Resource Planning in a World of Shared Teams
The plan says half an integration specialist for twelve weeks. Twelve weeks later the project has had about six days of him, most of them in half-day fragments, three of them in the same week in March when somebody else's release slipped. Nobody lied and nobody underperformed. The plan expressed his availability as a percentage, and a percentage is an average of something that never arrives evenly.
Shared people are now the normal case in most organisations, so this is not an edge condition to be worked around. It is the environment the plan has to be built for, and the techniques that make a plan survive it are mostly about replacing averages with commitments and making the sharing visible while it can still be changed.
Plan the days, not the percentages
Ask for named days. Two days a week for eight weeks is a plan, because it can be put in a diary, checked against, and missed visibly. Forty per cent for a quarter is an accounting entry. The difference costs nothing to obtain at the point of asking and it decides whether you will know in week two or week nine that the work is not happening.
Ask for blocks, and size them honestly. A specialist reassembling context on a complex system pays a real cost every time they come back to it, and on unfamiliar work half a day is mostly that cost. Two consecutive days are worth considerably more than four half-days, and where the work genuinely needs continuity, say so when the allocation is agreed, because retrofitting a block into somebody's diary afterwards is a much larger favour than asking for one at the start.
Three assumptions that break these plans
That availability is uniform. A plan that spreads a person's time evenly across a period will be wrong at both ends: they will be less available than planned when their other work peaks, and the peaks are usually predictable if anybody asks. Month-end, term start, regulatory submission, seasonal maintenance windows: the other demands on a shared person have a shape, and the shape is knowable.
That a percentage survives contact with a priority call. When two projects need the same person in the same week, somebody decides, and the decision is usually made by whoever asks most urgently or most senior. The Resources Performance Domain of the PMBOK® Guide Eighth Edition, at Section 2.6, treats capacity as part of the project's resource work, and in a shared environment that means planning for the contested week specifically, not for the average of the quarter.
That the person is interchangeable with their role. Plans are written in roles and delivered by people, and on specialist work the difference is large. Where the plan depends on one individual who holds knowledge nobody has written down, that is a single point of failure with a name, and it should appear on the risk register as clearly as a sole-source supplier would.
One specialist and four implementations
A SaaS vendor ran customer implementations in parallel: typically five or six live at any time, each with a project manager, a configuration analyst and shared access to two integration specialists. One of the two had built the connector framework and was the only person who could handle anything unusual in it.
Every implementation plan allocated him twenty per cent. Four of them were running simultaneously, which was accepted as normal, and in aggregate the allocation was eighty per cent, which looked prudent on the resourcing sheet. What the sheet could not show was that integration work is front-loaded: every one of those implementations needed him heavily in its first three weeks and barely at all afterwards. Two customers signed in the same month, and the two three-week windows landed on top of each other.
The visible symptom was four laptops on his desk and a fortnight in which he touched each project for an hour or two a day. Both new implementations slipped by around three weeks, and the slippage was blamed on customer data quality, which was a real problem and not the reason.
The change was made at the portfolio level, not inside any one project. Implementation start dates were spaced so that no two integration windows overlapped, which cost about two weeks of delay on one sale and removed roughly six weeks of slippage across the year. The second change was slower and mattered more: the framework knowledge was deliberately spread, with the second specialist paired into the front three weeks of each implementation until he could handle the unusual cases alone. Eight months later the constraint was two people wide, and the scheduling conversation stopped being a queue.
When the plan cannot be made safe
Sometimes the arithmetic does not work, and the honest response is to say so with the aggregate picture attached. A project manager reporting that their own plan is at risk is one voice among several. The same project manager showing that one specialist is committed to four projects in the same three weeks is describing an organisational problem, and it belongs in front of whoever owns the portfolio, because nobody inside a single project can solve it.
The other honest response is to change the shape of the work. Where a scarce person cannot be had when needed, the options are to move the work to when they are available, to reduce what only they can do by pairing somebody alongside them, or to accept a later date and say so now. All three are uncomfortable and all three are better than a plan that quietly assumes the problem will resolve itself.
For a PMP® candidate, the useful reading is that shared-resource planning is about commitments and timing rather than allocation percentages, so a situation where a part-time specialist has not delivered is worth examining for what was actually agreed. A response that asks for more of the person's time has not established whether the time was ever schedulable. Planning situations where the capacity exists in total and never exists when it is needed form the backbone of a structured PMP exam preparation course's resourcing material.
One exercise is worth doing this week. Take the three scarcest people in your plan and write down, for each, which specific days you are counting on and what else they are committed to on those days. Where you cannot fill in the second column, you do not have a resource plan for that person yet, and the people who can fill it in are usually one conversation away.
Interested in going further?
Planning around shared people is a scheduling problem with a governance problem underneath it, and the second one is invisible from inside a single project. Omega's PMP® Exam Preparation works through capacity decisions where the person exists and the availability does not.
Capacity, availability and the rest of the resource work are treated together in the PMBOK® Guide Eighth Edition.
Ad · Amazon affiliate link.
References
A142: The PMBOK 8 Resources Performance Domain: What It Really Covers
A150: Motivation Is Not a One-Size-Fits-All Technique
A145: Resource Levelling vs Resource Smoothing
A147: Conflict Is Not Always a Problem
A143: People Are Resources, But Not Just Resources
PMP and PMBOK are registered marks of the Project Management Institute, Inc.
