7 Capacity Planning Mistakes Every Resource Manager Should Retire

Every services firm's capacity plan starts life clean. Twelve weeks of allocations, target utilization by role, an owner per project, a Monday review on the calendar. By month three, it has drifted. By month six, it is a decorative artifact that the resource manager updates just before the QBR.

The decay is not random. It comes from the same seven mistakes, in slightly different combinations, at every firm. Retire these habits and the plan stays useful for the whole quarter.

Mistake 1: Are you planning at monthly averages instead of weekly percentages?

Monthly averages are the most expensive habit in capacity planning. They look tidy, they roll up cleanly to the finance forecast, and they hide almost everything that matters.

A person allocated 120% for two weeks and 60% for two weeks looks like 90% in a monthly average. Two of those weeks were problems. One was overload, one was bench. Both went uncorrected because the average was fine.

The fix is weekly percentages, floored at 5% increments. The grid gets more cells, but the signal per cell increases by an order of magnitude. Every reasonable staffing decision downstream, from bench alerts to overallocation flags, requires this resolution.

Mistake 2: Are you treating overallocation as a formatting problem?

The second-worst habit is smoothing 115% down to 100% in the plan because "the row does not fit above the ceiling."

If someone is planned at 115%, either they will work 115%, or someone else will pick up the slack, or the work will slip. The plan should say 115% until the decision is made, and the decision should be one of: hire, subcontract, reduce scope, move dates, or accept the risk explicitly.

Smoothing to 100% skips the decision and delegates the resolution to whoever finds out first. Usually that is the person themselves, on a Wednesday night, discovering they cannot deliver everything they were told they would.

Mistake 3: Does every delivery lead edit the plan directly?

Shared-editing plans die faster than any other kind. Every lead trusts their own rows, nobody trusts anyone else's, and reconciliation happens monthly by whoever has time.

The fix is one editor, one plan, and a fast request channel. The resource manager owns the grid. Delivery leads submit staffing requests in writing. The resource manager commits to a response window, usually four business hours, that is fast enough to remove the temptation to just edit the plan directly.

This is a process change, not a tooling change. Firms that skip the process change and buy tooling instead end up with a shared-editing spreadsheet inside a fancier interface, and the same decay.

Mistake 4: Are you planning only from booked pipeline?

If your plan only reflects signed deals, you are systematically missing two categories of risk.

• Deals in late-stage pipeline. If a deal at 70% probability is scheduled to start in six weeks, you need to hold capacity now. Not committing until it is signed means either it slips or you scramble.
• Renewals and expansions. A renewal that is 80% likely to expand by 30% in Q4 needs capacity earmarked in Q3. If you only staff from signed paper, you will always be a step behind.

The fix is a probabilistic plan layer. Show signed capacity, plus a shaded band for weighted pipeline capacity. The band moves as the pipeline changes. Nobody commits headcount against pipeline alone, but they can see the shape of the quarter honestly.

Mistake 5: Are you staffing by availability without checking skills?

The bench and the request are both real, but they never match up because the plan does not know who can actually do the work.

• The pattern. A delivery lead needs a senior engineer with healthcare experience. The plan shows three benched senior engineers. Two of them have never worked in healthcare. The lead hunts anyway.
• The cost. Either the wrong person gets staffed and the engagement takes longer than it should, or the lead subcontracts, or the request lingers and the bench grows.

Skills matching does not require an elaborate taxonomy. Fifteen tags per person, refreshed twice a year, is usually enough. The point is to reduce the search space from "who is free" to "who is free and can do this," which cuts most staffing hunts from days to minutes.

Mistake 6: Are you reviewing the plan monthly instead of weekly?

Monthly reviews are too slow for a services firm. By the time you notice the pattern, the hours are already gone.

The right cadence is weekly, with a specific structure. Twenty to 40 minutes on a Monday, resource manager and heads of delivery, agenda ordered by red cells. The point of the meeting is decisions, not updates. Every red cell either resolves in the meeting or leaves with a named owner and a deadline.

The quarterly review is separate. It re-baselines targets, checks role mix against demand, and reviews the utilization trend. Do not conflate the two. The weekly review is tactical, the quarterly is strategic, and putting both in one meeting makes both worse.

Mistake 7: Are you conflating capacity with utilization?

Capacity is a plan input. Utilization is a plan output. They are not the same thing, and treating them interchangeably confuses every downstream decision.

• Capacity is the maximum billable hours a person can deliver in a period, given their target utilization and PTO. It is a fixed input, decided by role and calendar.
• Utilization is the percentage of available hours actually billed. It is a measured output, calculated from time tracking against the target.

If someone says "we have 800 hours of capacity next week" they mean input. If someone says "we ran 74% utilization last week" they mean output. Mixing them, as in "our capacity is 74%," is the source of a lot of muddled conversations.

Which mistake should you fix first?

Here is the rough priority order for a firm that has all seven. Fix in this sequence.

1. Move to weekly percentages. Immediate, high-value.
2. Stop smoothing overallocation. Immediate, no cost.
3. Move to a one-editor plan with a fast request channel. One-week transition.
4. Weekly review cadence, structured agenda. One-week transition.
5. Add skills matching to the grid. Two-week setup, ongoing maintenance.
6. Add probabilistic pipeline layer. One-month setup.
7. Clean up capacity vs utilization language across the firm. Ongoing.

The first four are almost free and account for most of the improvement. The last three take more work but compound over quarters.

What actually matters

Every one of these mistakes has the same root: the plan is being asked to do less work than it should, either because the resolution is too low, the ownership is too diffuse, or the cadence is too slow. A capacity plan that runs at weekly resolution, with one editor, on a weekly cadence, with real skill data behind it, is not more expensive to maintain than a spreadsheet. It is just more honest. Every point of utilization the firm recovers over the next quarter comes from the honesty, not from the tooling.