Capstone: implementing lean on a mid-rise project

This capstone brings the course together by following Northgate Builders as it plans and runs lean practices on Willowmere Residences, the invented eight-storey building used throughout. The aim is not to present a perfect project but to show how the pieces fit: how the master schedule, pull planning, the look-ahead and weekly plans, the takt plan, buffers, prefabrication, productivity measurement and continuous improvement work together, what they cost, what they deliver and where they run into trouble. All figures are illustrative. The final sections give a template that students can use to prepare a lean implementation plan for a project of their own.

The project and the starting point

Willowmere Residences has a concrete podium with ground-floor retail, a lobby and amenity space, and a seven-storey residential tower above with twelve suites per floor, 84 suites in all. Northgate Builders holds a $38 million stipulated-price contract with a 20-month schedule. Northgate has used the Last Planner System on two earlier projects, with mixed results: PPC rose to about 70 % on one but the system faded on the other when the superintendent changed. Most of the trade partners on Willowmere have heard of lean but only the drywall and mechanical contractors have used it before. The owner, a rental housing developer, cares most about early occupancy, because each month of earlier completion brings rental income forward, and has offered to discuss a shared-savings arrangement for schedule improvements.

The implementation plan

Northgate's project manager and superintendent prepared a lean implementation plan before mobilization, using the structure below. The plan was short, about six pages, and was reviewed with the owner and with each major trade partner before subcontracts were signed, so that participation in pull planning, look-ahead and weekly planning meetings, takt control and hand-off checklists could be written into the subcontract scopes.

The master schedule and phases

The contract required a CPM schedule, which Northgate built at summary level around fourteen milestones: excavation complete, podium structure complete, tower structure topped out, building enclosed, permanent power, elevators available, each pair of residential floors ready for occupancy and substantial completion. Six phases were defined for pull planning: excavation and foundations, podium structure, tower structure, envelope, residential fit-out and podium fit-out. The residential fit-out phase, the longest and most repetitive, was planned with takt; the others were planned with pull planning and conventional CPM logic.

The residential takt plan

The residential fit-out was planned with the 28 zones and 9 wagons from lesson 7 and a two-day takt, giving 72 working days of flow, plus one buffer takt after the inspection wagon, for 74 working days. Crew sizes were balanced to the takt with each trade partner. Bathroom pods were set floor by floor as the structure rose, so that wagon 7's tile work was limited to kitchens and entries, and corridor racks were installed in a separate off-takt sequence ahead of wagon 2. The takt plan was posted on a large board in the site office and on a zone status board at each floor's entrance.

# Willowmere residential takt plan: duration, floor handovers and crew loading
zones, wagons, takt, buffer_takts = 28, 9, 2, 1
steps = wagons + buffer_takts
duration = (zones + steps - 1) * takt
print(f"Fit-out duration: {duration} working days ({duration / 5:.1f} weeks)")
# A floor (4 zones) is ready for occupancy when its last zone clears the last wagon
for floor in range(2, 9):
    last_zone = (floor - 1) * 4                    # zones 1-4 are level 2, 5-8 level 3, ...
    ready_day = (last_zone - 1 + steps) * takt
    print(f"level {floor} ready on working day {ready_day} (week {ready_day / 5:.1f})")
# Peak crew on the floors when all steps are working at once
crew = {"framing": 6, "mechanical": 7, "plumb + elec": 8, "insp/firestop/insul": 4,
        "board": 8, "tape": 6, "paint + tile": 8, "millwork": 7, "finish trades": 9}
print("Steady-state residential crew:", sum(crew.values()), "workers across", len(crew), "wagons")

Fit-out duration: 74 working days (14.8 weeks)
level 2 ready on working day 26 (week 5.2)
level 3 ready on working day 34 (week 6.8)
level 4 ready on working day 42 (week 8.4)
level 5 ready on working day 50 (week 10.0)
level 6 ready on working day 58 (week 11.6)
level 7 ready on working day 66 (week 13.2)
level 8 ready on working day 74 (week 14.8)
Steady-state residential crew: 63 workers across 9 wagons

The first residential floor is ready for occupancy on working day 26, about ten weeks before the last on day 74, and each subsequent floor follows eight working days later. For a rental building, this phased occupancy is valuable: suites can be leased floor by floor, subject to the municipality's rules for occupancy of a partly completed building, the fire separation and life safety systems on occupied floors and the owner's insurance.

Weeks 1 to 12: getting started

The first months were harder than planned. In the first weekly planning meetings, several forepersons made vague or oversized commitments, and PPC ran at about 55 %. The constraint log grew to more than sixty items, many of them design questions from the podium. The pull planning session for the tower structure surfaced a key problem: the formwork subcontractor's plan assumed a crane availability that conflicted with pod deliveries. Resolving it took two sessions and a change to the crane schedule, but it was found five months before it would have caused a delay. By week 12, PPC had risen to the low 80s, the top reasons for misses were known and being worked on, and forepersons had started raising constraints earlier.

# Dashboard for weeks 1-12 on Willowmere (illustrative data, see lesson 6)
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]
tmr       = [0.58, 0.61, 0.60, 0.66, 0.69, 0.70, 0.72, 0.75, 0.77, 0.78, 0.80, 0.82]
open_constraints = [38, 49, 61, 58, 54, 51, 47, 44, 40, 37, 35, 33]
for w in range(12):
    ppc = completed[w] / committed[w]
    flag = "" if ppc >= 0.75 else "  below target"
    print(f"wk {w + 1:>2}  PPC {ppc:5.1%}  TMR {tmr[w]:4.0%}  open constraints {open_constraints[w]:>2}{flag}")
first4 = sum(completed[:4]) / sum(committed[:4]); last4 = sum(completed[-4:]) / sum(committed[-4:])
print(f"PPC first four weeks {first4:.1%}, last four weeks {last4:.1%}")

wk  1  PPC 54.8%  TMR  58%  open constraints 38  below target
wk  2  PPC 57.8%  TMR  61%  open constraints 49  below target
wk  3  PPC 63.6%  TMR  60%  open constraints 61  below target
wk  4  PPC 64.6%  TMR  66%  open constraints 58  below target
wk  5  PPC 70.0%  TMR  69%  open constraints 54  below target
wk  6  PPC 70.2%  TMR  70%  open constraints 51  below target
wk  7  PPC 73.1%  TMR  72%  open constraints 47  below target
wk  8  PPC 76.5%  TMR  75%  open constraints 44
wk  9  PPC 79.2%  TMR  77%  open constraints 40
wk 10  PPC 80.0%  TMR  78%  open constraints 37
wk 11  PPC 81.5%  TMR  80%  open constraints 35
wk 12  PPC 82.1%  TMR  82%  open constraints 33
PPC first four weeks 60.3%, last four weeks 80.7%

The fit-out: holding the takt

The residential fit-out started on schedule. The first-run studies for board hanging and pod connection paid off: by the third zone, both wagons were finishing comfortably within the takt. The weakest hand-off was, as expected, the inspection wagon, where the municipality's inspection availability varied; the buffer takt absorbed most of the variability. Two events tested the plan. In week 31, a delivery of kitchen cabinets arrived with the wrong finish for floors 4 and 5. The millwork wagon kept its rhythm by installing on floors 6 and 7 first, which the constraint log had flagged as a possible fallback, and the replacement cabinets were installed off takt two weeks later. In week 36, the drywall contractor lost three workers to another project, and the board wagon fell behind for four days. The team used the buffer takt, the taping crew helped with board for two days and the drywall contractor's owner, who attended the weekly meeting, restored the crew within the week.

What it cost and what it delivered

Lean implementation is not free. Northgate's costs included training, a part-time lean lead, facilitation of pull planning sessions, the time of forepersons in meetings, the first-run studies and the second hoist. Benefits included lower labour cost, lower general conditions from the shorter schedule and, for the owner, earlier rental income from phased occupancy.

# Lean implementation: costs and benefits on Willowmere (illustrative)
costs = {"training (forepersons, 2 days x 30 people)": 30 * 2 * 8 * 85,
         "part-time lean lead (60 weeks x 16 h)": 60 * 16 * 110,
         "pull planning facilitation (6 sessions)": 6 * 4_500,
         "weekly meetings: forepersons' time (60 weeks x 12 people x 1.5 h)": 60 * 12 * 1.5 * 85,
         "first-run studies": 3 * 24 * 85,
         "second hoist (rental and operator, net)": 55_000}
benefits = {"field hours saved (4,800 h x $85)": 4_800 * 85,
            "general conditions (5 weeks x $23,000)": 5 * 23_000}
owner_benefit = 6 * 84 * 2_300 * 0.5      # avg. 6 weeks earlier, 84 suites, $2,300/month rent, ~half a month each
total_cost = sum(costs.values()); total_benefit = sum(benefits.values())
for k, v in costs.items():
    print(f"cost     {k:<64}${v:>10,.0f}")
for k, v in benefits.items():
    print(f"benefit  {k:<64}${v:>10,.0f}")
print(f"Contractor net: ${total_benefit - total_cost:,.0f}")
print(f"Owner's earlier rental income (rough): ${owner_benefit:,.0f}")

cost     training (forepersons, 2 days x 30 people)                      $    40,800
cost     part-time lean lead (60 weeks x 16 h)                           $   105,600
cost     pull planning facilitation (6 sessions)                         $    27,000
cost     weekly meetings: forepersons' time (60 weeks x 12 people x 1.5 h)$    91,800
cost     first-run studies                                               $     6,120
cost     second hoist (rental and operator, net)                         $    55,000
benefit  field hours saved (4,800 h x $85)                               $   408,000
benefit  general conditions (5 weeks x $23,000)                          $   115,000
Contractor net: $196,680
Owner's earlier rental income (rough): $579,600

On these figures, the contractor's benefits exceed its costs by a modest margin, and the owner gains considerably more from earlier occupancy. That imbalance is common: the largest benefits of lean planning often go to the owner through a shorter schedule, which is why shared-savings arrangements and collaborative contracts (CON-340) help align incentives. Northgate's agreement with the owner shared a portion of the schedule benefit, which made the investment clearly worthwhile for Northgate and its trade partners.

Lessons learned

Northgate's closing lessons-learned session, run as a large plus-delta with all trade partners, produced a short list of points that it adopted as company standards. Several echo the research literature on lean implementation.

Template: your lean implementation plan

Students completing this course should be able to prepare a lean implementation plan for a real or realistic project. The template below follows the structure Northgate used. A good plan is short, specific to the project, agreed with the owner and trade partners and backed by named people, dates and measures.