Part 6 · 2 chapters · ~12 min

Problems at Scale: Ambiguous, Organisational, Long-Running

What changes when problems span teams and quarters: defining ambiguous problems, mapping owners and incentives, gathering scattered information, choosing a first step that builds momentum, writing the plan, and keeping long efforts alive.

12

Why they feel different

Puzzle methods still apply, but the dominant difficulties are definition, people, incentives and information. Staff engineers are hired to solve exactly these.

PROBLEMS AT SCALE
why they feel different, and what changes in the method
ambiguousnobody agrees what the problem ismany ownersno single person can fix itlong-runningmonths, with people joining and leavingincentivesteams are rewarded for other thingsinformationthe facts are spread across teams and systems
swipe the figure sideways, or tap expand for full screen
1/5
ambiguous
At scale, defining the problem is most of the work. "Deploys are slow" might mean the pipeline, the review queue, the approvals, or fear of breaking things.
defining it is most of the workinterview, measure, write it down
13

A method for big, fuzzy problems

steps
  1. Interview five to ten people closest to the pain; write down their words.
  2. Measure what you can (lead time, incident counts, queue ages) to replace opinions with numbers.
  3. Write a problem document: the problem, the evidence, who is affected, what has been tried, options.
  4. Find a first step that is small, visible and helps the people you need on side.
  5. Set checkpoints (monthly) and report progress against the measurements.