This is a presentation I delivered during a webinar to two PMI communities of practice: Government and Requirements. The problems and potential solutions are common to just about any software solution delivery effort.
2. What’s About to Happen?
• What is the problem?
– And why is it an important one to solve?
• Some order-taking scenarios
• A better approach
• Specific techniques designed to enable business/service
success
3. The Problem: Consequences of public
sector failure are higher than private sector
• Governments typically have deeper pockets: result is that they
incur higher losses
– “a billion-plus of taxpayer’s money went to essentially
nothing.” (Lt. Gen. Charles Davis on recent USAF logistics
software project failure )
– US government is spending $81 billion/year on IT as of 2012
• Estimated that $20 billion of this is wasted per year
(Darrell Issa, Chairman of Senate Committee on
Oversight and Government Reform)
– The UK Office of National Statistics reported that public
service productivity fell 3.2% between 1997 and 2007; at the
same time annual ICT spending was between £13-21 billion
per year (according to Dr. Martin Read)
4. The Problem: Consequences of public
sector failure are higher than private sector
• Difficultly quantifying the benefits of public sector projects:
results in a tougher decision on when to quit spending money
• National security may be jeopardized
• Freedom of Information (FOI) access and Auditors: failure
made public
– Ontario’s Social Assistance and eHealth systems are
examples
• The problems tend to affect many people: private sector
comparisons include AT&T’s switching problems affecting 60K
long-distance customers in 1990 and 100K LinkedIn passwords
hacked in 2012
6. Requirements cited as the reason for failure
• Reasons vary, but at the top of everyone’s list are:
– Unrealistic or unarticulated project goals
– Badly defined system requirements
• Incomplete requirements
• Changing requirements and specifications
• Of course there are other reasons as well... but the point is that
we need to “get better at requirements”
– But why are requirements so bad?
7. Some typical reasons
• “No one told us they needed that”
– Ontario social assistance system was incapable of raising
rates (cost to fix: $10 million)
• “Nebulous requirements”
– The USAF logistics system failure example
• “Hi, we need a new field... status... child table”
– It seems to start off so innocently
• “We are not allowed to question the business – it offends them”
– IT departments do as they are told, acting as passthroughs rather than value-add organizations
8. A better way: an engineering approach
Business goals
and needs
Understand root
causes and discover
real business needs
Detailed software
requirements
Detailed technical
specifications
Explore and evaluate
alternative solutions
that meet the needs
Technical
capabilities
9. Traditional techniques for eliciting
requirements
• Designed to identify root causes of problems:
– Pareto (fishbone) diagrams
– 5-Whys
– Brainstorming
• Designed to understand a domain:
– Data modeling
– Interviews and observation
– Process and organizational modeling
• Confirm business goals/vision
• Designed to define software requirements:
– Use case modeling/user stories
– Non-functional requirements elicitation
10. My preferred techniques
• Understanding problems and desired outcomes
– Ensure how we measure we are successful at
resolving/achieving these
• The “spirit” of use cases
– Use cases happen to be documented, but they are not
documents
• Repeating back incorrectly or differently
– When people get a little miffed they tell you what they really
need
• Throwing real situations at requirements and draft designs
– Early feedback using real world scenarios
11. Problems and Desired Outcomes
• 2 questions I ask:
– What about the current approach is preventing you from
doing your work?
– 6 months after the project is complete, someone asks you,
“Was it successful?” What measures do you use to tell them
yes or no?
• The output of these sessions (samples follow) become the
driving forces behind the project
12. Sample Problem Statements
The problem of
Affects
The impact of which is
A successful solution would
Restricted store
hours due to
municipal
regulations and labor
restrictions
Customers
The inability to access
the store when needed
Enable customers to access
merchandise whenever they need to, no
matter where they are.
Marketing, Sales
Loss of positive brand
image since our products
are not available to our
customers when they
need them
Reestablish a strong brand image by
making our store the first place to look
when renovating.
Excessive inventory
carrying costs
Supply Chain
Operations
Limited ability to
carry large items in
stores
Limited ability to
carry large
quantities of all
items in stores
Enable our store to market all
household and hardware items, no
matter how large.
Facilitate “direct-from-supplier”
delivery.
13. Sample Desired Outcomes
Desired
Outcome
Objectives
Priority
Evaluation Criteria
Increased
Efficiency
No lost documents
Must
The number of tickets opened to find lost
documents on accounts is 20% of what it is before
the system is deployed.
Improved inspection review
process
Must
The turn-around time for reviews of inspections
must be 60% of what it is before the system is
deployed.
Service center team members
re-assigned
Must
Due to elimination of paper files, all staff
responsible for handling these can be re-assigned
to other work
Elimination of the binders
means that auditors do not
have to travel to the field
offices for audits; thus a
reduction in the travel budget
Must
The travel budget for audit is currently $27,600 for
the year 2010; this must go away in 2011
Remote machines are no
longer needed
Should
The 54 remote machines are removed from field
offices
Eliminate sending documents
to Iron Mountain, on a goforward basis
Must
No documents sent to Iron Mountain for accounts
effective 1 June 2010 and onwards.
Reduction in
Operational
Cost
14. The “spirit” of use cases
• The use case model actually reflects
the selected alternate solution
uc Inv oicing Use Cases
Invoicing System
This is not a
function. It
reprsents a
goal and a
story.
Submit Inv oice
Inv oice Submitter
Approv e Purchase
Order
Paymenrt System
Payment Manager
Reconcile Payments
The story has a
usual set of steps
(basic flow,
linear) and rare/
exceptional/
abnormal
alternatives
(lateral).
15. Summing up
• There is a higher risk that government projects fail to deliver
value, and the consequences of that failure can be exceedingly
high
• “Requirements” cited as the reason for failure
– Often due to pass-through order taking
– Real analysis of the business goals and exploration of
alternative solutions is key to turning this around
• There is hope: there are proven techniques for becoming a
“business-enabler” instead of an “order-taker”