Home›Blog›Physical percent complete: the work done, not the time spent
FundamentalsBeginner
Published September 12, 2026 · By Vincent KENNEL
Physical percent complete: the work done, not the time spent
A work package reported at 60 percent can be three different figures, all of them correct. What progress actually measures, how the crediting rule is chosen, and why the first written rule got it backwards.
In brief
Physical percent complete is the percentage of an activity's work content that has been achieved (APM Glossary). It is neither the time elapsed nor the hours expended. On a work package already under way it is not observed but credited, under a measurement rule agreed before work starts (ISO 21508:2018, 5.6).
A work package is reported at 60 percent. The scheduler reads a duration, the work package owner reads what has been consumed, the project manager takes the number for work done. All three readings can be correct, and none measures the same thing. The GAO Schedule Assessment Guide records activities that had used up 80 percent of their scheduled time for 10 percent of the work accomplished (GAO-16-89G, 2015, p. 123), and the NDIA planning guide a detailed design task at 75 percent of its time for 40 percent of the scope (NDIA Planning & Scheduling Excellence Guide, version 4.0, 2019, p. 176). Neither gap is a data entry error.
Figure 1: three measures of one work package, only one is progress
What progress measures, and what it does not
Three names for one family, two entries for one term
The vocabulary is not settled, and that is the first obstacle. The APM Glossary is the only reference work in our set to carry two separate entries: percent complete, the completion status of a partially completed activity, and physical percent complete, the percentage of the work content of an activity that has been achieved. The first says how far along something looks. The second says how much of the work exists.
The family of methods that produces those figures is named differently by each institution that writes it down. The NDIA calls it the earned value technique (Master Definitions List for IPMD Guides, revision of 18 April 2022, p. 12), NASA the performance measurement technique (NASA Schedule Management Handbook, Revision 2, 2024, p. 292), the APM performance measurement techniques, and the GAO titles its own table Seven Measures of Effort (GAO-16-89G, Table 8, p. 170). ISO 21508:2018, which covers earned value management, uses the phrase earned value technique without naming a single one, and states that it does not specify any particular objective measure to be used.
Hence a point worth holding on to. Physical percent complete is a narrow term: it names one technique, or the software field that carries it. The umbrella is performance measurement techniques, or earned value techniques. Two practitioners can agree on the words and still be discussing different objects.
Three quantities on the same work package
On one work package, at one date, three quantities coexist, and none is derived from the others. The first is time elapsed, the second hours expended, the third work accomplished. Only the third is progress.
The PMI states one half without hedging: fifty hours of effort expended on a work package planned for one hundred hours does not mean that 50 percent of the work has been accomplished (Practice Standard for Earned Value Management, 2nd edition, 2011, section 7.3.1.3, p. 38-39). The GAO states the other: an update applied to an activity duration denotes the passage of time from the start date, not actual work accomplished (GAO-16-89G, p. 123). Contract project control keeps the two apart structurally, maintaining two ratios separately, one setting what has been accomplished against what was planned, the other what has been consumed against the same plan (DCMA Earned Value Management System Program Analysis Pamphlet, 2012, p. 16). If both measured the same thing, one would be redundant.
The tool arbitrates nothing, and adds a fourth ambiguity of its own. Scheduling systems carry several Percent Complete fields at once, expressing percentages of duration, of work in hours, of earned value and of scope (NDIA PASEG, p. 108), and the field shown by default in most of them is a percent of duration, which may or may not be related to the physical work percent complete (NASA Schedule Management Handbook, Revision 2, p. 288). Whoever has not chosen a field has chosen duration without knowing it.
What the confusion costs
Where status is reported as a percentage of the hours used, a job estimated at forty hours with twenty hours spent on it is reported at 50 percent. That is one of the ways projects reach 90 percent so quickly, while the last 10 percent takes another 90 percent of the time (Lois Zells, in The AMA Handbook of Project Management, first edition, 1993, chapter 30). The pattern has a name and a measured shape: a project reaches about 90 percent completion on schedule, then stalls, and finishes in about twice the originally projected duration (Ford and Sterman, Overcoming the 90% Syndrome, 2003, p. 178). No account of behaviour is needed to produce it. Measuring the hours consumed and calling the result progress is enough.
The percentage shown by default in a scheduling tool is not progress. It is a percent of duration (NASA Schedule Management Handbook, Revision 2, p. 288). Entering a percentage complete is not recommended either, because the software adjusts the remaining duration to yield the figure entered, and task managers confuse inputs with outputs, assuming that seven days worked on a ten day activity means 70 percent finished (GAO-16-89G, p. 123).
The rules, and what each one credits
On a running work package, progress is not observed
Between the start and the finish of a work package, partial progress is not something anyone sees. What is observed is a start and a finish. In between, a convention applies.
ISO 21508:2018 puts two constraints on that convention: the measure should be agreed upon before each work package commences and should not be changed once work has started, and one measure of performance only should be used per work package (clauses 5.6 d and 5.6 g). The PMI says the same about timing and adds the honest part: the methods are selected during project planning, prior to commencing the work, and there is not one single best way to measure progress (Practice Standard for Earned Value Management, 2nd edition, 2011, section 7.3, p. 36).
One finding explains a good deal of the disagreement in practice: the list of these rules lives in no standard. It appears neither in ISO 21508:2018, which declines to name any measure, nor in the contractual criteria that made the measure auditable in 1977. It lives in glossaries and audit guides, the closed list of seven techniques in the NDIA Master Definitions List being the clearest of them (p. 12).
The four rules
0/100. Nothing is credited before completion, everything at completion. It suits a work package that starts and finishes inside a single reporting period, and is commonly used for the receipt of materials (GAO-20-195G, March 2020, Table 21, p. 231).
50/50. Half is credited at the start and half at completion, regardless of how much work has actually been accomplished in between (Practice Standard for Earned Value Management, 1st edition, 2005, p. 10). The PMI names the consequence itself: real progress is invisible during execution and can give a false sense of accomplishment (Practice Standard for Earned Value Management, 2nd edition, 2011, Table 7-2, p. 42).
Weighted milestones. The work package is cut into segments, each ending in an observable milestone worth an agreed share of the whole (Practice Standard for Earned Value Management, 2nd edition, Table 7-2, p. 42). What makes the rule usable is at least one interim milestone per reporting period (GAO-20-195G, Table 21, p. 231).
Units complete. Units produced are counted against units planned, which is why the rule belongs to construction and to manufacturing (GAO-16-89G, Table 8, p. 170).
Seven such techniques are documented. Four are given here; the remaining three sit above the level of this article.
Choosing the rule, and the residual case
The choice is deduced, not debated
Two parameters govern, and only two: the duration of the work package relative to the reporting period, and whether or not the work is measurable (Practice Standard for Earned Value Management, 2nd edition, Table 7-1, p. 42).
The thresholds are written down. The 0/100 technique suits only a work package planned to start and end within one reporting period, and fixed formula in general is limited to two (GAO-20-195G, p. 231). Beyond that, interim and completion milestones scheduled no more than one month apart are what allow the simple techniques to be applied at all; without them, performance may be misstated, producing artificial variances that then have to be reported and explained (NDIA EVMS Guideline Scalability Guide, Revision 2, 2020, p. 14).
Figure 2: the crediting rule is read off two entries
The residual case, and two guardrails
One cell remains: work that is long and cannot be measured. It has a name, level of effort, and the text that made the measure auditable treats it as a residue rather than an option. Only effort that cannot be identified as discrete, short-span work packages or as apportioned effort is classed as level of effort (DoDI 7000.2, edition of 10 June 1977, criterion 3.b(8)); apportioned effort, the neighbouring cell, is defined in the same text as effort not readily divisible into short-span work packages but related in direct proportion to measured effort (section 2.e). On level of effort, progress follows the calendar by construction and the schedule variance is nil whatever happens (DoD Earned Value Management System Interpretation Guide, 2019, Guideline 12, p. 32). Classing measurable work as level of effort amounts to guaranteeing a progress figure that matches the plan.
Two guardrails close the section. A measurement base only generally indicative of some progress, such as milestones of equal value not related to specific work, is not acceptable (AFMC Cost/Schedule Management Guide, draft, version G, 15 May 1995, p. 3-9): the decorative milestone is ruled out. And outside physical production the principle does not change, since a countable can be constructed and credit taken in binary at the level of the unit (An Industry Practice Guide for Agile on Earned Value Management Programs, version 1.3, 26 May 2019, section 3.3, p. 18).
Where these rules come from
1919, the fact is measured and nothing is valued
A printed form, R-91-Eng., for the week ending 28 January 1919 on the Philadelphia base works, computes the progress of a pile driving item in physical units against the plan: 20,594 piles driven based on 22,462 total piles as per plans, which is 91.6 percent, with the weekly progress computed separately and the words computation checked entered on the form (US Army, A Manual of Planning and Progress for Construction Operations, 1920). A report of 2 September 1919 in the same volume records heating at 71 percent, plumbing at 76 percent, sprinklers at 60 percent and electrical wiring at 36 percent.
The units complete rule is there, whole, forty years before anyone named it. And nothing on those forms is valued. Piles are counted.
March 1963, the first written rule, and its defect
The supplement of March 1963 to the DoD and NASA PERT COST guide defines the value of the work performed to date, and lays down that where a work package is in process, the part applying to completed work is approximated by the ratio of what has been consumed to the latest revised estimate for that work package (Supplement No. 1 to DoD and NASA Guide, PERT COST, Output Reports, March 1963, p. 5). The definition is repeated word for word in the USAF manual of December 1963 (USAF PERT Volume III, PERT Cost System Description Manual, advance copy for AFSC implementation, December 1963, p. VII-6 and B-9). The first normalised device for measuring progress therefore derived progress from consumption, which is precisely what the preceding sections establish should not be measured.
The correction is contemporary with the defect. That same 1963 manual attributes to the guide of June 1962 the recommendation of a work package capped in size and in duration, three months (p. III-4, note 1). Miller gives the reason: the accuracy of the calculation is a function of the length of the work packages, and the shorter the work package, generally speaking, the greater the accuracy (Schedule, Cost, and Profit Control with PERT, 1963, p. 107). He also states what a long activity cannot support, namely that one cannot talk about a mechanical design activity as being 40 or 50 percent complete, and that the actual point in time constraining the next step has to be isolated instead (p. 36). The short work package and the observable milestone are not late refinements: they were written alongside the defect they correct.
1977, the measure becomes auditable
DoDI 7000.2 requires that physical products, milestones, technical performance goals or other indicators be identified and used to measure output (criterion 3.b(2)). Its definition of a work package holds three consecutive clauses saying the same thing from three angles: scheduled dates with interim milestones representative of physical accomplishment, an assigned value expressed in measurable units, and a duration limited to a relatively short span or subdivided by discrete value milestones so that objective measurement is possible (section 2.kk, clauses 4, 5 and 6). Driessnack gives the intent in one line: performance is measured at the level where the work is accomplished, in a very straightforward, pragmatic way, not a subjective one (Interview of Lt Gen Hans H. Driessnack, 1987, p. 141).
What was a computational convenience in 1963 had become an auditable obligation by 1977. The measurement rule stopped being a matter of team preference.
In short, three questions before a work package starts
Progress is not a number produced at the end of the month. It is a convention chosen before the work begins, and it is chosen once.
Three questions settle it, and all three can be answered before anyone starts working. Does this work package fit inside one reporting period, or does it need interim milestones to stay readable? Does the work produce something countable, or only observable steps? And is the rule written down, known to the people doing the work, and single?
A work package whose third question has no answer will still report a percentage every month. It will simply be a percentage of something nobody chose.
Stop guessing. See the real impact.
Frequently asked questions
Q.Does a progress figure say whether the project is late?
No. It says what share of the work exists, not whether that share is the one planned for this date, and not what it has cost. Comparing progress against the baseline is a separate operation, and putting a figure in money on it is another one again.
Q.Can reported progress go down?
It can. In simulation of a real project, the perceived fraction complete falls over a ten week period as testing reveals errors other phases had not reported (Ford and Sterman, The Liar's Club, September 2003, p. 214). A figure that recedes is a deferred correction, not a measurement.
References
APM - APM Glossary - 2026-08-19
Concurrent Engineering - David N. Ford, John D. Sterman - Overcoming the 90% Syndrome - Iteration Management in Concurrent Development Projects - 2003
Construction Division, U.S. Army - Earle B. Morden, Charles Penrose - A Manual of Planning and Progress for Construction Operations - 1920
DoD - Cost/Schedule Management Guide - Draft, Version G, 15 May 1995
DoD - DCMA Earned Value Management System Program Analysis Pamphlet - 2012
DoD - DoDI 7000.2 - Performance Measurement for Selected Acquisitions - Edition of 10 June 1977
DoD - Interview of Lt Gen Hans H. Driessnack - 1987-10-20
DoD - USAF PERT Volume III - PERT Cost System Description Manual - Advance Copy for AFSC Implementation, December 1963
DoD - DoD Earned Value Management System Interpretation Guide (EVMSIG) - 2019
GAO - GAO-16-89G - Schedule Assessment Guide - Best Practices for Project Schedules - 2015
GAO - GAO-20-195G - Cost Estimating and Assessment Guide - Best Practices for Developing and Managing Program Costs - March 2020
ISO - ISO 21508:2018 - Earned value management in project and programme management - 2018
NASA - Schedule Management Handbook - Revision 2, in force from 15 March 2024
NDIA - An Industry Practice Guide for Agile on Earned Value Management Programs - Version 1.3, 26 May 2019
NDIA - Master Definitions List for IPMD Guides - Revision of 18 April 2022
