The Last Planner System III: weekly work plans, PPC and learning

The weekly work plan is where the Last Planner System becomes a set of promises. Each week, the last planners, usually trade forepersons, choose tasks from the workable backlog and commit to completing them in the following week. At the end of the week, the team records which commitments were kept and, for each one that was missed, the reason. The fraction of commitments kept, percent plan complete (PPC), is the system's best-known metric. But PPC on its own is just a number; the value comes from the learning loop that asks why commitments fail and changes the system so they fail less often. This lesson covers how to make good commitments, how to run the weekly planning meeting and daily huddles, how to calculate and interpret PPC and how to turn reasons for misses into improvement.

What makes a good commitment

A commitment in the weekly work plan is a promise by a named person to complete a defined piece of work by a stated day. Ballard proposed five quality criteria for assignments. They are worth checking for every line in the plan, because vague or unsound assignments make PPC meaningless.

Commitments should be made in the language of promises: 'I will finish rough-in in 501 to 503 by Thursday' rather than 'we should be able to get to some of floor 5'. A last planner who cannot make that promise, because the crew is short or a constraint is uncertain, should say so. A smaller set of reliable promises is better for everyone downstream than a long list of hopeful ones.

The weekly work planning meeting

The weekly work planning meeting usually takes 45 to 90 minutes, near the end of the week, with all the trade forepersons who will work on site the following week. It reviews the past week, then builds the next week's plan from the workable backlog. Many teams combine it with the look-ahead meeting from lesson 5, taking the look-ahead items second.

The meeting should be held in a room with the plan visible to everyone, on a wall or large screen, and it should be run by the people doing the work rather than by an office scheduler. Commitments are made by forepersons in front of their peers, which makes them more considered and more reliable than commitments made privately to a superintendent.

Daily huddles

Between weekly meetings, a short daily huddle, often 10 to 15 minutes at the start of the day, keeps the plan on track. Forepersons say what their crews will do today, confirm hand-offs and raise any problem that threatens a commitment, so that it can be solved before it causes a miss. The huddle is not a status report to management; it is a coordination meeting among peers. Problems that need more time are taken offline with the people involved. Many sites combine the daily huddle with the daily safety talk, which helps make safety part of planning rather than a separate topic.

Percent plan complete

PPC is the number of commitments completed as promised divided by the total number of commitments made for the week. A task counts as complete only if it was entirely finished; partial completion counts as a miss. This binary rule can seem harsh, but it is deliberate: PPC measures the reliability of promises, not the amount of work done, and a downstream trade that receives 90 % of an area cannot start on it.

Research by Ballard and others found that projects starting the Last Planner System typically have PPC around 50 %, and that PPC rises to 70 % or 80 % as teams learn to make sound commitments and remove the causes of failure. Higher PPC is associated with better productivity, because crews spend more time on work that is ready. Very high PPC, consistently above 90 %, can be a warning sign: the team may be committing only to easy work, padding its plan, or counting partial work as complete.

# PPC trend and rolling average on Willowmere Residences, weeks 1 to 12 (illustrative)
committed = [42, 45, 44, 48, 50, 47, 52, 51, 53, 55, 54, 56]
completed = [23, 26, 28, 31, 35, 33, 38, 39, 42, 44, 44, 46]
ppc = [c / n for c, n in zip(completed, committed)]
for w, p in enumerate(ppc, 1):
    roll = sum(ppc[max(0, w - 4):w]) / min(w, 4)
    bar = "#" * round(p * 40)
    print(f"week {w:>2}: PPC {p:5.1%}  4-week avg {roll:5.1%}  {bar}")
print(f"Overall: {sum(completed)} of {sum(committed)} commitments kept = {sum(completed) / sum(committed):.1%}")

week  1: PPC 54.8%  4-week avg 54.8%  ######################
week  2: PPC 57.8%  4-week avg 56.3%  #######################
week  3: PPC 63.6%  4-week avg 58.7%  #########################
week  4: PPC 64.6%  4-week avg 60.2%  ##########################
week  5: PPC 70.0%  4-week avg 64.0%  ############################
week  6: PPC 70.2%  4-week avg 67.1%  ############################
week  7: PPC 73.1%  4-week avg 69.5%  #############################
week  8: PPC 76.5%  4-week avg 72.4%  ###############################
week  9: PPC 79.2%  4-week avg 74.8%  ################################
week 10: PPC 80.0%  4-week avg 77.2%  ################################
week 11: PPC 81.5%  4-week avg 79.3%  #################################
week 12: PPC 82.1%  4-week avg 80.7%  #################################
Overall: 429 of 597 commitments kept = 71.9%

Reasons for non-completion

Every missed commitment is given a reason, chosen from a standard list so that reasons can be counted and compared over time. A typical list follows the seven preconditions from lesson 5, plus categories for planning errors (the task was overcommitted or poorly defined), changes in priority and weather. The reasons are recorded by the last planner who made the commitment, in the meeting, honestly and without blame.

# Pareto of reasons for missed commitments, weeks 1 to 12 (illustrative)
from collections import Counter
misses = (["prerequisite work"] * 41 + ["overcommitment"] * 22 + ["information"] * 17 +
          ["materials"] * 15 + ["crew size"] * 13 + ["external (inspection)"] * 11 +
          ["equipment"] * 6 + ["space"] * 5 + ["weather"] * 4 + ["priority change"] * 3)
c = Counter(misses)
total = sum(c.values())
cum = 0
for reason, n in c.most_common():
    cum += n
    print(f"{reason:<24}{n:>4}  {n / total:6.1%}  cumulative {cum / total:6.1%}")

prerequisite work         41   29.9%  cumulative  29.9%
overcommitment            22   16.1%  cumulative  46.0%
information               17   12.4%  cumulative  58.4%
materials                 15   10.9%  cumulative  69.3%
crew size                 13    9.5%  cumulative  78.8%
external (inspection)     11    8.0%  cumulative  86.9%
equipment                  6    4.4%  cumulative  91.2%
space                      5    3.6%  cumulative  94.9%
weather                    4    2.9%  cumulative  97.8%
priority change            3    2.2%  cumulative 100.0%

The Pareto ranking shows where to focus. On Willowmere, prerequisite work and overcommitment together account for close to half of all misses. Prerequisite work misses mean that hand-offs between trades are unreliable; the team should look at how areas are released and accepted, perhaps with a simple hand-off checklist signed by both trades. Overcommitment means forepersons are promising more than their crews can deliver; the response is to size commitments to demonstrated capacity, using each trade's recent completion record.

From reasons to root causes

A reason category is a symptom, not a root cause. 'Prerequisite work' tells the team that the previous trade had not finished, but not why. The learning step asks why repeatedly, often using the five whys technique covered in lesson 11. For one miss on Willowmere: the board crew could not start zone 4B (why?) because insulation was not complete (why?) because the insulator had been moved to another floor (why?) because the in-wall inspection on 4B was late (why?) because the inspection request was made the day before rather than three days ahead as the municipality requires (why?) because nobody owned inspection booking. The countermeasure is not to blame the insulator but to add inspection booking to the constraint log with a named owner and a lead time.

Using PPC well

PPC is a measure of planning reliability, not of productivity or progress, and it should not be used to rank trades or punish them. A trade that keeps 95 % of very small commitments may be contributing less than a trade that keeps 75 % of challenging but well-planned ones. When PPC is used for blame, people quickly learn to game it by committing only to trivial work or by redefining tasks after the fact. When it is used for learning, it becomes a shared measure of how well the team plans together. Some teams also track PPC by trade, so that each trade can see its own reliability, and milestone variance, which measures whether phase milestones are being met.

Digital tools

Many teams now run the Last Planner System on digital platforms that support pull planning boards, look-ahead plans, constraint logs, commitments and PPC charts, and that connect to the master schedule and to field management software (CON-260). Digital tools make it easier to share plans across a large site and to analyse reasons over time. They do not replace the conversations. The commitment made by a foreperson in front of peers, and the honest discussion of why it was missed, are the heart of the system; a tool that lets people update tasks silently from their phones can undermine that if it replaces the meeting rather than supporting it.