A practice facility with no offseason
The Clippers reached out to Tellus. I never asked how they found us; we were number one on Google in those years, so people found us. What they wanted was a redesign of the practice facility, room by room: the locker rooms, the performance and rehab space, the training and coaching areas, and the player lounge. Ownership and team operations approved the program. Tellus held the whole of it, the design and the general construction, which is a different seat from sitting on the owner's side of the table, and the difference mattered on this job for a reason I will come to.
Where they were is easy to describe and easy to underestimate. An NBA team uses its practice facility every day of the year that it is not on the road. The season runs into the spring, the playoffs run into June, the summer holds workouts, camps and rehab, and camp opens again in September. There is no window long enough to close the building, hand it to a contractor and take it back finished. The rooms we were rebuilding were the rooms the players and the staff were using while we rebuilt them.
What made it a design problem and not just a scheduling problem was the shape of the scope. This was not one big build. It was many small scopes spread across four rooms that shared walls, air, water, power and hallways. A program like that has no single critical path; it has dozens of little ones, and every one of them runs through a room somebody needs tomorrow morning. The people in it were the team's operations and facilities staff, who knew which rooms the players needed on which days; the ownership side, who approved the money and the program; the trades, none of whom could see the whole sequence from where they stood; and us, holding the drawings and the schedule at the same time.
What the client actually needed was not the fastest finish. It was a plan built around what the team needs open on any given day, and a definition of "done" that belonged to the people who use the rooms, not to the people who draw them or build them.
Hold the calendar fixed, let the finish date move
Every renovation schedule I had ever seen was built the same way: the finish date is the constraint and everything else bends around it. That is right for an empty building and wrong for an occupied one, because the thing the client cannot give up is not the date; it is the use of the building. So I turned the schedule around. The practice calendar was fixed, and the finish date and the sequence were the free variables, chosen to protect the calendar.
The obvious alternative was to run it the normal way and manage the conflicts as they came: off-hours work, weekend shutdowns, temporary partitions, a phone call when a room had to close for a day. That is how most occupied renovations run, and it works when the calendar has slack. This one had none. A conflict on an ordinary job becomes a change order; a conflict in this building becomes a practice the team cannot hold, and no change order pays for that.
Three decisions followed from turning the schedule around, and each one had an alternative I chose against.
One room down at a time. The first question I put to team operations was not "how fast can we go" but "how many of these four rooms do you need on any given day". The answer sets how many can be closed at once, and the plan was sized to one. The design rule was three of the four rooms open on any day the team was in the building. Taking two rooms down would have shortened the program; it would also have meant a day, sooner or later, when the players had nowhere to lift or nowhere to dress.
The shared systems set the order, not the Gantt chart. Before a drawing went out for bid we mapped what each room shared with its neighbors: the air handling zones, the electrical panels and the circuits that had to be shut down, the wet walls and drain lines, the exit paths, the fire alarm and sprinkler zones, the structure, the routes dust and noise would travel, and the routes the crews and materials would take to get in. Which rooms could be worked apart, and which had to be worked together, was a fact about the building, not a preference, so the phase order came from that map: the physics of the building set the sequence, and the schedule was written afterward.
Team operations signs "done". The industry's trigger for moving from one phase to the next is substantial completion, which is the designer's judgment that a space can be used for its purpose. On this job Tellus was the designer and the builder both, so substantial completion would have meant us certifying ourselves finished. I moved the call. A phase was done when the team's operations staff signed that they could run a practice in it. Only they could tell the difference between a room that matched the drawings and a room they could use, and only that difference mattered. The decision that actually decided the phase belonged with the person who could make it.
How I came at this one
The question I asked first was how many of the four rooms the team needed on any given day, not how fast the work could finish. That question fit because there was no offseason to hide a shutdown in; the calendar was the constraint and the finish date was the only free variable. Everything else in the plan was the consequence of answering that one question honestly.