Cynefin sensemaking framework and usage examples
- 2. www.luxoft.com
Maksim Gaponov – Agile/Lean Coach
Experience
10+ years in IT. Developer background and management experience as a team
lead, project manager, product manager and CTO. Worked both in large
companies and smaller start-ups. Over 7 years ofAgile experience, Maksim is an
seasoned coach, trainer and speaker at various conferences, including AgileDays.
Certificates
§ Member of Scrum Alliance, Certified Scrum ProductOwner
§ Member of International Consortiumfor Agile (ICAgile), ICAgile Certified
Professional, ICP-Business Value Analysis, ICP-Agile Coaching, ICAgile
accredited trainer
§ Member of Scrum.org, Professional Scrum Master (PSM I)
- 6. www.luxoft.com
Complex Adaptive Systems
w Understanding through constraints:
- Order: fully constrained; predictable &
repeatable outcomes
- Chaos: random, unconstrained, difficult to
create or sustain
- Complex: system partially constrains
behaviour, but behaviour modifies
constraint
w The inherent uncertainty of a complex
systems means we have to navigate a fitness
landscape of possibilities
- 7. www.luxoft.com
Let’s take away and map ordered ones
w Understanding through constraints:
- Order: fully constrained; predictable &
repeatable outcomes
- Chaos: random, unconstrained, difficult to
create or sustain
- Complex: system partially constrains
behaviour, but behaviour modifies
constraint
w The inherent uncertainty of a complex
systems means we have to navigate a fitness
landscape of possibilities
- 8. www.luxoft.com
Now let’s take away the chaotic ones
w Understanding through constraints:
- Order: fully constrained; predictable &
repeatable outcomes
- Chaos: random, unconstrained, difficult to
create or sustain
- Complex: system partially constrains
behaviour, but behaviour modifies
constraint
w The inherent uncertainty of a complex
systems means we have to navigate a fitness
landscape of possibilities
- 9. www.luxoft.com
Finally let’s take away the ones falling into complex description
w Understanding through constraints:
- Order: fully constrained; predictable &
repeatable outcomes
- Chaos: random, unconstrained, difficult to
create or sustain
- Complex: system partially constrains
behaviour, but behaviour modifies
constraint
w The inherent uncertainty of a complex
systems means we have to navigate a fitness
landscape of possibilities
- 10. www.luxoft.com
We’re spending most of our time in Complex domain
w Highly sensitiveto small changes (weak
signals too easily dismissed)
w Granularity, gradient & enabling
constraints
w Proximity & connectivityare key
w Shift from fail-safe design to multiple
contradictorysafe-to-fail experiments to
avoid prematureconvergence
w Avoid confusing correlation with
causation or simulationwith prediction
- 11. www.luxoft.com
Obvious domain
Nature
Known knowns
Familiar, certain, well worn
pathways
Constraints self-evident to all
Stable within universal
constraints
Self evident solutions
Response
Sense-Categorize-
Respond
Ensure sound process in place
Monitor for noncompliance &
deviance
Test for complacency
Protect some pet mavericks
Preparation
Watch for outliers
Usable process
Right support people for key
staff
Automate, but no automata
Anonymous appeal/whistle
blowers
- 12. www.luxoft.com
Complicated domain
Nature
Known unknowns
Ordered, predictable,
forecastable
Constraints evident and
enforceable
Stable within constraints
Evidence susceptible of
analysis
Response
Sense-Analyze-Respond
Determine experts or process
to resolve
Manage & enforce process
Monitor effectiveness of
constraints
Focus on exploitation not
exploration
Preparation
Right people & process,
right time
Process engineering with
feedback
Diversity of experts in network
Sound Analytical practice
Stand aside but stay in touch
- 13. www.luxoft.com
Chaotic domain
Nature
Unknowable unknowns
Temporary state - no time
No evidence of any constraint
High turbulence no patterns
Old certainties no longer apply
Response
Act-Sense-Respond
Speed of authoritative response
vital
Follow and enforce heuristics
Focus on constraints not
solution
Use the opportunity to
innovate
Preparation
Establishheuristics &
Parables
Human sensor networks
Crews as crisis management
teams
Simulation games for key
managers
Multi-perspective dissent
feedback
- 14. www.luxoft.com
Complex domain
Nature
Unknown unknowns
Messily coherent, patterns
discernible
Partial changing constraints
Flux within stabilities
Evidence supports
contradiction
Response
Probe-Sense-Respond
Monitor safe-to fail
experiments
All contradiction within
heuristics
Flex constraints to manage
emergence
Agility key to
amplification/dampening
Preparation
Create time and space for
reflection
Human sensor networks
operational
Social network stimulation &
the like to create networks
Scenario planning (inc. micro)
Build and monitor ‘requisite
diversity’
- 15. www.luxoft.com
Making sense of the world to act in it
There are three basic types of system: ordered,
complex and chaotic.
Complex systems have propensities and
dispositions but no linear material cause
In Cynefin order is divided into ‘obvious’ &
complicated and the fifth domain) disorder.
The boundary between Obvious and Chaotic is
shown as a cliff, or a catastrophic failure arising
from complacency.
- 17. www.luxoft.com
Using Cynefin for User Story splitting
w Obvious – Just build it. Or, if it’s too big, find all
the stories, and do the most valuable ones first.
w Complicated – Find all the stories, and do the
most valuable and/or most risky ones first.
w Complex – Don’t try to find all the stories. Find
one or two that will provide some value and
teach you something about the problem and
solution, build those and use what you learn to
find the rest.
w Chaotic – Put out the fire; splitting stories
probably isn’t important right now.
w Disordered – Figure out which domain you’re in
before splitting so you don’t take the wrong
approach. Richard Lawrence, http://www.agileforall.com/2014/08/cynefin-and-story-splitting/
- 19. www.luxoft.com
Using Cynefin on Retrospectives
w Obvious – just implement these
improvements, add them to working
agreements
w Complicated – find experts in your
environment or internal/external consultants,
ask them to help you
w Complex – brainstorm a set of small and
time-bound experiments, find a way to
measure outcomes, implement them
w Chaotic – distribute knowledge and wait
some time until the issue transitions into a
complex domain
w Disordered – brake these down until they
start falling into one of domains
- 21. www.luxoft.com
Interventions in the complex domain
w Each intervention should be
- coherent
- safe-to-fail
- Finely grained, tangible
w Overall the portfolio should contain
- some that are oblique in nature
- include “naive” approaches
- a few high risk/high return options
w Contradictions are good!