Team: who does the work, what they cost, and what to do when there isn't enough of it

Payroll is the largest line in any agency’s costs, usually by a wide margin. It is also the least flexible. A client can cancel a retainer with thirty days’ notice. A salary keeps running until you make a decision you will not enjoy making. Every article in this hub is about the same underlying problem: your revenue moves faster than your headcount can.

I have made each of these mistakes at least once. I hired for a project that ended four months later. I kept people on the bench for two quarters because I could not face the alternative. I said yes to work we could not staff, then tried to recruit a specialist in six weeks. The arithmetic was available every time, and I had skipped it.

Start with what a person costs and what they bring in

A developer on a $110,000 salary costs you about $137,500 once you add payroll taxes, benefits, equipment and software. They have roughly 1,800 hours a year available after holidays and time off. If they bill 1,350 of those hours at $150, they bring in $202,500. If they bill 1,000, they bring in $150,000 and you have a problem that no pricing change will fix.

That gap between available hours and billed hours is the bench, and it is the first thing to understand before you change anything about your team. The article on bench management works through what the bench costs in salary and in lost revenue, how much of it is healthy, and how to forecast it twelve weeks out so it stops being a surprise. It pairs with the Bill hub’s piece on utilization rate, which covers the metric itself.

Four ways to add capacity

When there is more work than people, you have four options. Each has a different cost, a different speed, and a different way of going wrong.

Hire. The slowest option and the cheapest per hour once the person is busy. If you are hiring outside your city, which most agencies now are, the process changes. How to hire remote developers covers where to find people, how to run a paid trial that tells you something an interview cannot, and the choice between contractor, employee, and an employer-of-record arrangement.

Augment. You rent a person from another firm by the hour or the month and manage them yourself. This is staff augmentation, and it is the pillar article of this hub because agencies sit on both sides of it. Some of you sell it. Most of you will buy it at some point. The margin math is different in each chair, and the contract terms that matter (conversion fees, notice periods, IP assignment, non-solicitation) are the same.

Go abroad. Whether you hire or augment, you will be offered people in other countries at lower rates. Nearshore vs offshore compares the two honestly. The short version: the hourly rate is the least important number in the comparison, and the hours of shared working day are the most important.

Partner. For some work, none of the above is realistic. A client asks for an AI system or a serious automation build, you have nobody who has done it, and the specialist you would need costs more than the project is worth. When to partner instead of hire covers white-label delivery: how the margin works, who talks to the client, who owns the IP, and the point at which the work is steady enough to bring in-house.

Here is how the four compare on the dimensions that decide it.

The mistake I see most often is choosing by hourly rate. An augmented developer at $85 an hour looks expensive next to an employee whose loaded cost works out to $76. The employee also comes with a recruiting cost, a ramp period, and a salary that continues after the project ends. For a six-month engagement, the augmented developer is cheaper. For a three-year one, the hire wins easily.

The question underneath all of this

Every one of these choices comes down to one question: how sure are you about the revenue? If a client has signed a twelve-month retainer, hire. If you have a strong pipeline and nothing signed, augment or partner until the signatures arrive. If you are staffing a kind of work you have sold once, treat it as an experiment and keep the commitment short.

Owners resist this because flexible capacity feels like renting when you could be buying. Renting is the correct choice when you do not know how long you will stay.

This is also where the commercial side of the business connects. A team is only as stable as the revenue under it, which is why retainer pricing and a reliable way of getting clients do more for job security at your agency than any amount of goodwill.

When there is not enough work

Sometimes the bench does not clear. A large client leaves, the pipeline is thin, and the runway arithmetic says the current team is not affordable for more than a few months. The hardest article in this hub is how to lay off employees. It is written for the owner of a small agency who has to do it personally, to people they know well. It covers how to tell whether it is truly necessary, how to decide who, what to check with an employment lawyer first, what to say in the room, and what you owe the people who are leaving.

It sits in the same hub as hiring on purpose. The decision to lay someone off is usually made, without anyone noticing, on the day they were hired against revenue that was never secure. Owners who understand their bench, forecast their capacity, and use flexible staffing for uncertain work have to make that decision far less often.

What AI changes

Clients now ask whether the team they are paying for needs to be as large as it is. Some of that pressure is fair. A developer with good AI tooling produces more per hour on certain kinds of work, and clients know it. I cover the larger argument in is software engineering dead. For staffing, the practical effect is that team size is a weaker signal of capacity than it used to be, and the mix of skills matters more. Many agencies will run smaller core teams with a wider ring of partners and specialists around them.

You will make the hire, augment, or partner call several times a year. Do the arithmetic each time.