A Refounding Is a Rewrite. The Job Descriptions Are a Spec Nobody Updated.
Watch the video version:
Early in my career, I had a test for whether a report mattered: stop producing it and wait to see if anyone screamed. One report looked like the easiest call I would ever make. It ran monthly, fully automated, to a distribution list on which not one name still worked at the company. I checked the list twice, because the answer seemed too clean, and then I deleted the code that produced it.
Months passed, and nobody screamed. Then January arrived, and finance surfaced. Every year they had been asking one of those departed recipients for twelve months of that report, and they consolidated it by hand into an annual view that had never been named as a deliverable anywhere. They needed it in two days, on a deadline that was theirs and not mine.
I was scared: two days to rebuild something I had personally deleted, for people whose deadline I had no standing to move. The raw data was still there, so I built a summarized year by hand and made the deadline. I had audited the recipients. I never audited the consumers, and it did not occur to me that those were two different lists.
This month Canva told its people it was refounding the company. Product, Growth, and Platform merged into one organization with fewer management layers, and 1,271 people, about 22% of the company, were laid off, reassigned, or relocated. I make no prediction about how it goes for Canva. The word is the interesting part, because every engineer has lived through a refounding. We call it a rewrite.
The case for a refounding sounds right to any leader who has carried an organization that kept fighting itself. Strip out the layers that built up over the years, merge the groups that kept colliding, and design the company you would build if you started today. The old structure was the problem, so a new structure is the fix. Technical leaders know this argument from the other side: propose rewriting a ten-year-old codebase from its spec, and the first question in the room is what the old system does that the spec never mentions. That question rarely gets asked about the org chart.
The org chart's spec is the job description, and it starts going stale the week it is written. A seat accumulates work: every gap someone filled, every favor that turned into a habit, every "can you just handle this" that stuck. Nobody goes back to update the job description. A refounding redesigns from the job descriptions, so everything the seat picked up since has no line in the new design. Nobody decided to drop it; it was never on paper to keep.
That is State Loss: the knowledge lived in the seat instead of the system, and the new structure has nowhere to keep it. The cost never shows up the day the memo goes out. It shows up the way my report did, months later, on someone else's deadline, owned by a person who was never in the room where the new chart was drawn. The Foreman lives for this moment. He treats every reorg as a reason to demolish and rebuild from the foundation, and he counts on nobody having written down what the old structure was holding up.
The leaders who come through a reorg with the work intact don't understand structure better than you do. They read the seats before they redraw them. If you are the one drawing the new chart, you are also the one person positioned to find out what each seat really does before it disappears.
Read the seats, not the job descriptions. The document tells you what a seat was hired to do. Only the person in it, and the people downstream of it, can tell you what it does now. Writing that down is The Persistence Layer at its smallest size: one seat, and the two lists I never compared.
Pick one seat on your team this week. Ask the person in it to list everything they do that isn't in their job description. Then ask the two teams downstream what would break if that seat went quiet for a month. Where the lists don't match, you have found the work your next reorg would delete, and you found it while it can still be written down.
LeadershipOS™ has the full Persistence Layer: how to move what lives in one seat into the system, so a reorg can't reset it: https://TheLeadershipOSBook.com
Twelve capable people, each making a reasonable decision, and the sum is a leader still at the desk at 6:43 pm. Why your team is performing exactly as designed.
