📈 Improvement Concepts

Concepts

The mental models behind StepChange: detect structural change, avoid tampering, and decide what to do next when outcomes don’t move.

▶ Open the StepChange Analyzer Interpret results What to do next
Foundation concept

Making the invisible visible

The symptom is visible. The cause is hidden. The constraint is structural. The contradiction is assumed away. Every great improvement framework is a different method for bridging that gap between what’s visible and what’s true.

Read the foundation concept →
Start here — five concepts that underpin everything else
  1. Why nothing changes — the failure pattern behind repeated effort without structural improvement
  2. Types of measures — outcome vs process vs balancing; lead vs lag
  3. Variation & SPC — noise vs signal and why tampering happens
  4. Joiner levels of fix — Level 1/2/3: the quickest predictor of whether an intervention will produce lasting change
  5. Necessary but not sufficient — why correct interventions still fail the sufficiency test

After those five, go to What to do next to apply them to your Bootstrap CUSUM result.

▶ Find the right concept for your situation
Bootstrap CUSUM shows a change point
Common cause vs special cause
Special cause variation
Change detected: what to do next
The Joiner Triangle
Joiner levels of fix
Making the invisible visible
Finding the real constraint
Divergent and convergent thinking
Stratify, experiment, disaggregate
Bootstrap CUSUM shows no change
Common cause vs special cause
Common cause variation
No change detected: what to do next
Decide to wait (don’t tamper)
Tampering & impatience
Stuck choosing or designing measures
Types of measures
Behaviour over time
Causal loop diagrams
Model for Improvement
Intervention not producing results
Necessary but not sufficient
Theory of Constraints
The Joiner Triangle
Joiner levels of fix
Finding the real constraint
Why nothing changes

Why things fail

Why nothing changes

Failure patterns behind repeated effort without structural improvement.

The Joiner Triangle

Quality, Scientific Approach, All One Team — one interdependent system. Taken separately they are not as effective.

Joiner levels of fix

Level 1/2/3: the quickest predictor of lasting change.

Deming’s 14 Points

Operational prescriptions that follow from systems thinking and variation.

The compliance trap

Why measurement pressure often worsens the outcomes it measures.

The Innovator’s Dilemma

Why systems resist disruptive improvement from within.

Common traps

The don’t-do-this list, in one place.


Measurement & analysis

Types of measures

Outcome vs process vs balancing; lead vs lag indicators.

Variation & SPC

Noise vs signal and why tampering happens when you confuse them.

Common cause vs special cause

Which type of variation do you have? The decision guide and what to do next for each.

Common cause variation

Stable systems performing at the wrong level: when to wait and when to redesign.

Special cause variation

Signals that something specific changed: eliminate before improving the stable system.

Gaming the Measure

Deming’s Point 11 — why targets get gamed and the balancing measures defence. Goodhart’s Law in NHS practice.

Hawthorne Effect

Behaviour changes when people know they are being measured. Why observed improvement is not the same as structural improvement.

Regression to the mean

Why improvement after a bad period may be statistical inevitability — and how Bootstrap CUSUM tells the difference.

Tampering & impatience

Why reacting to noise makes performance worse, not better.

Behaviour over time

Patterns before data; archetypes and feedback loops.

Causal loop diagrams

Feedback structure behind persistent performance patterns.


Improvement method

Model for Improvement

Aim, measures, changes; PDSA as structured learning.

PDSA cycle

Prediction → test → study → adapt. Bootstrap CUSUM is the Study step.

Making the invisible visible

Why every improvement framework solves the same problem — and where Bootstrap CUSUM fits.

Finding the real constraint

Six-phase process from symptoms to testable plan. TRIZ, SIT, four traps, contradiction analysis.

Divergent and convergent thinking

The two-diamond approach, evaluation matrix, PICK, POWER — 3 hours to a testable plan.

Stratify, experiment, disaggregate

Joiner’s three strategies for improving a stable system. What to do instead of tampering.

Going to the Gemba

The visibility intervention — going to where the constraint lives with the authority to act on it immediately.

Visual Management

Making work visible vs making the constraint visible — the honest distinction that determines whether visual management produces change or records the problem.

Evaporating Cloud

Makes the structural contradiction keeping a constraint in place visible — then dissolves it by naming and questioning the assumption underneath.

Bright Spots

Positive deviance: find who is doing it better and understand why.

Root cause analysis

Fishbone diagrams and multi-factor cause mapping.

The 5 Whys

Linear causality tool for straightforward, single-chain problems.

The PICK model

Impact vs effort prioritisation before choosing what to work on.


Systems & constraints

Theory of Constraints

Find and elevate the binding constraint. Improving non-constraints does not shift outcomes.

Necessary but not sufficient

Why correct interventions still fail the sufficiency test — and how Bootstrap CUSUM confirms it.

Systems thinking & silos

Why component optimisation harms the whole system.

Failure demand

Demand created by system failure and rework — the hidden load.

Focus & prioritisation

How to choose where to act when everything feels urgent.

Ashby’s Law

Requisite variety: why simple responses fail complex systems.


Next step

Run your time-series data through the StepChange Analyzer (Bootstrap CUSUM), then use Interpret results and What to do next to act on the finding without tampering.

▶ Open the StepChange Analyzer Interpret results What to do next 7-step method