Information Technology Reference
In-Depth Information
meeting with ITIL experts and the target organiza-
tion's practitioner which challenged the benefits
quantification. Hence, this KPI is included only in
the benefits quantification of the simulation with
estimated data because no real data was available
at the time.
Taking a look at the risk, it influenced greatly
the financial metrics included in the financial
analysis. Thus, when the risk increases, the NPV
and IRR decrease and the PBP increases.
Lastly, a higher level of a negative correlation
between variables is associated to higher values of
variance, standard deviation and skewness, which
means that the Monte Carlo simulation trials tend
to be more dispersed if there are negative correla-
tions between the variables considered.
As a consequence, different quantification log-
ics were applied in both simulations, for instance:
some KPIs are based on the employee productivity
whilst others are based on the cost per incident,
which is calculated by dividing the IT total costs
by the total number of incidents in a period of one
year. On the other hand, the employee productivity
is the result of the division of the revenue by the
total number of employees.
Using different forms of quantification isn't
necessarily incorrect. On the contrary, it makes
the estimator more correct as both forms are valid
and should be taken under consideration, since
the cost per incident is focused on cost reduction
and the employee productivity is driven towards
productivity gains.
Also, in view of the fact that the risk's influ-
ence over the investment analyses is enormous,
it is important that the risk analysis is performed
carefully and with the help of risk experts.
Simulation with Multiple Processes
Even though the correlations might not pay off
the superior project's investment costs, there were
more positive ROI scenarios in the Monte Carlo
simulation with correlations than in the one without
correlations, as it is illustrated in figure 5, which
is caused by the fact that the standard deviation,
variance and skewness are higher in the simula-
tion with correlations than in the one without, and
this causes the trials to be more dispersed in the
simulation with correlations.
simulation with Multiple processes
When the correlations amongst processes are
considered in order to perform a Monte Carlo
simulation, there are higher chances of more posi-
tive ROI outputs being generated as a consequence.
To prove this statement, the expected loss ratio
of the simulation without correlations is 7,35%
higher than the one of the simulation with cor-
relations, meaning that there are more 7,35% case
scenarios with a positive ROI in the simulation
that considers correlations. This result cannot be
expected in situations where correlations amid
processes are not considered at all.
The Monte Carlo simulation of the implementa-
tion of several processes could have an expected
loss ration higher than 50%, but the fact that the
benefits of each process positively influence the
benefits of all the other processes decreases the
expected loss ratio or, in other words, increases
the number of positive scenarios.
Even though the correlations might not pay
off the superior project's investment costs, it
eVALuAtion
This section is dedicated to evaluating how suc-
cessful the action was so as to test out how well
the proposed estimation process performed in
stipulating what action to take.
incident Management
process simulation
The simple fact that different KPIs were utilized
in the two simulations influenced greatly the ef-
fectiveness of the estimator itself.
Search WWH ::
Custom Search