Bob Edwards

Work Simulation
1st Industrial
Revolution
Industry 4.0
Complex Systems
A Sense Making Model for Systems (The Cynefin Framework)

Complex
Complicated

Ordered Systems

Chaotic
Simple

(David Snowden)
 tools, programs,
language, behavior

shared beliefs, values

and assumptions
programs, practices,
rituals

shared beliefs, values

and assumptions
 The problem with AND?

Efficiency Thoroughness Trade Off

(Hollnagel)
programs, practices,
rituals

shared beliefs, values

and assumptions Foundational Beliefs
H.O.P. Foundational Beliefs

➢ Work is Complex
➢ People Make Mistakes
➢ Blame Wastes Resources
➢ Context Drives Behavior
➢ Learning & Improving is Vital
➢ Leader’s Response Matters
Assumption: if you try hard enough
you won’t make
mistakes
Error
- Didn’t intend my actions
- Didn’t intend a negative outcome.

Mistake
- I intended my actions
- Didn’t intend a negative outcome.

Person Problem
- I intended my actions
- I intended a negative outcome.
Is error a choice?

Are mistakes a
choice?
“Mistakes arise directly from
the way the mind handles
information, not through
stupidity or carelessness.”
(Edward de Bono PhD)
Is error bad?

- Error is not intrinsically bad

- Error is normal
- We are error-guided creatures
- We depend on trial and error.
October 1879
(13 ½ Hours)
Error is not a Choice…

Error is not a Violation.

Error Trap?
Provocative Error Trap?
Norman Door
 the worker is the
Assumption: problem
“Fundamentally,
people come to work to do what?

Good work!“
Waldorf & Statler
Response to an event
We need to answer the question . . .

Do we want retribution?
or
Do we want restoration?
We can blame and punish?
or
learn and improve?

But we can’t do both!

(Conklin)
Human or system problem?
Saturday
Jan 13, 2018
8:07 am

Fundamental Attribution Error

- Lots of recent EMA activity due to North Korea threat
- Sirens tested in Dec for 1st time in 30 years
- Test occurred at shift change, non-standard drill time
- Leadership set the drill in motion and acted like it
wasn’t a drill (disguised voices)
- 5 others on the call – placed on speaker phone part
way through (Operator did not hear “exercise,
exercise, exercise”)
- Operator had confusion on 2 past drills (10 year
veteran)
- Most sirens did not work
- Shelters were locked
- Phone and internet overloaded (911 calls didn’t work)
- Leadership had to call Pacific Command to verify no
threat
- No military base activity taken
- People told to take alerts seriously next time.
“insufficient management
controls, poor computer software
design, and human factors”

“employee had performance

issues and had messed up on at
least two previous drills”

Resulting in:
- Firing the operator
- EMA manager resigning

Blame fixes nothing!

“It doesn’t take phenomenal ability to realize that
a person who is given blame learns how to avoid
the blame next time, while the person who gives
blame learns nothing. As a result, things continue
to go wrong . . .” Bill Salot
H.O.P. is NOT the absence of
rules or discipline

. . . its the notion that if you depend

on a person doing something 100%
right 100% of the time…

…you will be disappointed...

…A LOT
(Andrea Baker)
Don’t we need
to hold people
accountable?
(Andrea Baker)
The less we know. . .

. . . the more we tend to blame . . .

. . . the more we blame . . .

. . . the less we are going to learn.

Charles Major - Luminant

“No one, and I mean no one, has
punished their way to operational
excellence!”
Trust is a lot to ask?

Lets start with

establishing an
environment where
we can at least be
honest!
“...blame is the enemy of
understanding.”

(Andrew Hopkins)
People Are As Safe As
They “Think” They Need To
Be,
Without Being Overly
Safe…
In Order To Get Their Job
Done.

(Conklin/Edwards)
Drivers Are As Safe
As They “Think” They Need
To Be,
Without Being Overly
Safe…
In Order To Get To Their
Destination.

(Edwards)
Assumption: What we measure, we
improve.
 The METRIC Bias
“what gets measured gets improved”
Metrics?

Have we become managers

of metrics?

Instead of leaders of
people?
We often measure what is easy
to measure but may not be
that important, because often
times the things that are
important are hard to measure.
The Tyranny of Metrics, J. Muller
Great metrics
don’t necessarily
mean you have
great
performance.
 Fatality

Basic Safety
Reactive – Incident Based

Better Design
Injuries Proactive – Process based

Time
“Do we know how brittle
we actually are?”
(David Payne)
Can we change people’s
behavior?
Changing Behavior??
Behavior Modification
Behavior Change

Belief Systems (Values)

What gets rewarded –
gets repeated (positively and negatively)
(Edgar Schein)
“You cannot change
the human condition,
but you can change
the conditions under
which humans work”
James Reason
Our Goal . . .

. . . is to become less surprised by

human error and failure . . .

. . . and instead, become a lot

more
interested in and a lot better
at operational learning!
(Edwards)
La falla es una combinación de variabilidad normal entre
el Trabajo planeado vs el trabajo practicado en la realidad
Trabajo planeado
Realidad
Normalmente
Exitoso!

"Expertos de la
línea azul"
(Hollnagel, Conklin, Baker, Edwards)
3 Parts of an Event

(Conklin)
3 Parts of an Event

The Challenge:
Not to let
post-event hindsight
bias our judgment of the
pre-event context.
(Conklin)
For every complex problem,
there is an answer that is neat,
simple, easy to understand
and probably wrong!

(Edwards)
“Underneath every seemingly obvious,
simple story of error, there is a second
deeper story. A more complicated story . . . a
story about the system in which people
work.”
(Dekker, 2006)
Some tools lead us to a linear
understanding of the event . . .
. . . which may be enough

Root
5 4 3 2 1 Event
Cause?

The problem is, complex failures are not linear . . .

. . . and there may not be a single actual root cause.

 Start back in process . . .
. . . move towards the event
Weak Signals
Production pressure
Unclear Signals Adaptation
Fear of reporting Latent Conditions
System Strengths System Weaknesses
Errors
Resource constraints Hazards & Risks
Surprises Event
Local Factors Flawed processes Personal Factors
Incomplete Procedures Normal Variability
Data
Design shortcomings Near Misses Past Success No Surprise!
Poor communication Tradeoffs
Change in plans Goal Conflict
Incompatibilities
(Conklin/Baker/Edwards/Howe and more)
Difference between
Failure and
Success?
 Start back in process . . .
. . . move towards the event
Weak Signals
Production pressure
Unclear Signals Adaptation
Fear of reporting Latent Conditions
System Strengths System Weaknesses
Errors
Resource constraints Hazards & Risks
Surprises Success
Local Factors Flawed processes Personal Factors
Incomplete Procedures Normal Variability
Data
Design shortcomings Near Misses Past Success
Poor communication Tradeoffs
Change in plans Goal Conflict
Incompatibilities
(Conklin/Baker/Edwards/Howe and more)
Success

Failure is a combination of
normal variability

(Hollnagel, Conklin, Edwards, Baker)

A Sense Making Model for Systems (The Cynefin Framework)

Complex
Listening Faster Complicated
Fail Safe Experiments Good Practices
Learning from success Fail Safe Design

Ordered Systems

Chaotic
Simple
Novel Practices Best Practices
Emergency Response Plans Fail Safe Design
Incident command structures
Root
Cause? 1 2 3 4 5 Event

(David Snowden)
If we want better answers
and a deeper understanding. . .
we have to ask
better questions!

(Conklin, Edwards)
Expand the question
from “why did you
do that?” . . .

. . . to “how do we
normally do this
work?”

(Conklin)
Let’s learn and
improve . . .

. . . and build more

capacity
 The things people think and talk about,
what they discover and learn, are implicit
in the very first questions asked.
Questions are never neutral, they are
fateful.

(David Cooperrider, Suresh Srivastva, 1987)

Hard to crash
& safe to crash
Great performance is not
the absence of errors…

… it’s the presence of

capacity
(Conklin)
Defenses
• Types of Defenses
• Strength of Defenses
• Layers of Defense
• Sustainability of Defenses
Hierarchy of Controls ??
• Elimination
• Substitution
• Engineering Controls
• Administrative Controls
• PPE

More focused on ownership and effectiveness.

Procedures are important…

But they are not sufficient

enough to create safety

Our organizations have become

complex-webs of procedures that
are incomplete and difficult. (Conklin)
What does this mean?
Request?
Fact?
Threat?
Promise?
If you want great procedures,
Have those who have to use them,
help you write them.
Strength of Defenses
Critical Steps

Documented
Procedural
Steps

CRITICAL STEPS
All Steps and Tasks - Non-recoverable
- Must be done correctly
- Need Essential Controls

Essential controls must be in place around the Main chute fails to open 0.1%

conditions that must be satisfied in order to prevent Fatality rate 0.001%

catastrophe after the critical step. Source: www.uspa.org
Reliability &
Resilience
Aug 7, 2015. DL1889 Boston to Salt Lake.

Emergency landing in Denver

Photo: Jack Thompson/EPA
Response to this message
 When we believe we know the
answer . . .

. . . we stop asking questions

. . . we stop listening
. . . we stop learning!
The power to ask the right
questions . . .

. . . comes from acknowledging

that you don’t know the right
answer.
The worker is not the problem
to be solved . . .

. . . the worker is the problem

solver.
Sterigenics
“I have never been especially
impressed by the heroics of
people convinced they are
about to change the world. I
am more awed by those who
struggle to make one small
difference.”
(Ellen Goodman)
Resources www.hophub.org
www.hopcommunity.org

Todd Conklin, PhD Sidney Dekker, PhD Bob Edwards Chris Clearfield
Andrea Baker Andras Tilcsik

Weick & Sutcliffe Jerry Muller Edgar Schein, PhD Don Norman
Bob Edwards
H.O.P. Foundational Beliefs

➢ Work is Complex
➢ People Make Mistakes
➢ Blame Wastes Resources
➢ Context Drives Behavior
➢ Learning & Improving is Vital
➢ Leader’s Response Matters
What is Operational Learning?
Not a traditional investigation
Not worried about collusion
Not focused on the “one true story”
Not focused on the one “root cause?”
Not focused on blame
Tells the story of how work normally gets done.
Tells the story of complexity
Tells the story of normal variability and coupling
(Tells how the conditions lead to this type of event
if an event brought the Learning Team together)
 Ingredients for Operational
Learning (including Learning
Teams)

Safe place to Being Learning Valuing

talk teachable First Soak Time

Defining the Team

Problems Generated Ideas

(Baker/Edwards)
Determine Need & The HOP LT Cycle
Establish Team

Learn Gain
Understanding

Soak Time Define

Problems &
Prioritize

Try-Storm
Ideas
Sustain & Improve
Monitor Conditions
Test
Defenses
(Edwards, Baker, Conklin PhD)
 The things people think and talk about,
what they discover and learn, are implicit
in the very first questions asked.
Questions are never neutral, they are
fateful.

(David Cooperrider, Suresh Srivastva, 1987)

If we want better answers
and a deeper understanding. . .
we have to ask
better questions!

(Conklin, Edwards)
Expand the question
from “why did you
do that?” . . .

. . . to “how does
work normally get
done?”

(Conklin, Edwards)
Oct 12, 1997 at 5:18 PM
John Denver dies in plane
crash

Failed to fill up tanks before flight

Did not put plane on auto-pilot
Hit right rudder pedal by accident
Pilot Error! Images from Wikipedia.org
20 years experience
www.angelfire.com
2400 hours
Single / Multi / Aerobatic Bi-Planes / Lear Jet
Had flown it once for 30 minutes prior day
Quick one hour flight
Buzz neighbor’s house.
Switch location / labeling
Operation (up=off / down = right / right = left)
Fuel gauge marking
Fuel consumption rate.
When?
• Not for everything (resources!!)
• Based on severity (or potential)
• Post-event (Injury/Quality/Operations)
• Near Miss or Close Call
• Good Catch
• Interesting Successes
• High Risk Operations
• Challenging Design Problems
• Not for determining punishment
• Not for criminal behavior.
Everybody knows . . .

Audible Alarm
Learning Team Make-up

• Coach or Facilitator (and co-facilitator)

• Small enough to manage but large enough to
capture the context (i.e. 5 – 7ish)
• Those close to the event or issue
• Possibly someone from outside the process
• Support members as needed
• Leadership to sponsor it and kick it off (they
may or may not be able to stay, depends. If you
are not sure, have them step out)
Industrial Empathy
“Our goal is to learn enough that we
realize, given the conditions they faced
and the information they had, the tools
and equipment they used and the
pressure they were under, that we
would probably have made the same
decision.”
(Edwards/Baker)
Intersection
Break Room
Determine Need & The HOP LT Cycle
Establish Team

Learn Gain
Understanding

Soak Time Define

Problems &
Prioritize

Try-Storm
Ideas
Sustain & Improve
Monitor Conditions
Test
Defenses
(Edwards, Baker, Conklin PhD)
Soak Time Tips
• At least overnight (if at all possible)
• Allows time to process learnings
• Allows time to go look
• Allows the coach time to think of additional
questions.
Circle of control
Circle of Influence
Circle of Concern
 Micro-experiment
Try-storm ideas
Be completely present
with the Learning Team.
Coaching Learning Teams
• Be a coach (Take off your safety hat)
• Co-coach (Builds capacity and helps a lot)
• Good listener (Everyone can’t be a coach!)
• Be curious ( . . . and then be quiet)
• Nonjudgmental (or at least act like it)
• Open minded (You may not always like what you hear)
• Engaging (Value everyone’s input)
• Street credibility (Basic process knowledge helps)
• Encourage (Let them figure it out)
• Be interested in helping tell the blue line story
• Don’t be afraid to tell the real “blue line” story.
Seek out differing opinions . . .
Determine Need & The HOP LT Cycle
Establish Team

Learn Gain
Understanding

Define
Tell the Learning
Soak Time
Problems &
Prioritize
Team Story
Try-Storm
Ideas
Sustain & Improve
Monitor Conditions
Test
Defenses
(Edwards, Baker, Conklin PhD)
 It's hardly ever a
safety problem . . .

It’s usually an operational

problem!
Resources www.hophub.org
www.hopcommunity.org

Todd Conklin, PhD Sidney Dekker, PhD Bob Edwards Chris Clearfield
Andrea Baker Andras Tilcsik

Weick & Sutcliffe Jerry Muller Edgar Schein, PhD Don Norman