23. Hohmann Reorganizations
We did not wait for change on Mars; we invited it, scheduled it, and lived by it.
If you want a clean definition of early Martian management, it is this: we treated reorganization as a craft, not a crisis. Every new Hohmann window — those slow, stately openings in the orbital clock — brought a new cohort of ships, parts, fresh faces, and a reason to tune the machine. People think that sounds exhausting. It was, sometimes. But it was also merciful. The worst reorganizations are the ones that arrive as gossip. Ours arrived as pre-planned timetables.
I was there when the first of these took shape, a slender document named with un-Martian levity: the Reorg Trigger Table. The title undersold it. What the Table actually did was take the heat out of politics and pour it into arithmetic.
We stopped pretending that a settlement of a hundred could keep the same bones at five hundred, or that five hundred could endure the same spans and ladders at two thousand. We gave ourselves permission to change — and a metronome to change by.
And this was a critical beat of the MarsX plan that laid out that Founding Century.
♦
There is a rhythm to interplanetary life that feels inevitable once you have lived it awhile. Hohmann windows set the beat: departures, arrivals, months of drift in between. Seven to eight months aloft for crews and cargo, punctuated by those operatic weeks when hulls flare and docks fill. We aligned our reorgs to that music.
This was not romantic. It was logistics.
Between windows, you know who intends to leave and who intends to come. The rosters harden early. Skills maps sharpen. Mentors arrange handovers. Spares, benches, and secondments take on purpose. By the time descent burns begin, the org’s next shape is no longer hypothetical; it is latent — waiting for bodies to make it real.
We learned quickly that ad-hoc reshuffles in the middle of a window were pure entropy. The Mars team is small, even at four thousand; every unexpected rewire dumped sludge into the gears: permits in limbo, SIMOPS broken, safety envelopes confused. So we wired change to the windows. If something truly urgent emerged between — alarm floods, safety regressions — we made the minimum surgical tweak and queued the structural change for the next arrival wave. Stability during ‘run’, change during ‘handover.’
And in that cycle, we found a paradoxical calm: change as routine.
♦
A word on doctrine. Our maxim was “triggers, not politics.” You may find it quaint that we had to write that down. But in the early years we remembered Earth far too well. Influence tends to collect around loudness and proximity. Neither is a reliable proxy for truth.
The Trigger Table ended that.
Each trigger was blindingly specific: a metric with an owner, a threshold that had to be crossed, an action that would follow, a budget that unlocked automatically, a review date when we would check if reality matched the plan. Span-of-control above twelve for thirty sols? Split the unit and seat a second single-threaded lead. Permit queue ninety-fifth percentile creeping above five days? Stand up a second Permit Cell and lend it a records clerk. Two concurrent construction pads for more than forty-five sols? Fork Construction into Pad Works and Fit-Out; issue distinct permits and foremen warrants. Alarm flood out of policy for a day, or two rollbacks in a week? Add a Change Captain, freeze elective changes, run hygiene.
No speeches. No war rooms. A trigger fired; an action followed; a ledger entry explained why. If a pillar wanted to defer, they could — but only with a written rationale, a risk plan, and a sunset date. Exceptions expired unless renewed. In a city where nothing rots naturally, even deviations were compostable.
♦
Reorganizations on Mars were never revolutions. I know the Earth press preferred the romance of the Big Bang Rewire — visionaries saving the colony by firing half the org chart and inventing five new pyramids. It makes for good feeds. It also kills people.
Our changes were more musical than surgical. You felt them in spans tightening, in ladders gaining depth, in new 24/7 seats lighting up on MCC, in split interfaces that made boundaries visible where they had blurred. Sometimes it was a kind of federation: a spoke town crossing the readiness index and taking a local charter, with mini-pillars bound to central standards. More commonly it was unglamorous: a kitting focus cell so late kits stopped haunting lifeline jobs, an extra records custodian so Civic could keep audit SLAs without leaning on good will, a second console to handle the night pager load.
A few reorgs were mood shifts rather than maps: retiring a meeting that no longer earned its keep, adding an extra T2 sync during range week, rebalancing the huddle tree so workface was louder than headquarters. Those mattered more than they appeared. Rituals shape attention; attention shapes accidents.
♦
Please do not picture chaos. Imagine instead a checklist.
The reorg lifecycle began with sensing. The Integration Cell kept the score — span health, permit queues, alarm floods, kit lateness, downtime. Nothing heroic. Dashboards with owners. A weekly leaders’ review maintained a Reorg Watchlist. When a metric approached its line, the question was not “who’s at fault?” It was “which pre-written action is about to trigger, and what knock-ons will it cause?” That phrasing matters: shame corrodes; planning lubricates.
About sixty sols before a window, pillars drafted their changes. They attached the triggers. They checked the budgets already tied to those rows. They wrote the role cards. They updated the Target Operating Model diffs so flows, boards, and permits would route correctly on day zero. They prebooked training seats so new leads would be warranted, not just named. This was the most prosaic part, and the most important. A reorg is not a diagram; it is a thousand tiny affordances made ready.
A week or two before arrival, the leaders signed the Reorg Order. Not a proclamation — more a bill of materials for a new shape: scope, effective date, reporting deltas, budgets, warrant updates, TOM diffs. Civic published a short rationale with the trigger evidence. On window day, we executed in two waves. First, we changed the lines: who reports to whom, who sits which seat, which nets belong to which ladder. Second, we changed the tools: boards went live with new queues, MCC displayed the new seats, permits reflowed to duplicated cells, MoC paths re-pointed. Handover kits were already waiting on desks: why this change, what had been tried before, who to call.
Between thirty and ninety sols into the new shape, we measured. If the metrics fell back into band, the reorg ticket closed. If they didn’t, we tweaked. Not a second grand reorg — focused adjustments. A lot of management is resisting the urge to overcorrect.
♦
People always ask about morale. Change can feel like erasure if it arrives without context. We learned to narrate our own evolution. The ICell master board included a short “why we’re changing” line on each reorg item; not propaganda, simply the causal rationale: “Span exceeded policy; second lead seated,” “Queue breach; duplicating permits by class,” “Pad concurrency; splitting foremen.”
It helped that the civic ledger made decisions inspectable. When a trigger fired, any settler could look up the evidence. The extract was redacted only for personal medical data or legal sensitivity. Every action had a sunset review. More than once, a workface crew reminded leaders of their own policy: “This expires next month — are we renewing or reverting?” On Mars, the shop floor could read, and they expected you to keep time.
And something else: our leadership rotation policy interlocked with reorgs on purpose. The Chief Engineer and Chief Operator held two windows each, offset; the Mayor rotated mid-cycle. If a top role changed this window, the reorg scope narrowed to structural basics, with process tweaks deferred. One big move at a time. Continuity by design, not luck.
♦
Classically, the matrix — the cross-functional weave of Technical, Operations, and Civic — makes reorgs dangerous. Tug a thread on one pillar, and the other two tighten or slacken in surprising places. We avoided knots by keeping our envelopes sovereign. A reorg could add a Change Captain, but it could not grant Operations the right to alter lifeline set-points. It could duplicate Permit Cells, but it could not convert Civic into a delivery arm. It could fork Construction, but the Tech engineers still approved pressure envelopes.
We also refused to let matrix charters metastasize. On Mars, the temporary joint teams that built rings, refit reactors, and stood up depots were powers of speed — but they were time-boxed by charter. A reorg might spawn three or four new charters as we entered a busy season; it also wrote their sunsets. No shadow orgs. Deliver, hand over, vanish. This discipline prevented “matrix creep” after reorgs, one of the most corrosive anti-patterns we saw on Earth.
♦
Here is a truth outsiders underestimate: most reorg failures do not come from bad diagrams. They come from missing seats.
A console added without night parity is a promise undone at 03:00. A permit cell spun up without a records clerk births backlog. A split without a warrant creates two leaders in name and none in practice. We wrote the antidotes into the Trigger Table: every action listed dependencies — C3 seats, Supply kitting, records staffing—and budgets pre-tied to fire together. When someone tried to cut the “boring” support pieces, the order wouldn’t validate. The organigram would take, or it would not apply. Governance as an electrical interlock.
♦
Were there anti-patterns? Of course. Celebrity reorgs: a leader fond of a shape they had seen on a tour, eager to transplant it whole, triggers be damned. Big-bang rewires that tried to accomplish three windows of evolution in one week. Seat inflation: 24/7 posts established without duty-cycle evidence, because always-on feels like progress. Zombie exceptions: deviations granted under pressure and never sunset, undermining the very table that gave them a name.
We built the antibodies into the system. Civic required a Trigger Table citation for significant changes; no citation, no public rationale, no change. We capped concurrent changes and reserved the right to split waves. We required post-reorg reviews at D+30 and D+60; if a seat wasn’t used, we removed it; if a deviation had no sunset, it died. Not punitive — just honest.
♦
From a distance, this all sounds mechanistic. From the inside, it felt humane. The worst thing you can do to a growing settlement is ask people to endure avoidable ambiguity. Reorgs, done badly, are long winter nights. Done well, they are spring cleaning. Yes, the furniture moves. But you wake knowing where the exits are.
I remember an early window when we added a second lifeline console to protect outage windows during dust storm season. The pager fatigue on nights had crept into red. We had the trigger documented; we had the training prebooked. Still, the console shift leads were nervous. More hands can mean more noise. For a week, it did. Then the alarm flood fell into band, the new Change Captain absorbed the chatter, and the outage success rate stopped wobbling. One of the old hands — calloused, suspicious of all management, polite only to machines — brought the console crew a tray of mediocre cookies. “You could have done this months ago,” he said. “But at least you did it on purpose.”
That last clause — that was our north star.
♦
As headcount rose — from a hundred to five hundred, five hundred to two thousand, drift toward four — we watched our span discipline like the last candle in a storm. A target of one to ten in the field, tighter in EVA/high hazard seats; one to six for night console supervision; deputies warranted and currency maintained. Spans are not a matter of taste; they are the geometry of attention. When the table recorded a span breach for thirty sols, we split. No drama. No defense. Just the physics of care.
We also watched the federation boundary. Outposts earned their charters by readiness, not romance. When a site’s readiness index — the checklist of mini-pillar depth, records hygiene, crisis tether to central — crossed one, they federated. Local BAU authority rose; crisis command stayed centralized. Some argued for earlier autonomy; others for permanent centralism. The table did not care for either faction. It cared whether the conditions were met.
♦
None of this would have worked without our commitment to “one ledger.” The reason Mars could change itself without losing itself is that the record did not move. Permits, MoCs, SIMOPS, charters, reorg orders — everything bound to the same IDs and timebase. If it wasn’t in the ledger, it didn’t happen. Earth learned to love this discipline. So did we. The ledger was not an accounting fetish. It was the cup that held memory between windows.
Because reorgs are also acts of forgetting. You retire a queue. You rename a ladder. You kill a ritual that once saved you. The ledger remembers what you let go of and why. It lets new arrivals find the thread. It prevents the same errors from enjoying a second debut under new stationery.
♦
If this chapter is about anything, it is about the dignity of planned change. Most organizations change either too late or too loudly. We tried to do neither. We listened for the thresholds that predicted harm — fatigue, queue drag, collision risk, decision latency—and we tied them to actions we could afford and explain. We used orbital mechanics as a scaffold so our people were never surprised by the season of turning. And we staggered our top-role rotations so the org never turned all at once.
We did not always get it right. Sometimes we split too soon and paid in overhead. Sometimes we deferred and paid in accidents. But the table kept us honest, and the windows forgave our calibration errors. In fifty windows we grew from a small tribe to a small city. In each, the org changed just enough to survive the next. That was the art.
♦
There is a line we wrote in those years and taped, with more than a little pride, above the ICell board:
Use the window as a metronome, and the Trigger Table as the sheet music. When metrics cross a line, the org changes — because the plan said it would, not because someone shouted loudest.
It reads tidy on paper. It felt better in practice. The beat came predictably. The melody was familiar. And the music — the work — could go on.