2. COST ACCOUNTING CONCEPTS FOR IS ORGANIZATIONS › 2.2 Budgeting Concepts › 2.2.3 Steps in Developing a Budget

2.2.3 Steps in Developing a Budget

Once the enterprise has determined the level of detail to
incorporate into the budget, the actual budget development
process can be broken down into four steps:

o  Initial forecast development
o  Forecast update and validation
o  Consolidation and management reporting
o  Finalize

The initial forecast development is normally based on unit
resource volumes projected from historical accounting
information.  The projected volumes are then multiplied by
the applicable rates to compute anticipated expenditures.
The expenditures are then summarized and subjected to a
"reasonableness check" usually by comparison with the
previous budget.  While this methodology is basically sound,
it harbors two potential pitfalls.  First, extremely large
projected growth rates seldom pass the reasonableness test
even if no evident source of error can be found.  Second,
even if the projected expenditures are accurate in total,
they may vary substantially from the month-to-month actual
costs.

Poor correlation between the projections and the actual
expenditures often results from an ineffective forecasting
technique or model.  Many organizations simply do a linear
projection of the billable elements over the planning
horizon.  Unfortunately, IS usage is rarely linear.  In most
cases studied, there were periodic spikes in usage usually
tied to a seasonal fluctuation in the business.
Consequently, in nearly every month, there was a variance
from the budget that the users were required to explain.  In
the rare cases where the users are not concerned about the
variance, a linear forecast will work, but, for most
organizations, a time series analysis (which will incorporate
the seasonal fluctuations into the projection) is more
appropriate.

Now comes the most critical step in developing a realistic
budget--validating and updating the forecast.  The initial
forecasts should be displayed on easily read reports which
incorporate plenty of white space for users to write in
changes.  Many enterprises download the forecasts into a
spreadsheet because they are easy to manipulate and they
yield machine readable output that can be used to quickly
update the initial forecasts.
The mark-up reports or spreadsheets are then used to manually
factor in plans and changes which were not reflected in the
historical accounting data.  These changes can be caused by a
number of factors.  Some of the possibilities include:

o  Changes in business processes
o  Changes in business volumes
o  New applications
o  Enhancements to existing applications
o  Completion of system implementations
o  Introduction of new technologies (hardware or software)
o  Implementing an information center
o  New IS service offerings or changes in the current
   offerings

This step involves communicating (usually in meetings) with
end users, business planners, IS capacity planners, and
application development staffs.  Once the anticipated changes
have been incorporated into the plan you are ready for
consolidation and management reporting and iterative tuning
of the plans.

The consolidation process is simply updating the master copy
of the forecasts with all of the information collected in the
interview process.  This is a good point for a sanity check
that compares the consolidated forecasts of units with the IS
capacity plans to determine if the users forecasts of demand
bear any resemblance to the IS organizations ability to
service them.  If there is a significant discrepancy, work
with the capacity planners to determine which one is correct.
The two forecasts are frequently done independently and can
yield very different results.  Resolution of the difference
may involve going back to some of the original interviewees.
Once this is resolved you can test the rates by comparing the
user budget projections against the IS spending plans.
Again, these may be very different for many reasons.  If your
objective is full cost recovery, a rate adjustment may be
required to balance the user billing with IS spending plans.
There is also a possibility that you will have identified an
unexpected demand that will result in a change in the IS
spending plans to satisfy the requirement.
Now you're ready for management reporting, both internal and
external to IS.  The IS management team wants to understand
how their services will be consumed and by whom and user
management team will be interested in how the forecast fits
within their spending plans and business objectives.  During
this phase you should expect to adjust the plans several
times.  Today's emphasis on cost reduction will prompt
questions such as "How can we get more out of IS at less
cost?" and "Which programs can we eliminate to get IS costs
within our spending plans?".

Many organizations will begin to take shortcuts at this
point.  These generally occur when upper management edicts
demand a flat percentage reduction of IS costs and the
forecasts are simply reduced by that percentage.  Keep in
mind that your forecast represents the plans of the business
units.  They must be involved in and committed to any cost
reduction activity.  Costs don't usually go down because
someone thinks they are too high, some action plan must be
devised and implemented to make it happen.  We know of
organizations in which the credibility of IS is negligible
because of a flat reduction that resulted in users being
significantly over budget after the first month of the fiscal
year.

Organizations that have not attempted to rigorously develop
IS budgets in the past should not to attempt to apply this
process to the entire IS workload at one time.  Start with
the largest workloads and work with them until the model is
refined and reasonably accurate.  This approach should
represent a high percentage of the budget and is often
sufficient.  To arrive at a total budget, group the smaller
workloads into a single category that you address as a whole.
It is usually not worth the effort required to address
everything in detail.  Nor is it likely that you will every
be able to model everything accurately at a reasonable cost.
However, the point of diminishing returns will vary from
organization to organization.