Capacity planning for agencies: Why Start Before You’re Big
There’s a belief I ran into a lot in my own agency, and I still hear it from owners now. Capacity planning is something you put in place once you’re big. It’s for the fifty-person shop with a dedicated traffic manager and a wall of monitors, not for a team of twelve who can still sort most things out in a Monday meeting. So you leave it. You tell yourself you’ll set it up properly once things settle down, once you hire the next few people, once the systems feel like they need it.
I want to make the opposite case. Capacity planning is most valuable in the years before you think you need it, because those are the years your agency is forming the habits it will run on for the next decade. Put the discipline in early and your processes grow around your talent. Leave it until the wheels are coming off and you’re not planning capacity anymore, you’re firefighting the absence of it.
What capacity planning actually is
At its simplest, capacity planning is knowing how much work your team can take on, how much they’re already carrying, and what’s coming toward them, so you can make good decisions about who does what next. It sits one level up from task management. A task list tells you what needs doing. Capacity planning tells you whether the people you’d assign it to actually have room to do it.
For an agency specifically, it means seeing booked hours against available hours, per person and per department, across the current week and the two or three weeks ahead. That’s the whole idea. Everything else is detail.
I’ve written separately about what resource planning looks like in practice for a 10-to-50-person agency, the weekly rhythm, the different ways agencies handle it, the traffic-manager function when you can’t afford a traffic manager. This piece is about something that comes before all of that: the decision to start at all, and the case for starting sooner than feels necessary.
Why “we’re too small for this” is the expensive answer
The reasoning sounds sensible. A small team has fewer moving parts, so the picture is easy to hold in one person’s head. Why add a system for something you can already manage by talking to each other?
Here’s the part that reasoning misses. The point where holding it in your head stops working doesn’t announce itself. There’s no week where the agency crosses a line and you think, right, now we need capacity planning. Instead, it degrades gradually. You take on a couple more clients, hire two more people, add a retainer with weekly touchpoints. Each addition is manageable on its own. Then one week two projects peak together, your best designer is at 140%, someone’s on leave you’d forgotten about, and you realise the informal system stopped being reliable a while ago. You just didn’t have a moment to notice.
By then, putting capacity planning in is a change-management project. You’re introducing a new discipline to a stressed team in the middle of a crunch, which is the worst possible time to ask people to change how they work. If instead the discipline was already there, small and lightweight, from when the team was eight people, it simply scales up with you. The habit is set. The data has history. Nobody has to be convinced of anything.
Starting early is cheaper because you’re building a habit, not repairing a failure.
What you gain by planning capacity early
Your process forms around your talent. This is the one I care about most. When you plan capacity from early on, you learn how your actual people work: who’s fast at concepting and slow at finishing, which department is the real bottleneck, how long things genuinely take rather than how long you hoped. Your process then develops around those truths. Agencies that skip this end up with a process shaped by whatever chaos formed in the absence of planning, and then they try to force their talent to fit it. Building the other way around, talent first and process second, produces an agency that works with the grain of its people.
You grow more efficiently. Capacity data tells you when you actually need to hire, rather than hiring in a panic after a bad month or holding off too long and burning out the team you have. When you can see the team running at 95% for six weeks straight with a full pipeline behind it, that’s a hiring decision made on evidence. When you can see a department sitting at 60% while another drowns, that’s a redistribution or a cross-training decision. Growth stops being a series of reactions and becomes something you can pace.
You spot collisions while there’s still time to solve them. A resourcing conflict seen three weeks out is a scheduling adjustment. The same conflict seen three days out is overtime, compromised quality, and an awkward client call. The value of capacity planning is almost entirely about how early you see things. Early enough, and every problem has cheap solutions. Late, and your only options are expensive.
You protect your people before they break. Overload rarely arrives as a dramatic event. Someone absorbs an extra project because they had a gap, then a revision cycle runs long, then a retainer request lands, and they’re at 130% before anyone counted. Being able to see that pressure building, rather than hearing about it after a deadline slips or a good person hands in their notice, is worth more than any single feature of a planning system.
You can answer the capacity question honestly. When a prospect asks whether you can take on a project starting in three weeks, “let me check and come back to you” is a fine answer only if checking takes a minute. If it takes a day of asking around, you either guess and risk overcommitting, or you stall and risk the work. A live view of team capacity means you can say yes or no with confidence, which is its own competitive advantage.
Start small: capacity planning does not mean heavy process
The fear underneath “we’re too small” is usually a fear of bureaucracy. Owners picture timesheets policed to the minute, people resentful of being tracked, planning meetings that eat half a day. That’s not what early capacity planning has to look like, and heavy process at a small agency would do more harm than good.
Starting early means starting light. A realistic estimate of hours on each piece of work. Those hours mapped against each person’s available time. A quick look at the week ahead and the two beyond it. That’s enough to catch most of what matters at a team of ten to twenty. The discipline is in doing it consistently, not in doing it exhaustively. You can add depth as you grow, and because you started early, adding depth feels like a natural extension rather than a new imposition.
The habit is the asset. The tooling around it can stay simple for a long time.
What to look for in a capacity planning tool for an agency
Most general capacity planning tools were built for software teams or professional services broadly, and they treat an agency like any other project business. The things that matter for an agency specifically tend to be missing or bolted on.
A few things worth insisting on. The tool should think in departments, because that’s how agency work is briefed, resourced and reviewed, and department-level capacity is often more useful for planning than the individual view. It should spread estimated hours across a person’s real working days rather than dumping a lump of time on one date, so the plan reflects how work actually happens. It should show you real availability, which means meetings and leave and public holidays sitting on the same grid as booked tasks, because an afternoon of client calls is time that can’t also be spent on production. And it should let you see forward, two or three weeks minimum, because a tool that only shows you this week keeps you permanently reactive.
This is the thinking behind how Agencydesk approaches capacity, since it was built only for agencies rather than adapted from a general tool. The Resource Planner groups people by department, each with its own capacity indicator, so you can see which departments are near the limit before you commit to a deadline. Dragging time across a job spreads it evenly and caps it at each person’s working hours, and Google Calendar or Microsoft 365 meetings overlay on the grid alongside leave and public holidays, so you plan against genuine availability. Because the estimates come from the same job costing that drives scheduling and billing, the capacity view stays current as the work moves rather than needing a separate spreadsheet that drifts from reality by Wednesday.
The point isn’t the feature list. It’s that capacity planning for an agency only works when the tool reflects how an agency is actually shaped, and most don’t.
The honest version
You do not need to be big to plan capacity, and waiting until you are big is what makes it painful. Start when you’re small, keep it light, and let the discipline grow with you. Your process will form around the people you actually have, you’ll hire and grow on evidence rather than panic, and you’ll spend far less of your life firefighting problems that a forward view would have shown you weeks earlier.
If you want to see what planning capacity by department looks like against your own team, you can book a demo and we’ll walk through it with your structure in front of us. The Resource Planner guide in the help centre goes through the mechanics in detail if you’d rather read first.
Four principles for resource planning that actually hold up
Most resource planning advice stops at “know who’s available.” That’s the starting point, not the discipline. These are the principles that separate a plan that reflects reality from one that falls apart by Wednesday.
Plan against real availability, which means overlaying meetings and leave. A capacity plan built only on task hours is planning against time your team doesn’t actually have. If someone has six hours of client calls on Tuesday, those hours are gone, and a plan that shows them as available is fiction. Overlay the meeting schedule and the leave days onto the same view as the booked work. Without that, you’re planning around time you won’t have, and the plan will be wrong before the week starts.
Calibrate capacity to each person’s real working hours, not a flat eight-hour day. People don’t have eight billable hours in a day, and they don’t all work the same week. A useful planner normalises every person’s capacity to their own working hours per day and to a realistic productive quota rather than a theoretical maximum. When the capacity bars are calibrated that way, a person showing as full is genuinely full, and you can trust the number instead of mentally discounting it. This ties directly to recovery hours: if you don’t know a person’s realistic billable capacity, every plan you build on top of it inherits the error. (This is the same reasoning behind measuring utilisation against realistic hours rather than the full day, which I covered in the piece on utilisation and recovery rate.)
Make resource planning a shared, live activity, not a document one person maintains. The best resourcing happens when the team plans together, out loud, trading work across the week: this campaign goes here, that CI project slots into the quiet window, this person picks up the overflow. It has the feel of an open trading floor, everyone adjusting in real time against a picture they can all see. That only works if the plan updates live across every user as work is booked. A static spreadsheet that one person owns can’t support that, because by the time they’ve updated it, three other things have changed. A live shared view turns planning from an admin chore into a fast weekly conversation.
Drive the plan with accountability, not automation triggers. This is the one most people get backwards. It’s tempting to buy software with smart automation chains: the design task fires automatically when copy is marked complete, dev kicks off when design is done, and so on down the line. It sounds efficient. The problem is that a trigger chain hands control of the timeline to whoever is doing the current step. The chain finishes when they finish. But the client deadline doesn’t move to accommodate that. What actually works is the opposite logic: you book the work into people’s time and hold the team accountable to the booking. The deadline is fixed, so the plan has to be driven by commitment to dates, not by one task politely waiting for another to complete. That is the entire philosophy of resource planning. You decide when the work needs to happen, and you resource it to make that real, rather than letting the work decide its own pace. Here’s what that looks like in practice.
Take a job with a fixed client deadline of Friday. The chain is copy, then design, then development. Under the automation-trigger approach, design only starts when copy is marked complete. Copy runs two days late, so design fires on Wednesday instead of Monday. Development now has a day and a half for two days of work. The trigger chain worked exactly as designed, each step waited for the one before it, and the project still misses Friday, because the automation protected the sequence rather than the deadline.
Now run the same job on a booked plan. Copy, design, and development are each booked into specific people’s time against the Friday deadline, working backwards: development needs Thursday and Friday, so design is booked Tuesday and Wednesday, so copy is booked Monday and Tuesday. When copy starts slipping on Monday, the plan shows the collision immediately, because Wednesday’s design booking is now at risk. You see it on Monday afternoon, not Wednesday morning, and you have three days to act: add a second writer, trim a revision round, or have the deadline conversation with the client while it’s still a conversation rather than a failure. The deadline is the fixed point, and the plan is built to defend it.
Same team, same delay, same deadline. The difference is that one approach lets the work set its own pace and hopes the dates hold, and the other holds the dates and makes the work fit them. That is the whole of it.