Analysis and Design
Move from current state to future state, in that order
The analysis sequence is: agree the problem statement and scope, establish the baseline with reliable data, map the current state, find causes, prioritise, design the future state, then define how it will be measured. Skipping the baseline is the single most common design failure - you cannot claim improvement against a number you never took.
Techniques you must recognise
- SIPOC - high-level scope view: Suppliers, Inputs, Process, Outputs, Customers. Good for agreeing boundaries.
- Swimlane / cross-functional map - exposes handoffs and wait time between roles.
- Value stream map - distinguishes value-adding time from lead time; used to calculate flow and delay.
- 5 Whys and cause-and-effect (fishbone) - causal exploration, not proof.
- Pareto - separates the vital few contributors from the trivial many; drive design effort there.
- FMEA - anticipatory risk design: severity × occurrence × detection gives a priority number.
- Sampling and measurement system analysis - before you trust data, check that the measure is repeatable and the sample is representative.
Traps
Correlation presented as causation; conclusions drawn from a handful of observations; designing a solution before the cause is verified; and confusing a future-state design (what the process should do) with an implementation plan (how you will get there).
Rule of thumb
No design without a testable hypothesis. Write it as: "We believe that removing [cause] in [process] will move [metric] from [baseline] to [target] by [date], evidenced by [measure]." If you cannot fill every bracket, you are not ready to design.