders don't distrust HR; they distrust asymmetric assumptions. The typical recognition business case takes the highest available replacement-cost estimate, multiplies it by the highest available effect size, and attributes 100% of the result to the program. Every assumption breaks the same direction. Any CFO who's survived a vendor pitch can smell it.
The fix is to invert the posture. Take the lowest defensible number at every step, show your sources, and let the CFO argue the number up. A conservative model that still clears the bar is far more persuasive than an optimistic one that clears it by a mile — because the conservative one survives the meeting after you leave the room.
Step 1: Price a Departure Conservatively
Start with what one voluntary departure actually costs. The honest answer is a range, and the range is wide:
- Work Institute (the floor): replacing an employee costs roughly 33% of their salary — a deliberately low estimate covering recruiting, backfill, and lost productivity.
- SHRM (the middle): 50–60% of salary.
- Gallup (the ceiling): one-half to two times salary, once you count ramp time, institutional knowledge, and the drag on the team left behind.
For a CFO-grade model, use the floor. At a $65,000 average salary, one departure costs about $21,450 at the Work Institute rate. Nobody in finance will fight you on a number that conservative — which is exactly the point. (For the mid-range version most companies use for planning, the site-wide worked example is 100 employees × $65,000 × 15% turnover × 50% replacement cost = $487,500 a year. We'll carry both.)
Better still, don't use our salaries at all. Our employee turnover cost calculator lets you plug in your own headcount, average salary, and turnover rate, and slide the replacement-cost assumption from the 33% floor to the 200% ceiling. A model built on your numbers is the one your CFO will actually engage with — and for a deeper tour of what hides inside that replacement percentage, see what employee turnover really costs.
Step 2: Size the Baseline Exposure
Multiply out the annual exposure for a 100-person company at $65,000 average salary and 15% voluntary turnover (15 departures a year):
- Conservative (33% replacement cost): 15 × $21,450 ≈ $321,750 per year
- Mid-range (50% replacement cost): $487,500 per year
One more input matters, because it answers the "isn't turnover just inevitable?" objection: the Work Institute's exit research finds about 3 in 4 voluntary departures are preventable — people leave over things employers control, like development, management, and feeling valued. So of that $321,750 conservative exposure, roughly $241,000 is attached to departures the company could plausibly have kept. Turnover isn't weather. It's a controllable cost with a controllable share.
Step 3: Apply the Recognition Effect — Then Cut It
Now the treatment effect. Two anchor findings, both from large research bodies rather than vendor surveys:
- Gallup/Workhuman: employees who don't feel adequately recognized are about 2x as likely to say they'll quit within a year.
- Deloitte (Bersin): companies with strong recognition cultures see up to 31% lower voluntary turnover.
Here's where most business cases lose the room: they plug in 31% and call it savings. Don't. That figure describes companies with mature, strong recognition cultures — not year one of a new program. A CFO-grade model haircuts the headline effect for imperfect execution, ramp time, and attribution uncertainty. Cut it by two-thirds and model a 10% reduction in voluntary turnover: 1.5 fewer departures a year out of 15.
Even at the floor-level replacement cost, that's 1.5 × $21,450 ≈ $32,000 a year in avoided cost. At the mid-range 50% replacement cost, it's about $49,000. And remember what we did to get here: lowest replacement estimate, effect size slashed by two-thirds, no credit for engagement, productivity, or referral effects (the hidden costs of disengagement that never show up in a turnover line). This is the version of the number designed to survive hostile review.
Step 4: Count the Full Cost — Then Compare
A model that hides program costs isn't conservative, so count everything: software, rewards budget if you fund one, and the admin time to run it. Recognition software for a 100-person team typically runs anywhere from a few hundred dollars a year (flat-priced tools) to $3,000–$5,000+ a year (per-seat tools at $2–$4 per user per month), plus whatever you choose to spend on rewards.
Now the comparison, using the most pessimistic column of our own model:
- Conservative annual savings: ~$32,000
- Generous annual cost estimate: ~$5,000
- Net: ~$27,000 a year, roughly a 6x return — with every assumption leaning against you
The break-even framing is even sharper, and it's the sentence to lead with in the meeting: at floor-level replacement costs, the program pays for itself if it prevents one quarter of one departure per year. The question isn't whether recognition can clear that bar. It's whether anything else in the retention budget clears it as easily. (We run the fuller version of this analysis, with sensitivity ranges, in the ROI of employee recognition.)
The Objections a Good CFO Will Raise
"Correlation isn't causation."
Correct, and concede it immediately. The Deloitte and Gallup findings are observational: companies with strong recognition also tend to have other good habits. That's exactly why the model haircuts the effect by two-thirds instead of claiming the headline number. Then offer the real answer: run it as an experiment. Recognition is cheap enough to pilot for two quarters, and it generates its own measurement data — who's giving, who's receiving, which teams are engaged — so you can compare recognition activity against actual departures in your own workforce rather than arguing about someone else's study.
"How would we even measure it?"
Pick the metrics before launch: voluntary turnover rate (trailing 12 months), regrettable departures specifically, and program participation as the leading indicator. A recognition program nobody uses saves nothing, so participation is the honest early signal that the savings estimate is or isn't on track.
"Won't people just leave for money anyway?"
Some will, and no props emoji stops a 30% raise. But the Work Institute's preventable-departure data says compensation is the minority story; most people leave over how work feels — unseen, unvalued, going nowhere. Recognition attacks precisely that category, which is why the 3-in-4 preventable figure belongs in your model.
Where Propsly Fits (Yes, Propsly Is Ours)
Full disclosure: we sell the thing we're about to mention. But it's relevant to the math, because program cost is the denominator. Propsly is Slack-native peer recognition — teammates give each other props with a /props command, everyone gets 200 props a month to hand out, and every give lands in a public feed. The free tier covers unlimited users with leaderboards included, and Pro is $50/month flat for the whole workspace — $600 a year — for advanced analytics and automated gift-card rewards. Run that through the break-even above: at a $21,450 floor-level departure cost, Pro pays for itself if it prevents one departure every 35 years. That is not a typo; that's just what happens when the denominator is small.
Comparison shopping is rational — our recognition tools guide covers the whole market, including tools that aren't ours. The retention math holds for any of them; flat pricing just makes the CFO conversation shorter.
The One-Slide Version
If you take one thing into the budget meeting, take this structure:
- Exposure: departures/year × salary × 33% (Work Institute floor) — run your own numbers in the turnover calculator.
- Preventable share: ~3 in 4 departures (Work Institute).
- Effect: up to 31% lower turnover in strong recognition cultures (Deloitte) — modeled at 10% after a two-thirds haircut.
- Cost: full program cost, software plus rewards plus admin time.
- Break-even: a fraction of one prevented departure per year.
Every number sourced, every assumption leaning conservative, and a return that still clears the bar with room to spare. That's the retention math CFOs actually believe — because it's the math they'd have built themselves.