Repeat Fire · Transform
Step 6 of 6 · REPEAT

Transform: Improve the System

Assess delivered a verdict. Transform is where you act on it — not by starting the loop over from scratch, but by refining the system itself so the next cycle runs better than this one did. This is the step that turns REPEAT from a repeating checklist into something that actually compounds.

The job of this step
The verdict → a change to the system, not just the next move

Plan answers "what should I do this cycle, given the current gap." Transform answers a level up: "what should change about how I plan, given what I just learned." One is a move. The other is an update to the rules that generate moves.

Not Plan again

If Assess showed your contribution increase closed the gap but took longer than projected, Plan's job next cycle is to size the next contribution correctly. Transform's job is to notice that your return assumption was optimistic and adjust the assumption itself, so every future Plan step starts from a more accurate baseline instead of repeating the same miscalibration.

What actually gets transformed

CategoryWhat it covers
AssumptionsReturn rates, inflation estimates, expected retirement spending — inputs you set once and reuse every cycle.
ProcessCadence that's too infrequent to catch drift, or steps (like Evaluate and Plan) that keep sliding into each other.
StrategyBigger calls — liquidity-first vs. ladder, account mix, which FIRE flavor still fits where your life has gone.

Adapting to new circumstances, not just new numbers

Life changes in ways a spreadsheet doesn't capture on its own — a new job, a health event, a partner's income shifting, a kid, a move to a lower cost-of-living area. None of these show up cleanly in Review's numbers; they show up as context that makes last cycle's assumptions stale in ways Assess alone can't diagnose. Transform is where that context gets folded back into the system deliberately, rather than leaking in unevenly across a dozen small decisions.

Change less than it feels like

The instinct after a bad Assess result is often to overhaul everything — new strategy, new targets, new plan, all at once. Resist that. Transform works best as targeted correction: change the specific assumption or process element that Assess actually implicated, not everything adjacent to it. A loop that gets fully rebuilt every cycle isn't really a loop — it's a series of fresh starts wearing a repeatable framework's name.

The compounding benefit of REPEAT comes from small, accurate corrections applied consistently, not dramatic pivots applied occasionally.

Closing the loop

Once Transform is done, the system feeds directly back into Review — not because you're starting over, but because the next Review will now be measuring against updated assumptions, using a refined process, possibly pointed at a slightly adjusted strategy. Each pass through REPEAT doesn't just move you closer to your number; it makes the next pass sharper than the one before it.

What Transform is not

Transform doesn't execute anything itself — a changed assumption or strategy still has to travel back through Plan and Execute before it does anything in the real world. And it isn't a place to second-guess a verdict Assess already delivered cleanly; if the plan worked as expected, Transform's job isn't to relitigate that, it's to ask whether anything about the surrounding system should change given that it worked.

A worked correction
Assess foundContribution increase closed the gap, but took 40% longer than projected
Root causeReturn assumption of 7% was optimistic; actual blended return ran closer to 5.5%
Transform actionUpdate the return assumption used across every tool, not just this one plan
This is the step where updated assumptions matter everywhere — a revised return-rate assumption should get reflected in Your Number's investment-income projection, the Withdrawal Calculator's growth assumptions, and the Account Optimizer's bridge-year math consistently, not updated in one place and left stale in the others. This is now literal: update the return rate once on My Assumptions and every tool picks it up. Log the prediction vs. actual there during Assess, then use Apply actual here.