What a junior actually costs

A junior on $55,000 costs about $100,000 in their first year once you count everything, and roughly a third of that first year is spent on work that is not billable to anybody.

This is the calculation most agency owners do on the back of an envelope and get wrong in the same two places. Here it is properly, so you can argue with the numbers rather than the conclusion.

The salary is about 55% of it

Take a $55,000 account executive or junior bookkeeper. What actually leaves the business in year one:

• Salary — $55,000.
• Employer payroll tax — 7.65% FICA plus unemployment, roughly $4,700.
• Health insurance — the one everybody leaves out, and the largest single line after salary. Employer contribution for one employee runs $7,000 to $9,000.
• 401(k) match — 3% is typical, about $1,650.
• Recruiting — 15% to 20% of salary through an agency, so $8,250 to $11,000. Call it $9,000, and $0 if you hire through your own network, which is the single biggest lever on this list.
• Equipment and software seats — $2,500 to $4,000 in year one.
• Liability insurance, workspace, admin — $3,000 to $6,000 depending on whether you have an office.

That is roughly $85,000 in direct cost before anybody has done any work.

The part everyone forgets: your time

A new hire is not productive on day one and the ramp is not free — it is paid for out of the most expensive hour in the business, which is yours.

Realistically that is a full day a week for the first month, half a day for the next two, and an hour or so a week after that. Call it 15 days of founder time in year one. If your own time is worth $1,000 a day, that is $15,000.

$85,000 of direct cost plus $15,000 of your own time is about $100,000. Higher in New York or the Bay Area, or if you use a recruiter. This is why "we'll just hire someone" and "that's a $55k problem" are not the same sentence.

And the first three months are mostly not billable

They are learning your clients, your tools, your tone, and which of your processes are real versus which you described optimistically to them in week one. A generous estimate is 40% billable utilisation in the first quarter, rising to 70% by the end of the year. On a 1,600-hour year that is roughly 400 hours you paid for and could not bill.

When it is obviously worth it anyway

Hire when the work needs a person and would still need one if the tools were perfect:

• Judgement calls a client will argue with. Scope, price, bad news, anything where the answer is negotiated rather than derived.
• Relationships. A client who trusts a named human stays through a bad month. A client who trusts a system leaves during one.
• Work that grows the business rather than delivering it. Partnerships, positioning, the awkward conversation nobody has had yet.

Every one of those is worth $100,000 without much argument, because the alternative is you doing it and doing less of something else.

When it is a bad trade

The tell is that the work is the SAME shape every time and grows linearly with client count. Monthly reporting. Reconciling books. Chasing overdue invoices. Meeting notes. Preparing the same pack with different numbers in it.

Hiring for that buys you a person who does a fixed amount of it, and the amount grows with every client you win. So you hire again at fifteen clients, and again at twenty-five, and the margin that made the agency worth running gets consumed by the machinery of running it.

The question to ask about any recurring task is not "how long does this take?" but "does the time go up when I add a client?" If it does, that is a process problem wearing a headcount costume.

The comparison people actually want

$100,000 a year is about $8,300 a month. Software that handles the recurring half sits between $300 and $900 a month depending on what you run through it, which is roughly one twelfth.

That is not the honest comparison though, and it is worth saying so on a page that would rather you drew the flattering conclusion. The honest version has three parts:

• It does not replace the judgement. Somebody still approves what reaches a client. If your recurring work is 80% assembly and 20% judgement, you can automate the 80% and you still need somebody for the rest.
• Setup is real. Any system needs your processes written down, and most agencies discover during that exercise that half of theirs were never written down at all. Budget a week.
• A junior compounds. In three years they are a senior who runs accounts. Software in three years is software. If you are building a firm rather than a book of business, that difference is the whole argument.

A rule that works

Write down every recurring task for one month. Two columns: does the time go up when I add a client, and would a client notice if a machine did the first draft.

Everything that is yes-and-no is the automation list. Everything that is no, or yes-and-yes, is the hiring list. Most agencies find the first list is bigger than they expected and that it is also the list they hate.