How do you start a nanny agency? The decisions that matter, in order

You can take your first booking within a month. The six decisions that determine whether the agency is still trading in three years are: which model you run, how you vet, what you charge, what your guarantee is, how you find families, and when you stop using a spreadsheet.

Registration and insurance matter and are mostly paperwork. The decisions above are the business.

This is written for childcare and nanny agencies, but five of the six apply unchanged to any agency placing people into homes.

Which model should you run: placement, bookings, or both?

This is the first fork and it shapes everything after it.

Placement. You introduce a nanny to a family for an ongoing role and charge a one-off fee, typically a percentage of salary or a flat fee. Revenue is lumpy and large. You are a recruiter.

Bookings. You fill one-off or recurring care, after-school, date nights, holiday cover, and take a margin on each. Revenue is small, frequent and predictable. You are an operator.

Both, which is what most agencies that survive end up doing, because the two solve each other's problems. Placement revenue is large but unpredictable; booking revenue is steady but thin. A candidate who was not placed permanently can still earn through short-term work, so your roster stays warm and your pipeline is not wasted.

The practical catch is that the two run on different rhythms. Placement is a search that takes weeks. Bookings are a scheduling problem measured in hours. If you do both, make sure whatever you use to run the agency can hold one person's record through both, or you will end up with two systems and two versions of the truth.

What do you need registered and insured?

Country-specific, so check with your own regulator, but the shape is consistent:

• A registered business, and in some jurisdictions a specific employment-agency licence or registration. This catches people out: several countries regulate employment agencies separately from ordinary companies.
• Professional indemnity and public liability insurance. Non-negotiable, and families increasingly ask to see it.
• Employer's liability, if you employ rather than introduce. The distinction between introducing a nanny and employing one is the single most consequential legal question in this business, because it determines who runs payroll, who is liable, and how you are taxed.
• Written terms for both sides. Families and candidates, covering fees, guarantee period, replacement, cancellation and data.
• A data-protection position. You are about to hold identity documents, criminal record checks and children's details. Know where it lives and how long you keep it.

Get the employ-versus-introduce question answered properly by an accountant in your jurisdiction before you take money. It is cheap advice that is expensive to skip.

How thorough does vetting need to be?

More thorough than feels necessary, and documented.

The minimum is identity, right to work, a criminal record check at the correct level, two references taken by voice, and any role-specific qualification such as paediatric first aid. The check has a different name in every country and the level matters a great deal, so we wrote that up separately in background checks for care and domestic workers.

The part that actually goes wrong is not the first check. It is the certificate that expires eighteen months later while the person keeps getting booked, because the rota and the compliance record are two different documents. Decide now how you will prevent that, because retrofitting it at forty candidates is miserable.

What should you charge?

Placement fees in most markets sit between 10% and 20% of the nanny's gross annual salary, or a flat fee in the low thousands, with a London or major-city premium. For bookings you take an hourly margin over what the nanny is paid.

We worked the arithmetic through properly, including what a placement earns per hour once you account for the searches that do not convert: what should a nanny agency charge. The figures there are UK, but the models and the method apply anywhere.

Two things new agencies get wrong:

Pricing the margin too thin on bookings. A margin that does not cover the coordination each booking generates is not revenue, it is a marketing cost you are mislabelling. Work out what an hour of your time is worth and check that a booking covers the admin it creates.

Not having a rate card. Evenings, weekends, public holidays, multiple children, short-notice. If you decide these ad hoc, you will quote inconsistently, and families compare notes. Write them down before your first quote, not after your tenth.

What should your guarantee be?

The term agencies compete on without pricing it.

Most offer a free replacement if a placement ends within a set window, commonly eight to twelve weeks. Extending it sounds generous and cheap. Whether it is cheap depends entirely on how often your placements fail in that window, and you will not know until you have been recording why placements end.

Start at the shorter end. Lengthen it when you have the data to know what it costs you. Lengthening a guarantee is an easy announcement; shortening one is not.

How do you find your first families?

In order of what actually works for a new agency:

1. The people you already know. Your first three families will come from your own network. This is not a scalable channel and it does not need to be; it needs to produce evidence that you can do the job.
2. Local parenting groups and community forums, participating honestly rather than advertising. Every one of these has a rule against promotion and a tolerance for a person who is visibly useful.
3. Your own website with an enquiry form that goes somewhere. Not a contact email you check weekly. A form that creates a record, because the difference between replying in two hours and two days is most of your conversion.
4. Reviews from the first families, asked for directly and early. Social proof is the only thing that substitutes for a track record you do not yet have.
5. Directories, last, because they are low-effort and low-yield, and they work better once you have reviews to show.

The candidate side is usually easier than the family side, which surprises people. Good nannies are looking for an agency that communicates. Being responsive is most of the pitch.

When should you stop using a spreadsheet?

Later than software companies say, earlier than most agencies do.

A spreadsheet and a shared calendar genuinely work for the first handful of candidates and families. The point at which they stop is specific, and it is not a headcount: it is when a second person needs to be in the system at the same time as you. A spreadsheet cannot be two people's source of truth, and the first week of a coordinator working from a copy is the week your data diverges.

The other trigger is the first thing that gets missed. A booking that did not happen, an invoice that never went out, a certificate that lapsed. One is an accident. Two in a quarter is a system telling you something.

When you do move, the things that matter for an agency doing both placement and bookings are: one record per person that carries through from application to placement to booking, documents with real expiry dates and warnings, invoicing that comes out of completed work, and a way for families to request care without emailing you.

Rostavo does that for nanny and babysitting agencies, and the price is on the pricing page rather than behind a form, which in this category is unusual enough to be worth saying.

Rules differ by country and change. Check registration, licensing and employment status with a professional in your jurisdiction before trading. If something here is wrong or out of date for your market, tell us: hello@rostavo.com.