Definition
|
A risk is a potential problem, a situation
that, if it materializes, may adversely affect the project.
Risks that materialize are no longer risks, they are
problems.
All projects have risks, and all risks are
ultimately handled:
1). Some disappear,
2). some develop
into problems that demand attention, and
3). a few escalate
into crises that destroy projects.
The goal of risk management is to
ensure that risks never fall into the third category.
There are four steps to managing risks: identify them,
categorize them, mitigate them, and manage them. |
1.1
Identifying Risk
|
1.1 Identifying
Risk
Although all projects are different, the same
risks – those listed in Table 1.1 tend to recur. The list in
Table 1.1 is not exhaustive, and in identifying the risks
for a project, you must continually ask,
“What can possibly go wrong?”
If there is one risk that is universally the
most dangerous for all projects, it is the following:
Corporate management views the project
manager’s risk analysis as alarmist and will not take the
risks seriously until they materialize.
The only way to mitigate this risk is to
document all other risks, identify the actions you take, and
keep a management informed, especially as the risk becomes
more probable.
It is only by stressing your risk
analysis, by making explicit recommendations, and by
insisting that management understand the risks that you can
avoid having to say, “See, I told you so.”
|
1.2
Common Risks
Staff, equipment, client, scope, technology,
delivery and physical
|
1.2 Common Risks
Table 1.1 lists common risks that most
projects will encounter;
They form a starting point for developing a catalog of
risks. However, the list is not exhaustive; most project
managers will find several more risks that they can add, and
project experience will tend to increase this number. When
you are assessing the risks for your projects, always refer
to a list such as this. Otherwise, you run the project
management risk that not all project risks are identified.
Table 1.1: Sample list of project risks
|
Staff Risks
Key staff will not be available when
needed.
Key skill sets will not be available when
needed.
Staff will be lost during the project.
|
|
Equipment Risks
Required equipment will not be
delivered on time,
Access to hardware will be
restricted.
Equipment will fail.
|
|
Client Risks
Client resources will not be made available as
required.
Client staff will not reach decisions in a
timely manner.
Deliverables will not be reviewed according to
the schedule.
Knowledgeable client staff will be replaced by
those less qualified.
|
|
Scope Risks
Requirements for additional effort will surface.
Changes of scope will be deemed to be included
in the project.
Scope changes will be introduced without the
knowledge of project management.
|
|
Technology Risks
The technology will have technical or
performance limitations that endanger the
project.
Technology components will not be easily
integrated.
The technology is new and poorly understood. |
|
Delivery Risks
System response time will not be adequate.
System capacity requirements will exceed
available capacity.
The system will fail to meet functional
requirements
|
|
Physical Risks
The office will be damaged by fire, flood, or
other catastrophe.
A
computer virus will infect the development
system.
A
team member will steal confidential material and
make it available to competitors of the client. |
|
1.3
Categorizing Risks
|
1.3 Categorizing
Risks
There are numerous statistical methods for
defining degree of risks,
but the simplest categorization, and
therefore the most effective, is to describe risks as
extreme, high, medium, low, or minimal.
The degree of risk depends upon two
characteristics: the probability that the risk will occur,
and its impact on the project if it does.
Probability and impact are both categorized
as high, medium, and low, and their relationship, as
illustrated in Table 1.2 indicates the degree of risk.
Consider two risks: that a team member will
resign during the project and that a fire will consume the
office, destroying the installation and all the work that
has been done. Both risks are of medium degree. In the first
case, although the probability is high, the impact is low:
You assume that the team member will give adequate notice
and can be easily replaced. The second risk has a high – in
fact, potentially devastating – impact, but the probability
is low and the risk is easily mitigated by ensuring proper
off-site backup.
You categorize risks so that you can identify
those that are the most dangerous and therefore require the
most
attention. It is the extreme and
high risks that need your attention first.
Table 1.2: Categorization of degree of risk
Probability |
Impact |
High |
Medium |
Low |
High |
Extreme |
High |
Medium |
Medium |
High |
Medium |
Low |
Low |
Medium |
Low |
Minimal |
|
1.4
Mitigating Risks
By reducing
its probability, its impact or both
|
1.4 Mitigating Risks
You mitigate a risk by reducing its
probability, its impact, or both. Since every project is
unique, so are the mitigating actions. However, some
principles apply across projects and risks.
1.
Remove excuses:
When the project depends on someone (such as a supplier,
client, or line manager) to provide something (such as
staff, equipment, or material) in accordance with a
schedule, ensure that the provider knows the schedule,
knows what is expected, and understands the consequences of
a slippage. For major providers, such as the client,
make up a schedule giving the exact dates when the project
will require resources. If you have exact dates when the
project will require client resources. If you are not able
to give an exact date now, give a date by which you will be
able to.
You remove excuses by providing visibility
into the project, an active process in which provides are
forced to understand what is expected of them. For example,
if you have ordered a piece of equipment with a two-month
lead time to be delivered by a specified date, just putting
a required date on the purchase order is not enough. Four
weeks before delivery, call the sales representative to
verify the schedule. Three weeks prior, call to clarify, for
example, the power requirements. At two weeks, call to clear
up a technical question. One week ahead of time, call to
establish shipping procedures. With each call, of
course, you will ask if there are any problems that could
delay delivery, and you will emphasize how critical timely
delivery is. After this series of calls, the supplier has no
excuses to fall back on. There is no guarantee, of
course, that the equipment will actually be delivered on
time, but by actively reminding the supplier of the
schedule, you have reduced the probability of a late
delivery.
2.
Demand visibility:
when the
project depends on someone delivering something and there is
a process that the provider must follow before delivery,
you must understand at least the milestones of the
process. For example, if a piece of equipment must be
manufactured, identify the checkpoints in the manufacturing
process, have the sales representative attach dates to each
checkpoint, and call on those dates to ensure that the
milestones have been met and there are no delays.
If the process is repetitive, such as client
review and approval of project documents, understand the
process. What happens to a document when it is received?
Who reviews it? How are individual reviews reconciled? Is
there a final authority for approval? Who? What is the
priority of the project for the reviewers? With this
understanding, you will be able to suggest changes in the
process that will speed things up, if there are delays.
3.
Help people communicate:
When there is a surprise, the project manger
is frequently the last to know, even though the informal
communications network (or "rumor mill") among team members
and users contains various tidbits and snippets of
information that provide inklings of problems to come.
Helping people to communicate
increases the probability that useful information will find
its way to you.
The communications network can provide
advance warning that an employee is dissatisfied and looking
elsewhere,
that the performance of a system may be slower than
required, that software components may not integrate
smoothly, or that covert scope changes are being smuggled
into the system. In other words, the rumor mill is a prime
course of information about emerging risks.
The key rule to using the rumor mill is,
"Don't shoot the messenger." No matter how painful the
information, thank the deliverer; otherwise, like the jilted
spouse, you will be the last to know.
4.
Plan fallbacks:
If the technology does not perform
adequately, what can be done to improve it? If a critical
team member is lost to the project, how will those skills be
replaced? If the building burns down, how does the project
recover? Fallbacks are your plans for when the worst
happens.
Fallbacks must be capable of being put into action, either
now or when they are needed, and they must be capable of
being handled within the budget, schedule, and functionality
of the project. If this is not the case, they are not
fallbacks; they are wishes with nothing to anchor them but
the fervent hope they will never have to be exercised. |