How to Design a Leave Policy That Actually Gets Used

Most Policies Don't Fail on Generosity

Most PTO policies do not fail because they are stingy. They fail because they are ambiguous, manager-discretion-heavy, and quietly stigmatize the act of using them. A team can have 25 days on paper and still average 11. The number in the offer letter is not the policy. The behaviour around it is.

The single design choice that determines whether a leave policy works is whether the default is "save" or "use." Everything else -- carry-over rules, floating holidays, shutdown weeks -- is downstream of that choice. Get it wrong and a generous allocation just sits in a spreadsheet, accruing as a balance sheet liability. Get it right and a smaller allocation outperforms a larger one elsewhere.

This piece is for HR leaders, founders, and managers writing or overhauling a policy in 2026. The framing is operational, not theoretical: what design decisions actually move utilization, and what tooling and norms are required to make them stick.

Default to Use, Not to Save

The cleanest evidence sits in the unlimited-PTO experiment. The pitch was that removing the cap would liberate employees to take more time off. The result has been mixed at best. Industry data has produced different numbers depending on the source -- Namely's comparison found unlimited-PTO employees took 12.09 days versus 11.36 on traditional plans, and SHRM's reporting put the figures at roughly 16 versus 14. Both directions tell the same underlying story: removing the number does not, on its own, increase usage. It often reduces it. Without a target, employees anchor to whatever feels socially safe -- and "safe" is almost always lower than they would land with a posted allocation and an explicit nudge.

Netflix is the case study most people cite. The company has run unlimited PTO since 2003 and is on the record that the policy only works because of active leadership modelling -- executives publicly take time, talk about it on return, and reset the team's frame of what is normal. Reed Hastings has been blunt that without that modelling the policy backfires and people take less. The policy is not the cause of the outcome. The behaviour around the policy is.

Microsoft moved its US salaried employees to unlimited "Discretionary Time Off" in early 2023. The policy carries no mandatory minimum. It has the same dependency Netflix flagged: it works to the extent that leadership makes use visible.

Buffer is the most useful counter-example because they ran unlimited PTO and then added structure on top of it. They publish a transparent time-off dashboard showing how much vacation each teammate has taken in the last 12 months, and they ask everyone to take a minimum of 15 days off per year on top of public holidays. The minimum is not a cap, it is a floor -- they explicitly framed it as removing the guilt of taking time off when there is no defined entitlement to point to.

The pattern across all three: an unlimited or generous allocation is necessary but never sufficient. What converts the entitlement into actual usage is a combination of:

• A visible floor (a target number of days, a posted minimum, or a transparent log of peer usage) that gives employees something concrete to anchor to.
• Manager and executive modelling, where leadership takes long blocks and talks about them. If the head of engineering takes a week and is fully offline, the team learns it is allowed. If they take three days and answer Slack on two of them, the team learns the opposite.
• Operational coverage that does not require the person on leave to handle anything. The single biggest predictor of low utilization is the unspoken expectation that the work will pile up while you are away. If your absence creates a backlog you return to in full, you are not really off.

Policies that get used solve all three at once. Policies that solve only the first one -- the document -- get the 60% utilization rate that is depressingly common across companies offering 20+ days of annual leave.

What That Implies Operationally

Once "default to use" is the design intent, the rest of the policy follows.

Carry-over rules should permit a modest rollover -- five to ten days -- with a Q1 or Q2 deadline. Use-it-or-lose-it on December 31 produces a Q4 stampede, which is the worst possible time of year for it. Unlimited rollover produces hoarding and a balance sheet liability that hits at termination. The narrow middle works best, paired with automated balance reminders at the 6-month and 9-month marks so people are not surprised in December.

Company-wide shutdowns -- the week between Christmas and New Year, a summer Friday cluster, a quarterly mental health day -- function as bonus rest because they do not draw from individual allocations and they remove the coverage problem entirely. There is no one to backfill, no Slack queue accumulating, no guilt calculus. Everyone is off at once. This is the single highest-leverage addition to most policies because it eliminates the friction at the team level rather than the individual level.

Floating holidays handle the cultural-calendar problem that standard public holidays create. An employee observing Diwali, Eid, Lunar New Year, or Chuseok should not have to spend personal leave on days that matter deeply to them while Christmas and Easter are handed out automatically. Two to three floating days a year, assignable to any date, fixes this with negligible administrative cost. It also flattens demand around the standard public-holiday clusters.

Approval friction is the underrated structural problem. A policy that takes longer to get approved than the leave itself trains people not to take short breaks. The ceiling should be a direct manager confirming within 48 hours, not a multi-step review. Auto-approving requests submitted 30+ days in advance solves the same problem from the other direction and rewards the foresight that makes coverage planning easier for the manager too.

Two specific practices worth flagging because they undermine policies even when the policy itself is sound:

Blackout periods that overlap with every productive bridge window -- where strategic days off bracket public holidays to create longer breaks -- effectively reduce the usable allocation to a number much smaller than the headline one. If your blackout calendar blocks every Thursday-to-Monday around a holiday, employees learn that "20 days" is a theoretical figure they can never actually deploy.

Counting public holidays inside the leave allocation -- advertising "25 days" when 8 are bank holidays -- is the kind of bait-and-switch that corrodes trust the moment someone does the arithmetic. State the allocation as days beyond statutory holidays and beyond any company-wide shutdown.

Pro-rata rules for part-time, compressed-hours, and contract workers should be published explicitly. Burying them in a footnote produces confusion, which produces underuse.

Build a Bridge-Friendly Policy

Strategic leave planning -- bridging public holidays with one or two PTO days to create longer breaks -- is one of the cheapest ways to raise the perceived value of an allocation without adding days. It only works if the policy supports it.

In practice that means publishing recommended bridge windows at the start of each year, naming the dates where one or two leave days connect a public holiday to a weekend. When the company surfaces the opportunity, employees treat taking it as expected rather than opportunistic. A quarterly five-minute leave-planning slot in team standups normalises strategic time off as a shared planning activity rather than something individuals quietly arrange. Sharing planning tools at onboarding -- like LeaveWise for country-specific bridge windows -- sets the tone from day one.

The ROI Case

The business argument for high-utilization leave is cleaner than people assume. Three components, in order of how easy they are to defend:

Retention. PTO satisfaction consistently ranks among the top factors in the standard employee-survey data on stay-or-leave decisions. Replacing a mid-career employee runs roughly 50% to 200% of their annual salary. A few extra days of taken leave is much cheaper than one rehire.

Productivity. The research on rested-versus-fatigued performance is unambiguous. People who take regular planned breaks produce higher-quality work, make fewer errors, and sustain output across a 12-month cycle. The team that sprints for 10 months and stumbles through the last two does not net out ahead of the team that paces itself.

Employer branding. As compensation packages converge, the texture of the leave policy -- how it actually feels to use, not how it reads -- becomes the differentiator. Candidates ask current employees whether people really take their days. The answer is the brand.

Start with the Calendar

The best leave policies don't just grant time -- they guide people toward using it well. That starts with making the calendar visible, the bridges obvious, and the act of planning rest as normal as planning work.

Help your team find the best windows →