Twenty percent allocation is a scheduling fiction. It is a number typed into a resource column, not a description of human availability. The number survives because it is easy to enter, easy to approve, and impossible to verify until the work is late.
This is a case-file analysis of how that fiction is manufactured, where it hides in project artifacts, and what a usable procedural close looks like.
Case file: the 0.2 FTE that never existed
This case file is fictional. The names, dates, documents, and outcome are invented to make a planning risk visible; it is not testimony about a real project.
Artifact 1 — Project charter clause:
“The Data Migration Lead will be assigned at 20% allocation for the duration of the project (approximately 26 weeks), reporting to the Program Director for project matters and to the IT Operations Manager for line matters.”
Artifact 2 — Resource schedule metadata:
Resource: D. Okafor
Allocation: 0.2 FTE
Start: 2026-01-05
Finish: 2026-07-03
Calendar: Standard (Mon–Fri, 8h)
Booking type: Soft
Notes: [blank]
Artifact 3 — Status line, week 9:
Data Migration: Amber. Lead reports competing priorities from Operations. Revised estimate for workstream completion: +3 weeks. No change to overall project end date assumed.
Artifact 4 — Redacted email, week 11:
From: IT Operations Manager
To: Program Director
Subject: Re: D. Okafor availabilityWe agreed 20% in the charter. In practice I need him on incident response and the audit remediation. I can protect maybe one day a fortnight, and not in the same week every time. If you need more than that, we need to talk about backfilling my team.
In this fictional case, the schedule slips because 0.2 FTE was treated as available work time when the managers had only agreed to an allocation on paper.
What fractional allocation actually means
A resource schedule can record a fraction of a workweek without proving that the hours will be free when the project needs them. In the example, a 20% assignment means a planned share of time. It is not yet a commitment to eight protected hours every week. The charter still needs a pattern, a priority rule, and an answer to what work will move to create the capacity.
The arithmetic is seductive. Twenty percent of forty hours is eight hours. Eight hours a week for twenty-six weeks is 208 hours. At a blended rate, that is a budget line. The arithmetic is correct, but it does not establish whether those hours are available. That is a separate planning question.
Three structural facts break the arithmetic:
- In this case, the fractional allocation is additive, not subtractive. The lead remains on incident response and receives a project responsibility on top of it. The 20% is drawn from slack that may not exist.
- Context switching has a cost that the schedule does not model. A person splitting time across two roles loses time to reorientation, handoff, and the cognitive tax of holding two sets of priorities. The schedule records eight hours; the effective output is lower.
- Priority is not allocated. This charter assigns time but does not assign precedence. When Operations has an incident and the project has a milestone, the person resolves the conflict by whatever local rule applies — the local decision rule they understand, which may differ from the project plan.
None of this is a secret. It is simply not represented in the artifacts that govern the project.
Where the fiction hides
Fractional assignments are encoded in documents that look precise. Each artifact has a specific failure mode.
The charter clause
A charter clause that states a percentage without stating precedence, backfill, or escalation is not a commitment. It is a hope with a number attached. The clause above says the lead reports to two managers. It does not say which manager wins when the two sets of priorities collide. That omission is the failure mode.
The resource schedule
Scheduling tools accept fractional allocations because the data model supports them. A booking type of “soft” is a flag that the allocation is provisional, but the flag is rarely read by anyone outside the scheduling function. The schedule shows a continuous 0.2 FTE line. The reality is a discontinuous, unpredictable set of hours that the tool cannot represent.
The status report
In this case file, “competing priorities” is the first written signal that the agreed allocation is not being honored. The report records the symptom without naming the cause, because naming the cause would require a decision the governance body has not made. The revised estimate is absorbed into the project’s contingency, which hides the structural problem behind a schedule adjustment.
The email trail
The email above is the most honest artifact in the file. It states the real availability — one day a fortnight, irregular — and it states the condition for change: backfill. That condition was never met. The project continued to plan against 0.2 FTE.
Why matrix organizations make this worse
In a functional organization, a project manager negotiates with one line manager. In a matrix organization, the project manager negotiates with a line manager who has their own operational mandate, their own performance metrics, and their own definition of priority. The project manager often has no formal authority over the resource.
The fictional charter gives the project lead and the operations manager competing claims on the same person’s time. That tension need not be a defect if the charter states who has priority and how a conflict is decided. Without a real capacity commitment or an escalation path, the 20% entry postpones the decision rather than resolving it.
The line manager is not being unreasonable. They are protecting their operational commitments. The project manager is not being naive. They are working with the allocation they were given. The failure is in the governance layer that approved a number without approving the conditions that would make the number true.
A procedural close
The following procedure is a recommendation, not a standard. It is designed to be run before the allocation is approved, not after the project is late.
Step 1: Replace the percentage with a named pattern
Do not accept “20%.” Ask for a pattern. For example: “Every Tuesday and Thursday morning, 08:00–12:00, excluding the first week of each month.” A pattern is verifiable. A percentage is not.
If the line manager cannot name a pattern, the allocation is not real. Record that finding in the charter.
Step 2: Name the precedence rule
Add a clause to the charter that states which manager has priority when the two roles conflict, and what the escalation path is when the rule is unclear. A precedence rule does not eliminate conflict. It makes the conflict visible and resolvable.
Step 3: Require a backfill condition
If the fractional assignment draws from a role that has operational commitments, the charter should state what is removed, deferred, or backfilled to create the capacity. “Nothing” is an acceptable answer only if the line manager confirms that the role has slack. That confirmation should be in writing.
Step 4: Track actual hours against the pattern
Do not track against the percentage. Track against the pattern. If the pattern is “Tuesday and Thursday mornings,” record whether those mornings were actually available. A simple log — date, planned hours, actual hours, reason for variance — is enough. The log is the evidence that the allocation is or is not being honored.
Step 5: Escalate at the first variance, not the third
The status line in week 9 was the signal. It was absorbed into the project’s contingency. The procedural close is to escalate the variance at the first occurrence, with the log as evidence, and to require a governance decision: change the allocation, change the scope, or change the date. Absorbing the variance is not a decision. It is a deferral.
Step 6: Close the file with a named failure mode
If the project fails, the post-mortem should name the failure mode. “Fractional allocation without precedence or backfill” is a nameable, preventable failure. “Resource issues” is not. The purpose of naming it is to make the next charter harder to write the same way.
What the artifacts should have said
A charter clause that survives contact with reality looks like this:
“The Data Migration Lead will be assigned at 20% allocation, defined as Tuesday and Thursday mornings, 08:00–12:00, excluding the first week of each month. The IT Operations Manager confirms that incident response and audit remediation will be covered by the existing on-call rotation and the temporary contractor engaged for Q1–Q2. When project and operational priorities conflict, the Program Director has precedence for the duration of the project. Unresolved conflicts escalate to the Steering Committee within five working days.”
That clause is longer. It is also true. The difference between the two clauses is the difference between a plan and a wish.
FAQ
Is fractional allocation ever workable?
Yes, under specific conditions: the person has slack in their existing role, the pattern is predictable, precedence is defined, and the allocation is tracked against actual hours. Without those conditions, the percentage is a planning artifact with no operational meaning.
What if the line manager refuses to commit to a pattern?
Then the allocation is not available. The project should plan without it or escalate the refusal to the governance body. Planning against an uncommitted allocation transfers the risk to the project schedule, where it will surface as a delay.
How do I know if a 20% allocation is real?
Ask three questions: What is the pattern? What is removed to create the capacity? Who has precedence when priorities conflict? If any answer is “we’ll figure it out,” the allocation is not real.
Does this apply to part-time staff or contractors?
The mechanics are similar but the constraints differ. A contractor with a defined contract and no operational role may have genuine availability. The failure mode described here applies when the fractional assignment is additive to an existing role with its own demands.
Closing note
The twenty percent allocation is not a lie told by any individual. It is a fiction produced by a system that rewards precise-looking numbers and punishes unresolved conflicts. The number is entered, approved, and scheduled. The conflict is deferred. The project absorbs the variance until it cannot.
The procedural close is not a guarantee of success. It is a way to make the fiction visible before it becomes a post-mortem. That is the most an investigator can offer.