Find the Irreversible Parts First

Rank the decision by “can I undo this,” then set the order

The decision I'm making is: [decision]. First pull out the irreversible parts and put them in order. Requirements: 1) Mark each action that can't be undone, with the concrete cost of undoing it; 2) What future options each one locks shut; 3) The cheapest information available before you push it — a pilot, a trial order, a quote, a small rollout; 4) The sequence: which must happen now, which can wait for more information, and the latest date for each; 5) Which single one, if wrong, invalidates the whole decision. Do not judge whether the decision should be made.

How to use

Ask “after this step, can I still change it?” Do the reversible steps first; hold the rest until you know more.

Good for

The decision has been broken into seven or eight actions and the team wants to run them in parallel, but you suspect one of them ends the option to turn back.
You're about to sign a long contract, hire a key role, or publicly commit to a delivery date, and want to know which step belongs last.

Not for

The whole thing is reversible and redoing it is cheap — sequencing is pointless; start doing.
Reversibility is fine and the real problem is money — that's the cost-underestimation check, not this one.

Example

Input

Decision: give up the office, go fully remote, and put the saved rent into a new business line. Rent is ¥960k a year; the new line should pay back in about 18 months.

What you get

Two one-way doors here: terminating the lease and announcing the remote policy. Once the lease is gone, finding comparable space takes 4–6 months plus deposit and build-out; going fully remote locks shut “we'll train juniors by sitting together.” The cheapest information is eight weeks of hybrid before you terminate — watch how fast new hires ramp and how well existing staff collaborate — and it costs little more than a small meeting room. The new line is actually reversible; pilot it small, and don't tie it to the same date as the lease.

Common mistakes

Labeling everything irreversible, so nothing ever gets done — the test is “what would undoing cost in money and time,” not “is there any risk.”
Listing actions without the constraints — budget, contract term, public promises — leaves no way to tell which door is actually shut.

Try tweaking it like this

Swap “sequence” for “a latest date for pushing each door,” turning analysis into a calendar.
Add “if I could keep only one irreversible action, which one,” forcing the real trade-off.