You are on page 1of 9

Measuring & managing software assets & business outcomes

Risk, revenue opportunities, and organizational agility in complex codebases


silverthread
silverthread
93% of large software projects sufer & fail from complexity
Example: Denver Airport baggage handling system failure Projects w/ budgets > 10M
Increased costs
Reduced performance
Lower cost-efectiveness
Carts jammed in tracks
Misaligned conveyor belts
Mutilated and lost bags
The Standish Group,
with a database of 50,000 development projects
http://www.computerworld.com/s/article/9243396/
Healthcare.gov_website_didn_t_have_a_chance_in_hell_
6.4% successful
52% challenged
Budget overrun
Schedule
overrun
Impaired
functionality
41.4% failures
Abandoned
Started again
from scratch
The fully automated system at the New Denver Airport is far
more complex than predecessor systems . . . This
enormous increase in complexity is the root of the problem.
- de Neufville
http://ardent.mit.edu/airports/ASP_papers/Bag%20System%20at%20Denver.PDF
2
silverthread
About Silverthread
3
How we can help The problem
Key software assets represent enormous
investments of money, time, and human
resources.
Some are deeply embedded in business
processes, critical to success.
Many codebases overburdened with
complexity:
High degree of risk
Hard to execute on new revenue
opportunities
Cost & schedule overruns
Difcult to evolve and adapt
Difcult to plan, manage, and understand
Costly to maintain
We can help evaluate a software asset, and
the wisdom of continued investment
Identify structural problems with large
nancial impact
Industry benchmarking
We can help you manage development
strategy to maximize ROI going forward
Analyze and visualize development
process, code structure, business metrics
Target the most important problem areas
Manage the process of transforming,
adapting, or simplifying the codebase
Track quality & productivity improvement,
quantitatively assess the nancial benets
silverthread
We can identify and explore specic problems that increase
the risk of defects & project or organizational failure
Benets of the approach
Find high-level problems known to
cause project and rm failure,
quality problems, risk
Identify specic organizational and
technical problem areas to probe
further
Measurement highly automated,
objective, and fast
Metrics valid across systems,
benchmarks against known
lemons and gems allow you to
know how this one compares
CodeMRI: See problems from 30,000 feet
Hidden wart causing major
nancial & organizational problems
4
silverthread
Modularity Hierarchy Layering
Prevent changes in
one area from
impacting others
Bound cognitive
burden placed on
individuals
Bound
communication
requirements

Facilitates bottom-
up, top-down design
Complex problems
broken up into small
simple ones
Infinitely scalable
can grow without
increasing local
complexity
Prevent feedback in
development
process
Create powerful
concepts and
abstractions
Make overall system
more flexible
Insulate people in
different domains
Hide information
Reuse
Long term waste
reduction
Build on and help
harden mature
capabilities
Many eyes detect
defects faster
New efforts move
more quickly
File hygiene
Easier to understand
code purpose
Easier to understand
data transformations
and follow execution
paths
Easier to enhance
Easier to debug
Easier to test
Better test coverage
possible
Our studies have demonstrated that when these properties break down, projects have
quality problems, experience cost & schedule overruns, produce more defect-prone
code, have higher long-term maintenance cost, and experience greater staf turnover.
We evaluate & benchmark codebases, focusing on things in the code
that directly impact business risk, revenue opportunities, agility, cost
5
silverthread
Our case studies show the nancial and risk related impact
of complexity across the entire software lifecycle
Defects & risk of failure Long-term maintenance cost Cost & schedule overruns
Analysis of critical defects found & xed
after a safety-critical system went live

20k 8K
Developer
Productivity
down 60%
In code with a measurably better architecture:
10% higher developer productivity during
maintenance
14% less time to release security patches
25% fewer incomplete or incorrect xes
Developers in low-complexity code
2.5X more productive
At least 30 FTE / 12% worth of efort
wasted per year
Over $4 million annual waste
3X defects in complex code
00.629% 13.004%
0.0319 0.822
Code with complexity
problems
Mean
defects xes
per LOC
% of elded
system with
critical bugs
Well structured
code
Study conducted at Fortune 100
engineering & systems conglomerate
Study conducted at Fortune 50
consumer oriented software rm
Well structured
code
Code with complexity
problems
Analysis of how complexity in a large
codebase impacts code production
Analysis of maintenance cost variance
between similar products (orig vs. rewritten)
Study conducted at software rm with
~100 products & $1B revenue
6
silverthread
7
Productivity &
cost metrics
Development
process metrics
Quality and risk
benchmarking
Cost / Benet &
ROI estimates
Codebase
deep dive
Improvement
tracking
Instrument &
Gather Data
Analyze
Visualize
Measure Report
Communicate Fix
Evaluate
We can help you decide whether to
continue investment in a software
asset
We can help you manage development
strategy to maximize ROI going forward
Silverthread approach: How can we help?
Silverthread capabilities and tools
silverthread
8
Carliss Baldwin
Founder & Board Member
Professor of Finance at
Harvard Business School
Author of Design Rules:
The Power of Modularity
CEO
Previously Software
Architect & Engineer at
MathWorks, General
Dynamcs, Veridian, PSR
PhD in Engineering, MIT
Dan Sturtevant
Michael Davies Alan MacCormack
Founder & Board Chairman
Founder, Endeavour
Partners
Faculty member at MIT &
London Business School
Founder & Board Member
Professor of Technology
Operations & Mgmt,
Harvard Business School
Expert in high-tech
product development
Director of Development
Previously Engineering
Manager & Senior
Engineer at Red Hat,
Egenera
Kenyon College
Dave Allan Jodie Zong
Developer
Previously at Hewlett-
Packard Co.
Northwestern University,
Computer Science &
Economics
Natalia Mocan
Operations & Finance
Previously at Grant
Thorton, Russian School
of Mathematics
Univ of Massachusetts,
Finance & Economics
Sunny Ahn
Strategic Advisor
Partner at Endeavour
Partners, Former CEO of
Context Connect
Expert in business
development and
corporate strategy in tech
silverthread

Reducing risk and accelerating results
by understanding and simplifying complex software
Cambridge, MA

You might also like