Capacity Planning: Who Has Time for What Next Week

Why teams promise more every week than they can deliver

On Monday the week looks good. The quote for the major customer, preparing the trade fair, two customer meetings, the new price list. By Friday two of them are done, the rest has slipped into next week. In between there were calls, a complaint, a colleague out sick, and three questions from the workshop. None of it was unusual, and yet none of it was in the plan.

If this keeps happening, promises lose their value. Customers hear “next week” and count on the week after, colleagues plan their work around deadlines that will not hold anyway. This article answers the question: how does a small company plan reliably who has time for what next week? To do so, we clarify the terms, introduce a thinking model from decision research, weigh two views on the accuracy of estimates, and work through a weekly capacity calculation in full.

A plan that allocates every hour has no room left for reality by Monday noon.

What is capacity planning, and which time can actually be planned?

Capacity is the working time a person or team has available in a given period. Capacity planning matches this time with demand: which tasks are pending, how much time do they need, and who takes them on? In a small company, the sensible period is usually the week, because it coincides with the weekly planning described in the article on weekly team planning.

What matters is which time can actually be planned. Vacation, sickness, and fixed appointments come off contractual working time first. Not all of what remains is usable, because breaks, short check-ins, and switching between tasks cost time. The article on how to calculate process costs sets this practical capacity at around 80 percent. And finally there is a base load: daily business that comes up every week without anyone planning it, such as inquiries, complaints, and questions. Only what is left after that is free for projects and initiatives.

The number at the end surprises many owners. But it explains why a plan that assigns 35 hours of project work to one person cannot hold. The base load cannot be estimated, it has to be read from the past, for example from a ticket system as described in the article on bundling customer inquiries, or from one week of time records.

The thinking model: the planning fallacy

In 1979, Kahneman and Tversky described a tendency they called the planning fallacy. People underestimate the time they need for an upcoming task even though they know that similar tasks took longer in the past. The cause lies in how planning is done. Whoever estimates imagines the path from start to finish, and that image is idealized. Interruptions, questions, and competing tasks hardly appear in it.

Wiese, Buehler, and Griffin (2016) highlight a distinction that is central to capacity planning. Performance time is the time someone actually spends working on a task. Completion time is the point at which the task is finished. Between the two lie other tasks, interruptions, and waiting times. According to research, estimates of completion time are considerably more prone to optimism than estimates of performance time. Whoever says “That takes four hours” may well be right. Whoever turns that into “That will be done tomorrow” often is not.

An analogy makes this tangible. According to the navigation system, the drive to the customer takes 40 minutes. Anyone who wants to be on time still leaves an hour earlier, knowing that refueling, looking for parking, and the traffic jam on the arterial road are not included in the 40 minutes. For office tasks this experience is often missing, because nobody records how long things really took.

What research shows about better estimates

In four experiments, Wiese, Buehler, and Griffin (2016) tested a method often recommended in project practice: backward planning. You start at the target date and go through the necessary steps in reverse order back to today. Participants who planned backward predicted longer completion times in all four studies than those who planned forward or without instructions. In the fourth study, where the prediction could be compared with the actual result, these predictions were less biased. Backward planning drew more attention to obstacles, interruptions, and competing demands.

In their introduction, the authors also summarize earlier findings. People who are asked to explicitly recall earlier completion times estimate less optimistically. In large construction projects, so-called reference class forecasting reduces time and cost overruns. It aligns the estimate with the actual results of comparable earlier projects rather than with one's own plan. Breaking a task into sub-steps, by contrast, showed mixed results and can even increase optimism when a very detailed plan creates a false sense of certainty. Buehler, Griffin, Lam, and Deslauriers (2012) also found in four studies that participants who imagined a task from an observer's perspective predicted longer and therefore less biased completion times. They paid more attention to possible obstacles.

Two views: do we always underestimate?

The first view is: estimates are fundamentally too optimistic, so every estimate should be increased across the board. The broad research on the planning fallacy supports this.

The second view disagrees. Thomas, Newstead, and Handley (2003) found no underestimation but an overestimation of duration for a simple, clearly structured task. Those who had done the task before estimated more accurately. The authors suspect that the duration of the task plays a major role: for short, simple tasks people tend to overestimate the time.

Both findings can be reconciled, and a useful rule follows. Short routine tasks that the team knows tend to be estimated generously and need no markup. Longer, new, or dependent tasks are underestimated, above all regarding when they will be finished. For these, looking at earlier similar cases is worthwhile. A flat markup on everything is therefore less accurate than distinguishing by type of task.

Why full utilization destroys deadlines

A second reason for missed deadlines has nothing to do with estimates. It lies in utilization. Queueing theory, a branch of mathematics, describes what happens when tasks arrive irregularly and take different amounts of time. The closer utilization gets to 100 percent, the faster waiting times grow, and not evenly but ever more steeply.

At 80 percent utilization, a new task waits on average four times as long as its processing takes. At 95 percent, already nineteen times. For a customer waiting for a quote, that is the difference between two days and two weeks. It follows that a buffer in weekly planning is not a convenience but the precondition for urgent work to find room at all. Limiting parallel work, as described in the article on OKR, Kanban, and weekly planning, works in the same direction.

How to plan capacity for next week

1. Available hours per person

   Subtract vacation, sickness, and fixed appointments from working time. Take 80 percent of the rest as practically usable.

2. Subtract the base load

   Subtract the average daily business of the last four weeks. Without measurement, start with a third of usable time and correct it after the first weeks.

3. Estimate initiatives

   Estimate the performance time for each initiative. For longer or new tasks, look at how long the last similar ones took, and plan backward from the deadline.

4. Allocate only up to plannable time

   Allocate initiatives only up to each person's plannable time. Whatever does not fit is postponed or handed off, and on Monday, not on Friday.

5. Compare plan and actual on Friday

   Note for each initiative how long it actually took. This list is the reference class for the next estimate.

The last step is the most important and the one most often skipped. After a few weeks, the team has its own collection of actual durations that improves every estimate. When initiatives no longer fit into the week, that is not a planning error but information for a decision: postpone, hand off, or drop something else. The article on how to delegate tasks describes handing off with a clear handover. For larger initiatives with several people involved, the article on planning projects in small teams helps.

Conclusion: how does a small company plan reliably?

The question was
a small company plans reliably who has time for what next week. The answer consists of three honest calculations. First, only the time left after appointments, breaks, and daily business counts, and that is often less than half of working time. Second, longer and new tasks are underestimated above all regarding when they will be finished, and looking at earlier similar cases helps more against this than a flat markup. Third, every plan needs a buffer, because full utilization lengthens waiting times disproportionately. Planning does not become reliable through more detailed plans but through the weekly comparison of plan and actual.

What you can put into practice right away

1. Calculate each team member's plannable time for next week: working time minus appointments, 80 percent of that, minus base load.
2. Place the planned initiatives next to it and check where more is assigned than can be planned.
3. Decide on Monday what gets postponed or handed off.
4. On Friday, have the actual duration noted for each initiative.
5. Repeat this for four weeks and use the list of actual durations for the next estimates.

After four weeks you will know how large the base load in your team really is and how far estimates and results diverge. With these two numbers, promises to customers become reliable again.