Catalystic Leadership Field GuideCLS-001

Model for Improvement

Langley · Moen · K. Nolan · T. Nolan · Norman · Provost

A simple discipline for getting clear about the aim, defining what improvement would look like, and learning quickly through small tests.

See

The model at a glance

Clarify
Predict
Test
Learn
1

What are we trying to accomplish?

Clarify and align around the aim.

2

How will we know that a change is an improvement?

Define what evidence would count as improvement.

3

What changes can we make that will result in improvement?

Choose changes worth testing.

PLANDOSTUDYACT
PDSArepeat
Think clearly enough to learn quickly.
Understand

A rhythm between thinking and doing

The Model for Improvement creates a simple rhythm between thinking and doing. The three questions help a team get clear about what it is trying to accomplish, what evidence would count as improvement, and what changes are worth trying.

PDSA turns those ideas into small tests so the team can learn before making a larger commitment.

The three questions slow us down enough to think. PDSA gets us moving quickly enough to learn.
Two common traps

Acting before thinking

Jumping to solutions before the team is clear about the aim or what improvement would look like.

Thinking without acting

Analyzing and refining until the test becomes too large or too slow to generate learning.

The iteration
A test does not have to settle the question. Often, what we learn from one PDSA cycle tells us what to test next.
The discipline

Before testing, make a prediction. After testing, compare what happened with what you expected.

Before the testWhat do we expect to happen? What evidence will we look for?
After the testWhat actually happened? What did we learn?
“What happened compared with what we predicted?”
Work

Use it on a live improvement challenge

1Get the aim clear

  • Can you say what you are trying to accomplish in one sentence?
  • For whom? By how much? By when?
  • If you accomplish this, will it matter to the organization?

2Decide what evidence is good enough

  • How will you know things are actually better?
  • What simple measure could tell you?
  • What do you predict will happen?
  • What result would surprise you or change your mind?

3Make the test smaller

  • What test could we run by tomorrow—or in the next two hours?
  • Could we cut it in half—and cut it in half again?
Before you test, write down four things: Aim · prediction · measure · test.
Then compare: prediction · result · learning · next test.
Five-minute leadership conversation
  1. What are we trying to accomplish—and will it matter?
  2. How would we know the situation had genuinely improved?
  3. What do we predict will happen—and what small test could we run by tomorrow or in the next two hours?
The purpose is not to make the test impressive. It is to make the learning useful and fast.
Deepen

Use the model without turning it into paperwork

Common impediments
“We don’t actually agree on the aim.”
“We have a solution, but we haven’t agreed on what improvement would look like.”
“We don’t have a way to tell whether we improved.”
“We’re measuring activity rather than results.”
“Our test is too large to learn quickly.”
Common misuses
Jumping to Question 3Starting with “What should we do?” before agreeing on the aim or the evidence.
PDSA as paperworkThe cycle becomes documentation rather than learning.
Testing where expertise is already sufficientWhen a process is well understood, implement, verify, standardize, and exploit.
Oversized testsDesigning a pilot that takes months or affects the whole system when a much smaller test could answer the important question.
Catalystic application

The Model for Improvement gives people a practical way to improve the work they are closest to. At every level of an organization, people can clarify the aim, decide what improvement would look like, test a small change, learn from what happens, and decide what to try next. Improvement does not have to be delegated upward or handed off to experts. It gives people permission to discover better ways of working.

Related Catalystic lenses
Cynefin — Is the answer knowable through expertise, or must it be discovered through interaction with the system?
Statistical Control Charts — Is the process stable enough to interpret?
Prediction and Anomaly — What surprised us?
Explore / Exploit — Have we learned enough to standardize?
Practice note

Tom Nolan was known for spending substantial time asking different versions of “What are you trying to accomplish?” Leaders often described those conversations as unusually valuable, because the questioning surfaced differences in purpose before the group moved to solutions.

Source: Gerald J. Langley, Ronald D. Moen, Kevin M. Nolan, Thomas W. Nolan, Clifford L. Norman, and Lloyd P. Provost. The Improvement Guide: A Practical Approach to Enhancing Organizational Performance. 2nd ed. San Francisco: Jossey-Bass, 2009.