Your 2027 recognition budget doesn't need a better benchmark

Every recognition benchmark you've read was published by someone selling recognition 

Go looking for a 2027 recognition budget number and you'll find three:

• $100 to $150 per employee per year for performance-based recognition.
• $200 to $350 per employee per year for a full recognition culture, service awards and milestones included.
• 1% to 2% of total payroll, the most-cited figure of the three, usually attributed to SHRM but in practice sourced from vendor blog posts citing other vendor blog posts.

Every one of those reaches you through a company that sells recognition programs. Ours included. That isn't a scandal, it's how industry research works. Vendors have the data because vendors have the customers.

The problem is what happens when you put one of those numbers into a document and hand it to a CFO. 

The stat you thought was independent

The most-repeated recognition ROI figure in the industry says a 10,000-person company can save $16.1 million a year in turnover costs. It usually travels as Gallup research, which lands very differently than vendor research. Gallup is close to gospel in HR, and mostly deserves to be.

That study was co-published. Gallup ran it jointly with a recognition software company, which makes it the vendor's research as much as Gallup's.

There's also a condition attached that almost never survives the retelling. The $16.1 million applies to an organization of 10,000 people with an already engaged workforce. That qualifier does an enormous amount of work. It means the finding isn't "recognition saves you $16.1M." It's closer to "among companies already doing engagement well, recognition is worth this much on top." Strip the condition out and you have a much bigger claim and a much less true one.

Now picture your CFO's analyst spending ten minutes with it. They find the co-author. They find the dropped qualifier. They find the leap from a 10,000-person org to your 900-person one.

You don't lose the recognition budget in that moment. You lose something worse: the assumption that HR's numbers can be taken at face value. Next quarter's ask starts in a hole. 

Compensation already learned the lesson recognition hasn't 

Here's what makes 2027 different, and it has nothing to do with recognition.

WTW's Salary Budget Planning Report, based on 1,650 US organizations surveyed in spring 2026, puts average 2027 salary increase budgets at 3.4%. That's a shade under 2026's actual 3.5%. Flat, in other words. Gallagher's survey of roughly 1,200 employers lands in the same territory at 3.3% to 3.4%. WorldatWork has it at 3.6%. Pick whichever source you trust most; the answer doesn't change. There is no more money.

WTW's Lori Wisper described the outlook as the "land of 3%" for the foreseeable future.

The percentage isn't the interesting finding. What employers are doing with a flat budget is. WTW found companies moving away from broad-based increases toward targeted, performance-driven allocation: higher starting salaries for specific roles, retention bonuses for specific people, spot awards. About a third of employers had already adjusted their compensation programs, with more planning to.

Read that again with recognition in mind.

Compensation has accepted that a flat budget requires sharper allocation. The dollars aren't growing, so the discipline has to move to where they land. That is now the mainstream position in comp planning.

Recognition budgets, at the same companies, in the same planning cycle, are still set the way they've been set for a decade. A per-employee figure, multiplied by headcount, spread evenly across departments. Peanut butter. 

That's the actual 2027 story. Not that recognition budgets are too small. Recognition is still running the allocation model compensation just walked away from. 

What your CFO is actually weighing 

Recognition business cases usually open with program metrics. Participation rate. Recognitions sent per employee per month. Percentage of managers active on the platform. Sentiment scores.

None of those are decision inputs for a CFO. They're operating metrics for you.

Finance is weighing something narrower and more boring. Our longer guide to making the business case for employee recognition works through the whole conversation, but it comes down to this:

• Turnover cost avoided, fully loaded. Recruiting spend, time to productivity for the replacement, and the drag on the team during the gap.
• Whether the number is traceable. Can they follow it back to a source inside the company?
• What the money isn't doing instead. Every discretionary line competes with every other discretionary line.

A benchmark answers none of that. An industry average per-employee figure tells your CFO what other companies spend. It doesn't tell them what your company loses. Those are different questions, and only one of them is being asked in the room. 

There's a benefits squeeze running underneath all of this too. Per-employee health benefit costs are projected to rise sharply in 2026, the steepest increase since 2010 according to SHRM's benefits research. That pressure has to come out of the total rewards envelope somewhere, and discretionary programs are the softest target on the page. Recognition needs a defense that survives the comparison. 

Build a number your CFO can trace

Here's the method. It takes an afternoon and produces a figure with no vendor fingerprints on it. 

1. Work out your fully loaded cost per departure

Pull your last 12 voluntary exits. For each one, add up:

• External recruiting cost: agency fees, job board spend, or an allocated share of internal recruiter time.
• Hiring manager and interviewer hours, at loaded rates.
• Onboarding and training cost.
• The productivity gap between the departure and the replacement reaching full contribution. 

Average it. You now have a cost-per-departure number specific to your company. It will be rougher than you'd like. That's fine, and you should say so in the document. A rough number your CFO can trace beats a precise number they can't.

For most mid-market organizations this lands somewhere between half and two times annual salary depending on the role. Don't take that range from me. Calculate yours.

If you'd rather work against a structure than a blank spreadsheet, our turnover cost calculator asks for the same inputs. Use your own figures in it. The output is only as traceable as what you put in. 

2. Name the population you're actually protecting

Not headcount. Not "all employees." The specific group where turnover costs you the most.

Usually that's some mix of roles with the longest time to productivity, roles where a departure leaves a client-facing gap, and teams already showing elevated attrition. In most mid-market companies that's a few hundred people, not the whole org.

This is the step most business cases skip, and it's the one that puts recognition on the same targeted-allocation footing comp is already using. You're not proposing to spread money evenly. You're proposing to protect a specific population. 

3. Set a retention delta you'd be willing to defend

Honesty helps you here.

Don't claim recognition will halve attrition. Claim something small, defensible, and easy to check. A one to three percentage point reduction in voluntary turnover inside your target population is a conservative ask most CFOs will accept as plausible, and it's a figure you can be held to.

Then multiply: target population × retention delta × fully loaded cost per departure = turnover cost avoided.

Compare that to the program cost. If the ratio isn't compelling at a conservative delta, your problem isn't the business case. It's the program design, and you want to know that before the meeting rather than after. 

4. Put your measurement plan in writing before anyone asks for it

Say how you'll know whether it worked and when you'll report back. Voluntary turnover in the target population, measured at 6 and 12 months, against the trailing 12-month baseline.

That's the highest-leverage sentence in the whole document. It turns a spending request into a testable proposal, and it signals that you're not afraid of the result.

Gathering a clean baseline before launch is the part teams most often skip, and it's the part that makes the 12-month number arguable later. We walk through it in more detail, including which metrics to capture and when, in our piece on recognition ROI in manufacturing. The sector is specific; the baseline discipline isn't. 

Where a benchmark still belongs, and where it doesn't

Benchmarks aren't useless. Saying so would be overcorrecting, and you'd be right to distrust it.

They're useful as a private sanity check. If your self-built number comes out at $18 per employee per year, the industry range tells you the program is probably underscoped. If it comes out at $600, probably overscoped. That's real information.

The rule is about sequencing, not validity. Use benchmarks to check your math. Never cite them as the justification.

The moment an industry average shows up in the CFO-facing document as the reason for the number, you've imported the credibility problem you were trying to avoid. Do the sanity check in your own drafts and keep it out of the deck.

Our HR calculator hub is built for exactly that kind of private math. Budget, turnover, disengagement, absenteeism, and administrative time, all of it driven by your inputs rather than an industry average. 

The question a spend figure can't answer

There's one thing no benchmark can tell you, and it's the thing that most often explains why a recognition budget produces nothing: whether your program is built to convert money into retention at all.

Our Recognition Maturity Model scores programs across the dimensions that determine whether recognition changes behavior. Manager enablement, frequency, visibility, alignment to values, and measurement. It's scored out of 32. The average score across participating organizations is 18.4, the Developing band, at an average organization size of about 2,400 employees. Scores range from 8 to 32 among companies spending comparable amounts per head.

That range is the finding. Two mid-market companies can spend identical dollars per employee and land eleven points apart, because the money is doing completely different work in each. One is funding a service award catalog nobody visits. The other is funding manager-driven recognition that lands weekly.

We're a vendor and this is our data, so apply the same discount you'd apply to anything else in this piece. But notice that it answers a different question than the benchmarks do. A benchmark tells you what to spend. A maturity score tells you whether spending will do anything. If your program sits in the Developing band, more budget is the second problem. The first is that the delivery mechanism leaks. 

What proof looks like at your scale

Most recognition case studies you'll run into are enterprise. Ten thousand employees, a Fortune 500 logo, numbers that don't transfer to a 900-person company with a four-person HR team.

Here's one closer to your size. ATCC, a 620-employee organization, reached 83% platform registration and saw a 10% improvement in retention within a year of launching their program.

The number isn't the useful part. The shape is. A specific population of 620 people at one organization. A specific starting point, a specific ending point, and a one-year window. Checkable, bounded, and small enough to believe.

That's the structure your own business case should copy. Not "recognition improves retention," which is an abstraction your CFO has heard and discounted. Instead: our voluntary turnover in this population was X, it is now Y, over Z months, and here is the fully loaded cost of what we didn't lose.

The 83% registration figure deserves one more note, because it's exactly the kind of metric I said three sections ago that CFOs don't weigh. They don't. But you should, because it's the leading indicator. High registration doesn't prove ROI. Low registration predicts its absence. Keep it in your dashboard and out of your business case.

The one-page version

If you take one thing into your 2027 planning cycle, take this sequence:

1. Fully loaded cost per departure, calculated from your last 12 exits.
2. The specific population you're protecting, not total headcount.
3. A conservative retention delta you'd be willing to be measured against.
4. The multiplication.
5. A measurement plan with a date on it.

That's the whole document. One page. No industry averages, no vendor logos, no $16.1 million.

Salary budgets are flat at 3.4% and the discipline in compensation has already moved from how much to where it lands. Recognition is running about a year behind. The teams that close the gap this planning cycle will do it by bringing finance the one thing finance has been asking for all along: their own numbers.

Stop looking for a better recognition benchmark. Build a number your CFO can't trace back to a vendor. Including us.