Best overtime policy for small business: cap it, approve it, or hire, and the number that decides
The best overtime policy for a small business is written pre-approval with a published weekly threshold, reviewed quarterly against a crossover number you calculate once. Overtime is a variable cost and headcount is a fixed one, so overtime genuinely is the cheaper answer until the recurring shortfall gets close to a full schedule. For a $22 an hour role that crossover sits near 32 overtime hours a week across the team. Below it, approve the hours. Above it, you are financing a job you have not posted.
Most overtime policies are written as a rule about behavior, which is why they fail. They say overtime requires manager approval, the manager approves it because the work is real, and nobody ever revisits the arrangement. Two years later a company of 30 is paying for the equivalent of a person and a half in premium hours and has no idea, because the cost is spread across every paycheck instead of sitting on a headcount line.
The fix is not a stricter rule. It is a number that tells you when the rule should change.
What an overtime hour actually costs you
Start with the true marginal cost, because the sticker price understates it. An overtime hour is 1.5 times the regular rate, and the regular rate is total weekly earnings divided by hours worked, not base pay. Employer payroll taxes then apply to the premium exactly as they apply to everything else: 6.2 percent Social Security up to the wage base, 1.45 percent Medicare, plus federal and state unemployment insurance and workers compensation premiums that are usually rated on payroll.
Call the payroll tax load 7.65 percent for FICA alone and the arithmetic is simple. At a $22.00 regular rate, an overtime hour bills at $33.00 and costs you about $35.52 loaded, which is 1.61 times base pay. Workers compensation and unemployment push it higher in most states. If your regular rate is inflated by a production bonus or a night differential, as it should be, every one of those hours costs more again. The overtime calculator works the regular rate out properly, including the bonus case that most payroll shortcuts get wrong.
What an overtime hour does not carry is incremental benefit cost. Health premiums, the PTO grant and the equipment are per person, not per hour. That is the entire reason overtime can beat hiring: you are buying labor without buying another person.
What a new hire costs, per hour
Now price the alternative on the same basis. At the same $22.00 base rate, a full time hire costs $22.00 plus 7.65 percent FICA, or $23.68 an hour. Add employer-paid health coverage at, say, $6,000 a year, which is $2.88 an hour across 2,080 hours. Add recruiting and onboarding, call it $4,000 all in for a role at this level, amortized over the first year at $1.92 an hour. That lands near $28.48 an hour.
Per hour, hiring wins easily: $28.48 against $35.52. That is the comparison most articles stop at, and it is the wrong one, because you cannot buy 8 hours of a full time employee. You buy 40, every week, whether or not the work is there.
The crossover number, and how to calculate your own
Set the two weekly costs equal. A full time hire costs 40 hours at the loaded new-hire rate. Overtime costs however many premium hours you actually need at the loaded overtime rate.
Crossover overtime hours per week = (full time hours x loaded new hire hourly cost) / loaded overtime hourly cost.
With the numbers above: (40 x $28.48) / $35.52 = 32.1 hours. Below about 32 overtime hours a week across the affected team, paying overtime is cheaper than adding a head. Above it, the new hire is cheaper and you are simply choosing not to make the change. Run the same formula with your own rates, your own benefit load and your own recruiting cost. The number moves a lot: a role at $45 an hour with richer benefits crosses over closer to 30 hours, and a role with no employer-paid health coverage crosses much later.
Two adjustments matter. First, if the overtime falls on one person rather than being spread across six, treat the crossover as much lower, because a single employee working 12 extra hours every week is a turnover risk and turnover costs more than either option. Second, the formula assumes the extra hours are steady. Genuinely seasonal peaks favor overtime even well past the crossover, because a fixed hire idles in the trough.
The four options, honestly compared
| Option | Cost shape | Best when | Main risk |
|---|---|---|---|
| Approve freely | Fully variable, invisible on the org chart | Short, genuine peaks under a quarter long | It never gets reviewed, and the cost compounds silently |
| Pre-approval with a published threshold | Variable, but measured and owned | Almost every team under 200. This is the default answer | Managers approve by reflex unless the threshold triggers a real review |
| Part time hire | Semi-fixed, usually without benefits eligibility | A steady 15 to 25 hour shortfall, especially at fixed times | Scheduling overhead, and coverage gaps at exactly the busy moments |
| Full time hire | Fixed, plus recruiting and onboarding | Sustained demand past the crossover for two quarters or more | Idle capacity if the peak was temporary after all |
The part time row is the one small employers skip most often, and it is frequently the best answer. A 20 hour hire buys the same coverage as 20 overtime hours at roughly $23.68 loaded instead of $35.52, and halves the effective crossover. It only fails when the extra work is unpredictable or has to be done by someone with deep context.
A cap controls scheduling, never payment
One legal point before you write the policy, because getting it wrong is expensive. A cap on overtime is a management control and it has no effect on what you owe. The FLSA reaches hours suffered or permitted to be worked, which includes hours nobody authorized. If an employee works 46 hours after being told not to, you owe the 6 hours at time and a half, and you may discipline them for it. What you may not do is refuse to pay.
Comp time is not the escape hatch either. FLSA section 7(o) permits compensatory time off only for state, local and interstate government agencies, and nonprofits do not fall inside that exception. Within a single workweek you can shift hours around freely, because overtime is measured weekly, but you cannot bank Tuesday's extra hours against next month. The details are in our piece on comp time and why private employers cannot use it.
The other silent cost multiplier is classification. Chronic overtime pressure tempts managers to move someone to salary to make the problem disappear, which does nothing unless the role genuinely passes the duties test and clears the salary threshold. A misclassified salaried worker still accrues overtime, quietly, for as long as the mistake lasts. Check the role against the exempt and non-exempt tests before anyone changes a title.
New for 2026: overtime is now a reporting obligation too
There is a fresh administrative cost to factor in. The One Big Beautiful Bill Act added section 225 to the Internal Revenue Code, giving employees a deduction for qualified overtime compensation for tax years 2025 through 2028. IRS Notice 2025-69 confirmed that no changes were made to the 2025 Form W-2, so employers were not required to break the figure out last year, and it also confirmed that Forms W-2, 1099-NEC, 1099-MISC and 1099-K are updated for tax year 2026 to provide separate reporting of qualified overtime compensation.
That is now. To report it you have to isolate the half time premium required by section 7 of the FLSA, separately from straight time on the same hours, and separately from any overtime you pay voluntarily or under state law, because those do not qualify. If your payroll system reports overtime as one blended figure, this is a change you need to make before year end rather than in January.
When overtime is really a pay band problem
Persistent overtime in one team is usually one of three things, and only one of them is a staffing shortage. It can be a scoping problem, where a role has accumulated duties nobody removed. It can be a process problem, where the work arrives in an unmanageable shape. Or it can be a pricing problem: you have not filled the open role because the posted range is below what the market pays, so the existing team absorbs the gap in premium hours.
The third one is worth checking before you spend another quarter on either side of the crossover, because it looks exactly like the other two from the inside. Compare the loaded cost of the overtime you are already paying against the midpoint of a defensible band for the role. Teams are often startled to find they are spending more on premium hours than the raise it would take to fill the position. Building the range from public federal wage data rather than from last year's budget is the whole point of a structured salary band, and if the answer is that you should hire, the next task is getting through the candidate pipeline quickly enough to matter, which is where it helps to run first round screening interviews automatically instead of losing three weeks to scheduling.
What to actually write down
Keep the policy to one page. Name the person who approves overtime and require the approval in writing before the hours are worked. Publish a weekly threshold per employee, commonly 8 hours, that triggers a documented review rather than a refusal. State plainly that unapproved hours will be paid and may be disciplined, because that sentence protects you twice. Commit to a quarterly review at the team level against your crossover number, and put the number in the document so the review has something to compare against.
Then hold the review. The policy is not what saves money. The recalculation is.
Sources: Fair Labor Standards Act, 29 USC 207; 29 CFR 778.109 and 778.113; IRS Notice 2025-69; Internal Revenue Code section 225 as added by Public Law 119-21. Cost figures are worked examples using stated assumptions, not survey data. This is general information about published law, not legal advice. Confirm your own policy with employment counsel.