Technical craft

Technical Decision-Making

4 lessons · ~24 min total

Decide how to decide

~6 min

Match the decision process to the stakes: one-way doors get rigor, two-way doors get speed.

Key takeaways

Classify reversibility before choosing how carefully to decide: rigor for one-way doors, speed for two-way doors.
Most decisions are two-way doors wearing one-way-door anxiety; the implementing engineer can just pick.
One-way doors hide where dependents accrete: identifiers, public contracts, data shapes.

~5 min

Capture context, options, and rationale in a lightweight record so decisions survive turnover and stop being relitigated.

Key takeaways

Undocumented decisions get relitigated or silently worked around; a one-page record stops both.
Write the record at decision time: context, options, choice, and the consequences accepted knowingly.
The record's key line is the trade-off plus its expiry condition, so reopening needs an expired premise, not just taste.

~6 min

Evaluate third-party against in-house on total cost of ownership, not engineering pride.

Key takeaways

Compare decades, not the build estimate against the invoice: the ownership tail dwarfs version one.
Build what customers choose you for; buy what they merely expect to work.
The number always missing from the build pitch is the permanent fraction of an engineer that owns the system forever.

~7 min

Spot sunk-cost momentum and unwind a failing technical decision cleanly.

Key takeaways

Sunk cost is excluded by rule: only the road ahead versus the alternatives ahead belongs in the decision.
Design exits at entry: written kill criteria, reversible increments, the old path alive until the new one carries load.
Make stopping creditable: harvest what shipped, credit the effort, and record what changed.

Try a sample check

A sample from the Strategy & Prioritization track. The real checks unlock with an account.

Strategy & Prioritization

Your first lesson takes five minutes, and you can do it right now without an account.