Lunch & Learn series on From Code to Team (Second Edition), Chapter 9. 20 minutes: a story before the first slide (under 4), 12 talk, 4 discussion. Verbatim script; timing marks are the running clock. Slide numbers match the deck.
Room: mixed — leads, aspiring leads, and engineers who never intend to lead. Everyone in this room is on a project right now. Ask them how far through it is and you will get 3 different numbers.
0:00Told with the title slide up and nothing else. Under 4 minutes. Not on any slide. Presenter's voice; swap in your own, keep the shape.
"When can the kids sleep in it?"
That is the question my wife asked, standing in the backyard in March, looking up at a treehouse I had been building since the previous October. And I want you to notice that I did not have an answer — and that I had, that same week, told 3 different people how far along it was.
I told my brother it was 90% done, because the platform, the walls, and the roof were up, and that is most of a treehouse. I told the neighbor it was about 60%, because the ladder was temporary, the railing was 2 planks and a promise, and the window I had cut was covered in a bin bag. And I told my father — who builds things — that we were "basically in the back half now," which he correctly heard as I have no idea.
Three numbers. All honest. None of them the answer to her question, because her question had a definition in it and mine did not. The kids can sleep in it means a door that shuts, a railing a 6-year-old cannot fall through, a floor that does not let water in, and a ladder I would let my mother climb. I had not written that down. I had a picture in my head — a good picture — and I had assumed the picture traveled with the word treehouse.
So we wrote it down. On the back of an envelope, on the kitchen table: the kids sleep in it, the night after it rains, and nothing is wet in the morning. One sentence. Past tense. And the moment it existed, 3 things happened. First, I could count: 4 things left, not a percentage. Second, the thing I had been quietly avoiding — the railing, because railings are boring — turned out to be on the critical path, and the thing I had been enjoying — a rope pulley for sending up snacks — turned out not to be on the list at all. And third, 2 weeks later, when the neighbor leaned over the fence and said, "You should put a slide on it," I had an answer that was not "we'll look at it." The answer was: "Sure. A slide is a weekend. Which weekend do you want — the one where the railing goes on, or the one after the kids have slept in it?"
He chose the one after. The kids slept in it in May. It rained. Nothing was wet.
The pulley is still not built. It was never in scope, and the only reason I know that is that scope was, for once, a sentence on an envelope instead of a picture in my head.
That is what today is about: why 3 honest people give you 3 different percentages, what it takes to write it down so that nobody can argue with it — and why the slide always arrives in a corridor, 4 minutes before you have to answer.
3:30Ask a room how far through a project is, and count the answers. I have run that experiment accidentally on migrations, platform rebuilds, and integration programs across 3 companies, and the shape is always the same: 3 people answer, the 3 answers are meaningfully different, and not one of them is lying. Somebody says 90 percent, counting the things that have been built. Somebody says 60, counting the things built minus the ones built and then quietly rolled back. And somebody says, "we are basically in the back half now," which is a report on a feeling, delivered accurately — and the feeling is that we have been doing this for a while.
The first few times I heard that as a communication problem and went away to fix the reporting. It is not. Three honest people cannot produce 3 different percentages unless the denominator is missing — and it is missing because nobody wrote down what finished looked like. A lead with a clear picture in his own head assumed the clarity traveled with the words. What traveled was a direction, and a direction is a thing a dozen people can each complete differently in their imagination while agreeing enthusiastically in the same meeting.
5:00A project is not a sprint and not a quarter. It has a shape, an end state, and a landing, and it will outlive at least one re-org and one person you were counting on. The trap catches good engineers hardest: handed a project, you reach for the artifacts — Gantt, RAID log, RACI — because they are visible and produce the sensation of control. Most are ceremony. The parts that determine whether software lands are conversational: agreeing what done means, finding the frightening thing while there are still options, knowing in progress from stuck, and negotiating the trade out loud instead of absorbing it.
For the engineers who never intend to lead: most of this session is about the conversations around the code, and you are in half of them. Knowing what a definition of done looks like, and how to say "I don't think this is going to land" in week 4, makes you the engineer every lead wants on the critical path.