PresentationPDF Available

Complexity & Innovation in Engineering Management

Authors:

Abstract

Complexity in engineering management, and tools for managing complex IT projects.
KU Leuven
Stefan Morcov
13. Nov. 2020
1
A staggering 31.1 percent of projects are canceled
52.7% of projects cost 189% the initial estimate
16.2% of software projects are on-time/on-budget
(Standish Group, 1995)
2
IT projects often exhibit traits of complexity and are difficult to manage and
control
Projects derail and face significant challenges
Some projects are particularly large and complex. Their cost may be huge
Arianne 5 exploded in 1996, due to an erroneous data conversion from 64-bit floating
point value to 16-bit signed integer, because of reusing legacy software from Arianne 4
- 370 mil Eur loss.
X-ray machine Therac-25, 11 machines in 1982, 6 accidents, 3 deaths
Schengen Information System (SIS II) launched in 2013 - 6 years late, 8 times more
expensive than the initial estimate, at a final cost of €500 million
3
So, why do we implement
complex (engineering) projects?
Complexity works!
Complex projects
create complex products, for complex markets,
in complex organizations, with complex processes.
They deliver Value.
4
5
6
7
Innovation occurs at the edge of chaos
Systems must be taken outside equilibrium to innovate
Systems acquire complexity to evolve & survive
8
Tester
QA
Analyst
Domain expert
UI/graphics
UX/usabilty
Architect
Project manager
Sys admin
DBA
Technical leader
Support
n*(n-1)/2
Difficult to understand, foresee and keep under control, even when
given reasonably complete information about its components”
Structural complexity =
complicated =
consisting of many varied
interrelated parts
Dynamic complexity: ambiguity, uncertainty, propagation and chaos
nonlinearity, complex feedback loops and significant impact of small
factors (Lorenz's butterfly effect)
11
2 consortium partners
5 key subcontractors
62 key experts
Customers:
admin: EACEA
business:
DG Education
DG Employment
tech: DIGIT
Central Support team
+ 38 National teams
4000 participants attended the Annual
Conference 2020
Large & very large
Long decision cycle - management
by committee
Critical projects political agenda
Complex stakeholder map
Varied political interests formal
and informal
Changing political priorities
Some organisations are very
vulnerable to political change, some
are more stable
Very complex regulations
Complex / complicated
14
Project
complexity
Market
complexity
Product
complexity
Process
complexity
Organisation
complexity
1. Know your alphabet: rules, standards, procedures
(Project) management / engineering school
2. Know when to ignore rules and cut corners
Experience
3. Make new rules
Create knowledge
15
Systematic project management framework
Iron-triangle
Project objectives * Delivery * Invoices & cost
16
Methodologies
Project management: PMBoK, IPMA, Prince2, PM2@EC, RUP@EC
Quality: ISO 9001, Total Quality Management, CMM, Six-sigma
IT Service Management: ITIL, COBIT
Software development: Waterfall, incremental, RUP, Agile, SCRUM, FDD, XP,
RAD, prototyping, OOAD, UML, dependency modelling, traceability matrices
Software estimation: FPA, COCOMO
Testing, ISTQB
Security: ISO 27001/2, CRISC, CISM, CISA, CISSP
17
18
Invoicing, cost,
profit
Project objectives
Statement of work
Deliverables
Features, requirements,
stakeholder needs& interests
Success criteria
Planning
When?
How much?
What?
(and what
not
)
Scope /objectives << Stakeholder management
Any system does something right
not necessarily what it was supposed to do
Communication, configuration
Change
Risk management
Project manager know-how and
experience
19
EVA = Earned-Value Analysis: combines scope, cost and time
Configuration management
Change management
Communication management
Leadership, motivation and HR
20
Tool: Stakeholder map
Analyze, document and manage:
Who impacts the project: name, title, role
How: objectives/interests
Contact points
Communication channels/frequency
Monitor for change !
21
Tool: risk register
1. Identify risks & opportunities
2. Analyze (qualitative / quantitative)
3. Plan mitigation strategies
Reduce, avoid, transfer, accept (risks)
Exploit, create, share, accept (opportunities)
4. Monitor & manage
22
(Marle&Vidal 2016)
Specialized & general tools
1. Decomposition (X-BS, modularization, COTS components, reusability)
2. Dependency modelling
3. Causal-loop, systems thinking
4. Brainstorming
5. Mind-maps
6. SWOT / PEST / STEEP
7. Checklists
8. Decision trees & matrices
9. Cost-benefit
10. Cause-effect / Ishikawa, Paretto (the 80-20 rule)
11. Delphi, Focus groups
12. Toyota way, 5 why-s
23
Which do you think is the best
project management software
tool?
MS Project / MS Project Server
Jira
...
Google Sheets / MS Excel
24
Project success vs. project management performance
(~process performance)
Traditional scope & QA management defines success as
compliance with
written specifications
25
Agile management: iterative, wave-crest planning /scope/analysis,
Manage a complex project as a program
Mongolian horde concept: just throw-in more people
First rule of cyclism: always run up-hill and against-wind
We can fill-in the details later
The 95%-ready paradox: ready 95%, 97, 99, 99.5, 99.75, 99.9%...
The KLOC paradox: measuring software code quantity vs. quality
26
“…is a temporary endeavor
undertaken to create a unique product/service(PMBoK)
In theory, practice is simple; but in practice, it’s not.
Clear theoretical solutions are applicable only to perfectly spherical cows in a
vacuum under zero-gravity conditions.
27
Mix. different contract types (fix-price, time&means, mixed)
Mix different suppliers
Informal communication. Teamwork
The right team. Balance people vs. Structure
Management vs. leadership
Do not oversimplify. Manage complexity &risk
A well-managed project is not a project without a mess, but a project with a
well-managed mess
28
Systems (projects, products) are:
29
1.
Simple
Managing ≈ no effort
1.
Complicated (≈ structural
complexity)
Managing ∝ effort
1.
Complex
Managing ∝ effort times x
1.
Really complex
Managing ∝ exponential to effort
= consisting of many
varied
interrelated parts
Tools for managing structural complexity:
Configuration management
Change management
Dependency (DSM, DMM, MDM) & traceability matrix (requirements, stakeholders,
changes)
(Maurer 2017) (Marle&Vidal 2016)
30
Ambiguity, uncertainty, propagation, chaos
Nonlinearity, complex feedback loops and significant impact of small factors
Lorenz’s Butterfly Effect
Lorenz: Deterministic non periodic flow, 1963
~Ray Bradbury short story, 1952
Nicolas Taleb: Black Swan (2007)
Fractals & long-tail vs. Gaussian distributions
Vulnerability management = Resistance + Resilience
Anti-fragility (Taleb 2012)
31
32
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Time/cost
Team size
Team composition and
performance
Urgency/flexibility of
cost/time/scope
Problem/solution
clarity
Requirements volatility
and risk
Strategic/political
sensitivity/importanc…
Level of organizational
change
Level of commercial
change
Risk, external
constraints and…
Level of IT complexity
Average
Hass complexity scale - example
Prj 1
Prj 2
Prj 3
Prj 4
Prj 5
33
Positive complexity Appropriate complexity Negative complexity
Benefits > Cost Benefits ≃ Cost Benefits < Cost
Desirable Accepted Undesirable
Innovation & creativity
Additional features
Large budget
Component reusability
Political priority
Integration/compliance with new
technologies
Unclear objectives scope agility
Many varied technologies
Many interrelated components.
Unclear objectives.
New technology.
Large number and variety of stakeholders
Response
strategy
Complexity Effect
Positive Appropriate Negative
Create, enhance X
Use (exploit) X
Accept / ignore X X X
Simplify / reduce
X
Avoid / eliminate
X
Deploying additional tools is expensive
More processes = more centralization
=> less innovation, less flexibility,
increased vulnerability
34
Innovation occurs at the edge of chaos
A system must be taken outside the equilibrium state in order to
innovate
Systems must acquire complexity in order to remain viable
The law of requisite complexity
Positive/appropriate complexity
35
36
Questions, suggestions, feedback welcome !
http://StefanMorcov.com
37
1. Define problem
2. Collect (ALL) ideas
3. Categorize ideas (e.g. card-sorting)
4. Prioritize ideas (e.g. by voting)
38
39
“divide et impera”
X-BS (Breakdown Structure)
Work-BS, Risk-BS, Cost-BS, …
Standardization, modularization & reusability of
designs/components (see Ford)
40
SWOT
Internal: Strengths, Weaknesses
External: Opportunities, Threats
PEST - Political, Economical, Social, Tehnological
STEEP - Social, Tehnological, Economical, Ecological, Political
41
42
43
44
Design Structure Matrix (DSM)
between same components
Dependency Mapping Matrix (DMM)
between different
components
Multiple-Domain Matrix (MDM)
(Maurer 2017)
DSM DMM MDM
45
46
47
(Sheard, Sarah A, and Ali Mostashari. 2010)
48
Comitology
: how committees, government cabinets, and
other such bodies are created and eventually grow irrelevant
Larger organisation structure = less power
Large bureaucracies tend to diffuse responsibility and
decision
Complex and long procedures for evaluation and career
plan
Low incentive for taking personal responsibility: limited
benefits vs. high risk
Stable, conservative organizations, averse to risk
49
Work expands so as to fill the time available
This law is true for both private & public organizations
Case-studies: UK Ministry of Naval Forces, UK Colonial Office
The staff rose by 57% per year irrespective of any variation in the
amount of work (if any) to be done
Driving forces :
An official wants to multiply subordinates, not rivals
Officials make work for each other
Which are the reverse driving forces ?
50
EPALE: the European Portal for Adult Learning
eTwinning and School Education Gateway portals
The European Digital Skills and Jobs platform
EU-Academy: the eLearning platform of EC
Future of Europe participative democracy
Business Continuity for the Publications Office systems
eLearning for IUCLID @EFSA
Research projects in AI (biometrics and risk-identification)
51
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.