What should a nanny agency charge in the US? Fees in 2026, and what you actually keep

Most US nanny agencies charge a placement fee of 10% to 20% of the nanny's first-year gross salary, with a minimum between $3,000 and $5,000, and many add a non-refundable registration fee of $100 to $500 before the search starts. In New York and other high-cost markets, 22% to 25% appears on the top end.

For most agencies the minimum matters more than the percentage, and almost nobody models that before setting it.

These figures come from US agencies' own published fee schedules and from industry guides, read in October 2026. Agencies that publish their fees are a self-selecting group and the ones quoting privately tend to be dearer, so read this as the visible end of the market rather than a survey.

What do US nanny agencies actually charge?

In dollars, that puts a typical full-time placement somewhere between $5,000 and $18,000, depending on the salary and the market.

A worked example, because the percentage is the number people argue about and the minimum is the number that decides the outcome. A nanny on $65,000 placed at 15% earns the agency $9,750. The same agency placing a part-time nanny on $28,000 at 15% earns $4,200, which is below a $4,500 minimum, so the minimum is what gets charged. Set that floor too low and part-time work quietly becomes the work you do for free.

Percentage or flat fee, and what is the minimum really for?

The minimum is not a pricing detail. It is the thing that decides whether you can afford to serve the bottom half of your market.

A search takes roughly the same amount of work whether the role pays $28,000 or $80,000. Sourcing, screening, scheduling, reference calls and handholding both sides to a signature do not scale with salary. A pure percentage therefore underprices exactly the placements that are hardest to fill, because lower-paying roles attract fewer candidates and take longer.

That is what the minimum is for. Set it at the point where a search stops being worth doing, not at a round number.

A flat fee is easier to quote and easier to run, and it stops the family treating your fee as a reason to negotiate the nanny's salary down. The cost is that a $40,000 placement and an $80,000 placement pay the same, and the second one is usually more work, not less.

A percentage captures seniority, and in a market where experienced nannies command $80,000 and up it captures a lot. It also means a salary negotiation between the family and the nanny moves your fee, which cuts both ways and is worth a sentence in your terms.

Why do some US agencies price the guarantee, and should you?

This is the structural difference between the US market and most others, and it is the most useful idea on this page.

Several US agencies tier the fee against the length of the replacement guarantee. One published example charges 15% for a 90-day replacement period and 20% for 180 days. The family chooses how much protection to buy.

That is a better design than the usual one, for three reasons.

It prices a real cost. A replacement is a second search at zero additional revenue. Most agencies offer a guarantee, absorb that cost invisibly, and never find out what it is. Tiering forces you to work it out.

It is honest about risk. Longer guarantees cost more because longer guarantees get claimed more. Saying so out loud is easier to defend than a single fee that quietly includes insurance the family did not know they were buying.

It lets you compete without discounting. A family pushing on price can be offered the shorter guarantee instead of a lower fee. You keep the rate and they keep the choice.

The prerequisite is knowing your own failure rate, broken down by when placements end. If you cannot say how many of your placements fail in days 91 to 180, you cannot price that tier, and you should start recording why placements end before you try.

Is your nanny an employee or a contractor?

An employee, essentially always, and this is the question that ends US agencies.

The IRS position on household workers is long-standing and not ambiguous: when a family controls what work is done and how it is done, the worker is an employee, not an independent contractor. Nannies fail the contractor test on almost every factor. Issuing a 1099 to a nanny does not make them a contractor; it makes the classification wrong on paper as well as in fact.

This matters to you even when the family, not the agency, is the employer:

• Placement agencies introduce; they do not usually employ. The family becomes the household employer, owing FICA, filing Schedule H, and in many states unemployment insurance. If your terms are vague about this, families assume you handled it.
• Agencies running temp or on-call work often do employ, and then the obligation is yours. The line between "we introduced them" and "we scheduled, supervised and paid them" is the line between those two situations, and agencies drift across it without noticing.
• Misclassification is recoverable against the employer, with interest and penalties, years later.

The practical version: say in writing which of you is the employer, for each kind of work you do, and point families at a household payroll service rather than letting them improvise. Check the current rules on the IRS household employer page rather than on any agency's blog, this one included.

State law adds to this rather than replacing it. New York, California, Massachusetts, Illinois and others have domestic worker protections covering rest breaks, overtime and written agreements, and they differ. If you place in more than one state, you have more than one rulebook.

Do you charge sales tax on a placement fee?

Usually no, and occasionally yes, which is the worst combination because it means you have to actually check.

Most states do not tax professional and personal services, and a placement fee generally sits there. A minority tax some services, and a few tax employment and staffing services specifically. The answer depends on your state and sometimes on how your service is characterized, which is a question for an accountant licensed where you operate.

What matters for planning is that this is not the UK or EU situation, where crossing a single revenue threshold changes your price by a fifth overnight. In the US the exposure is smaller and it is geographic rather than size-based. Find out once, write it in your terms, and revisit it if you expand into another state.

What do you actually keep from a placement fee?

Take a $7,500 placement fee. Before any of it is profit it pays for:

The searches that did not place. A reasonable agency fills somewhere between half and two thirds of what it takes on. At two in three, that $7,500 is covering three searches worth of work.

Screening. Background checks, driving records, reference calls, verification. The cash cost is a few hundred dollars. The time cost is larger and it falls on the searches that go nowhere as well as the ones that convert.

The guarantee. The single largest hidden cost in the model, and the one most agencies have never measured.

Your own time. A full search that reaches offer is rarely under fifteen hours across sourcing, screening, scheduling, reference-taking and managing both sides through to a signature.

At fifteen hours for a filled search, plus the two thirds of an unfilled one it subsidizes, a $7,500 fee at a two-in-three fill rate is roughly $7,500 for 22 hours of billable-equivalent work, or about $340 an hour before overhead. That sounds comfortable until you take off insurance, software, advertising, and the searches that collapse at offer stage.

The number worth tracking is not your fee. It is fee divided by hours across every search you started, including the ones that died. Agencies that count only completed placements consistently overestimate what they earn and underprice accordingly.

What about temp, on-call and backup care?

Different business, same clothes. Margins per booking are thin, admin per booking is nearly as heavy as for a permanent search, and the cash arrives weekly rather than in lumps.

Agencies do it anyway because it is the only recurring revenue in the model. A family that uses you for backup care four times a year is worth more over five years than one permanent placement, and after the first time it costs nothing to acquire them.

Two traps. The first is pricing it like an afterthought: if your hourly margin does not cover the coordination each booking generates, it is a marketing cost you are recording as revenue. The second is the employment question above, because temp work is where agencies most often become the employer without deciding to.

When should you raise your fees?

Three signals, in the order they usually appear.

You are filling nearly everything you take on. A fill rate above about 80% usually means you are being chosen on price. That is the cheapest moment to raise fees, because the demand is already there.

Your minimum has not moved in two years. Percentages track the market automatically. Minimums do not, and a floor set in 2024 is worth noticeably less now.

You are turning work away. Obvious, routinely ignored.

What does not work is raising fees quietly. Families compare, agencies talk, and an increase you cannot explain in one sentence reads as opportunism. An increase you can explain, usually a longer guarantee or deeper screening, lands fine.

What this means for how you run the agency

Two things follow, and neither is software.

Track every search, not every placement. Your real rate is set by the work that did not convert, and if the only thing you record is filled placements you will drift below where you should be without seeing it happen.

Record why placements end, and when. That is the input to pricing the guarantee, and it is the one number that lets you tier a fee the way the better agencies in this market already do.

Both of those are a column in a spreadsheet until the spreadsheet gets long, and then they are the first two columns to stop being filled in. We built placement tracking around exactly this: the reason is required when a placement is closed, and the report tells you what proportion failed inside your guarantee and what those re-searches cost you. That is the figure the 15%-for-90-days and 20%-for-180-days split above is priced from, and almost nobody has it.

Both are record-keeping, and both fall apart in a spreadsheet somewhere past a few dozen searches a year. That is when agencies start looking for a system, which is the business we are in, and we would rather you got there from your own numbers than from an ad.

See also: what UK nanny agencies charge, which covers the same ground for a market where VAT changes the arithmetic.

Figures come from US agencies' published fee schedules and industry guides read in October 2026, and describe what agencies advertise rather than what every family pays. Nothing here is tax or legal advice: household employment rules are federal and state, and they change. If your experience of the market differs, tell us: hello@rostavo.com.