Prioritization & Decisions

Tap menu for chapters & tools

Chapter 05

Prioritization & Decisions

Choose fewer things and finish them

20 min read

A team lead’s scarcest resource is not engineering hours - it is attention. Poor prioritization creates thrash: half-finished projects, context switching, and a culture of urgency theater. Strong leads make trade-offs explicit and defensible.

Prioritize with a shared model

  • Impact: user value, revenue risk, reliability, learning.
  • Urgency: real deadlines vs. artificial pressure.
  • Confidence: how sure are we about the problem and solution?
  • Cost of delay: what breaks if we wait two weeks?
  • Team health: will this push people into chronic overtime?

Decision hygiene

Not every decision deserves consensus. Use a lightweight RACI or DACI. Classify decisions as reversible (move fast) vs. irreversible (slow down). Write Architecture Decision Records for choices that will be re-litigated later.

  1. Frame the decision in one sentence.
  2. List constraints and non-goals.
  3. Capture options and discarded alternatives.
  4. Record the decision, owner, and revisit date.
  5. Communicate widely enough that people stop guessing.

Say no without burning trust

“No” is a leadership tool. Soften it with alternatives: later, smaller, different owner, or a clear experiment. Explain the capacity math. Invite the requester to help deprioritize something else if their item must win.

A roadmap with twenty P0s is not ambitious. It is a refusal to lead.

Capacity math people can see

Priority fights end faster when capacity is visible. Estimate in engineer-days, subtract meeting and support tax, then show what fits. If leadership wants more, they must name what drops. Hidden overload is how roadmaps become fiction.

  • Write available days for the next two weeks, not hope.
  • Separate committed work from stretch bets.
  • Track WIP age so thrash shows up early.
  • Re-plan when a P0 arrives; do not pretend time is elastic.

Decision journal for leads

Keep a short personal log: decision, options, what you chose, what you expected. Review monthly. You will spot patterns (speed bias, people-pleasing, over-engineering) faster than any 360 review.

  1. Capture the one-sentence decision.
  2. Note the top risk you accepted.
  3. Set a revisit date.
  4. Write what would change your mind.
Two-way door reminder

If you can reverse the choice in a week with low cost, decide today. If the choice locks a public API, a hire, or a migration, slow down and write it down.

Diagrams for this topic

Visual models you can redraw on a whiteboard or in a design review.

Diagram

Two-way vs one-way doors

Two-way door

  • Cheap to reverse
  • Decide fast
  • Experiment
  • Example: feature flag default

One-way door

  • Hard / expensive to reverse
  • Slow down
  • Write options
  • Example: public API contract

Diagram

Priority stack ritual
  1. 1List asks
  2. 2Force rank (no ties)
  3. 3Capacity math
  4. 4Publish what drops
  5. 5Review weekly

Worked examples

Concrete situations: what goes wrong, what to try instead, what changes.

Everything is P0

Setting. Product wants three launches. Security wants a patch. You have one squad.

Common miss

Say yes to all and hope for heroics.

Stronger move

Show capacity: 6 engineer-weeks free. Rank outcomes. Recommend cut list in writing with residual risk.

Outcome. Stakeholders choose; team stops thrashing.

The invisible cut list

Setting. Leadership wants a date. Your team knows scope will slip but nobody writes what will drop.

Common miss

Commit to the date and “figure it out later.”

Stronger move

Publish a ranked cut list with residual risk for each cut, and ask for an explicit choice on date vs scope.

Outcome. Stakeholders own the trade; the team stops thrashing in secret.

Resources specific to this chapter

Go deeper with books, videos, and tools matched to this topic, not a generic dump.

Practice

End-of-chapter drill

Force-rank your team’s top five asks. Publish what is explicitly not happening this sprint and why.

Hint: Use capacity math. No ties.

If everything is P0, you have not decided.