summary
Traditionally, understanding the benefits of end user
computing has been more enticing than understanding the cost structure. Unfortunately, the
benefits have been hard to pin down (particularly with client/server architectures) while
more direct and indirect costs have been incurred than anticipated. Executives are
realizing that better methods are needed to understand and project costs.
There has also been a major shift in the speed of technological change. End user computing
technology has historically undergone major transitions at five year intervals. Now, Bill
Gates has stated that everything Microsoft currently makes will be obsolete in three
years. This means that organizations are facing an era of “continuous migration”
where costs that were assumed to be one-time variable costs are now actually periodic
continuing fixed costs.
Organizations must recognize that:
 New end
user technology is no longer being replaced at once across an entire organization. This is
too costly. It now propagates through an organization creating “The Lasagna
Effect” … a heterogeneous mix of technologies.
Many
organizations “sub-optimize” by making IS decisions without the involvement of
all the organizational stakeholders and without focusing on all organizational
costs/benefits.
Many
organizations use approximate or incomplete end user computing analysis methods which do
not accurately identify all costs. Also, these analyses are rarely validated –
reopened to see if they delivered promised costs/benefits.
The
traditional “Total Cost of Ownership” cost model and decision processes are no
longer as accurate in the new technological environment of rapid and continuous
transition. New models and methods based upon transitions must be developed.
A
Transition Cost Model must have the following main characteristics:
– transition based, not single
technology
– individual based, not machine
based
– activity based costing, not
traditional chart of accounts costing
– three year life, not five year
life
A plan must
be initiated immediately if an organization is to minimize costs and maximize the end user
computing technology to be acquired in the critical next two years.
Executive management must understand that technology is now an ever present continuation
of transition. Projections generated by a Transition Costing Model will help management to
make the right economic decision on technology.
research results
The development of a valid End User Computing (EUC)
costing methodology for the 90’s involves significant investigation. To determine the
“state of the art,” research focused in three areas:
Literature
search
– trade literature
– general business and news publications
– academic publications
– industry study reports
– investment analyst reports
– industry trade newsletters
– Internet resources
– Customer costing studies
Interviews
– customers undergoing transitions
– vendor and reseller sales staffs
Accounting
“best practices” review
– vendor costing tools (Microsoft, Apple,
Novell,
Symantec, Macromedia)
– industrial engineering discounted cash
flow models
– relevant FASB Original Pronouncements
With this information we were able to identify the key attributes of a valid costing
model.areas of concern in
EUC evaluation
There are a number of common situations facing the clients we
interviewed, situations which weaken a “best practice” cost analysis effort:
IS issues dominate. IS-focused decisions can result in what reengineering
experts call “suboptimization” – decisions which benefit one portion of an
organization to the greater detriment of other departments. For example, trimming the IS
support budget does not eliminate the user need. This unavoidable support function
transfers to end users.
Decisions made by fiat. Many PC platform meetings end up sounding like
religious discussions. The test of a decision by fiat is to ask: “Where is your data?
Have you put a pencil to that? Can you show me any study results?”
Decisions made by “common sense.” There are financial and
technological concepts which make intuitive sense, but are inaccurate. An example is dual
platform supposedly doubling support costs. A 1995 Gartner Group Consulting Services
survey of enterprise organizations established that “there is no detectable technical
support cost ‘premium’ associated with dual platform (MacOS and Windows 3.xx)
environments, compared to single platform environments.” Data often undermines
“common sense.”
Incomplete chart of accounts to capture “hidden” costs. Most
organizations face a major hurdle in determining the actual costs of EUC. They do not have
a Chart of Accounts coding system which can recognize all costs of end user computing.
Financial models are required to estimate true costs of EUC.
Heavy dependence upon outside consulting data. In the absence of live
accounting data, industry estimates provide a starting point. But does the model
accurately reflect the cost categories a specific organization? Are the costs for each
category, such as support personnel wages, the same? There should be a formal process of
verification.
Using a model as fact. Another concern is the precision of the data. It
is easy to become over-reliant upon exact figures. Estimates, by their very nature, cannot
be combined to generate a precise result.
Limited application of “life cycle cost” concepts. Industry
experts estimate that EUC acquisition costs now make up only about 14-17% of the total
life cycle costs. The big leverage is not in the IS department, but at the end user’s
desk where the benefits of EUC technology are being delivered.
Generally low internal analysis sophistication. There is an
overall lack of sophistication of cost modeling and analyses most organizations provide to
management.
Little analysis of the “tomorrow” results of “today”
decisions. Technology is moving so fast that many buyers must think of each
architecture as an annuity for the vendor – particularly with OS’s and
application suites. Most costing studies ignore this inevitable series of transitions.
Studies are typically done based on a single hardware and software platform over a five
year life.
Little accountability for past decisions. A common theme in the
literature is, “We’ve spent all this money on end user computing, where are the
benefits?” Paul Strassmann, in his classic book The Business Value of Computing,
could find little linkage between total computing expenditures and any measures of
business success. Few organizations go back to validate promised benefits.
There is nothing glamorous in understanding cost components, but this understanding is
this basic “blocking and tackling” which provides a solid foundation for
decision-making. To do this, an End User Computing costing approach more accurate than the
current Total Cost of Ownership model is needed.
how technology is deployed
Major PC platform changes have traditionally occurred
every five years or so. Organizations upgraded their technologies for leading edge users
and distributed older technology to “unwashed” users – those computer
illiterate users receiving their first personal computer. In a DOS environment,
organizations were able to maintain a relatively stable, homogeneous personal computing
platform.
Today’s desktop environment is just the opposite. Technology decision makers are
faced with constant transition. Typically, machines quit satisfying users long before the
end of their traditional life cycle. The question is, What is done with these relatively
new, eminently usable yet technologically “past their prime” computers? Few
organizations choose to replace all end user computing in a single purchase. The answer
for most organizations is a regular, partial infusion of new technology.
Most organizations choose to continuously acquire technology
for certain types of users. What separates end users is their level of sophistication,
extent of needs, and where they are in the technology adoption cycle:
The theory is to purchase new end user computing technology, then pass it down through the
user adoption cycle. This “hand me down” approach supposedly maximizes the
useful life of the technology by getting equipment at the right level of capability to the
right person at the right time. And it supposedly reduces total life cycle costs by
minimizing acquisition costs over time. But what it certainly does is increase other costs
by forcing organizations to manage a highly mixed (heterogeneous) end user computing
environment.
the “Lasagna Effect”
So there are two variables over time, technology and user
adoption. The inescapable result of letting new technology trickle down throughout an
organization is:
When you bring in a new end-user technology,
the old one sticks around.
Like a pan of lasagna, there are horizontal
transitions from serving to serving (occurring horizontally along the adoption cycle from
pioneer to foot dragger.) There are vertical transitions between layers (occurring
vertically down the technology migration path.) Both of these transitions create costs
which must be accounted for and analyzed. Taking a “serving” highlights the
situation most organizations find themselves in. The cross-section shows the range of
installed end user computing systems which must be integrated, used and supported at a
period in time.
In this scenario, transitions are no longer one-time,
variable costs. The continuous nature of transitions make them a fixed cost, an
“annuity” cost of keeping up with the end user computing technology. At any
point in time, someone is going through a transition. A proper cost model must therefore
use transitions as its basis.
from TCO to transition costing
The most prevalent financial model used for EUC costing,
when a model is used at all, is a Total Cost of Ownership (TCO) model. Transition based
modeling better reflects the reality in today’s organizations.
Three year study life. “Everything we [Microsoft] sell will
be obsolete in three years.” Most IS professionals are loathe to forecast technology
beyond the three year timeframe. The variability of delivery dates, technology
developments, design changes, volatility of manufacturers, etc., impede an accurate five
year analysis of technology.
Numbers adjusted over time. Many factors is the cost model will
vary over time. This is not only due to the traditional cause, inflation, but also due to
technology and organizational factors. These will require adjusting cost category
“constants.”
Person based. The modern user may have three systems: work
desktop, work laptop and home desktop (when doing work at home.) While some costs accrue
by person and some by machine, a valid EUC cost model must treat the “person” as
the base unit of measurement – not the machine.
Transition of technology. No single technology platform today is
likely to carry a user through a five year life. A cost model must analyze which platforms
a user will have over the next three years and what transitions will be required to
“keep up” with evolving needs, capabilities and standards.
Heterogeneous installed base. Organizations do not have a single
technology installed for all users. The cost model must incorporate the intermixing of
installed technologies and the partial deployment of new technologies.
Activity based costing. TCO models use a traditional functional approach.
Activity Based Costing utilizes a process model for determining accurate and complete
costs of cross functional activities such as the deployment of personal computing in a
transition environment.
Best projections. Providing cost estimates for any future action is, by
nature, inexact. It is an error to count on exact totals from potentially hundreds of
estimated amounts. The error is compounded when analysts then project the result for an
installation with thousands of users.
So these are the major differences between the traditional Total Cost of Ownership life
cycle model and a Transition Costing model. Transition Costing focuses on:
Measuring
costs of technology over a three year life
Adjusting
cash flows to reflect reality over time
Making the
user the unit of measurement, and assuming that a user can on average have a multiple of
machines
Recognizing
that users are somewhere in the technology adoption cycle, that technology is changing so
fast that the study life will contain transitions beyond initial acquisition which will
greatly affect costs
Recognizing
that organizations are experiencing the “Lasagna Effect” – living with
multiple technologies in a heterogeneous installation
Estimating
costs with process oriented Activity Based Costing in order to recognize and compare all
costs
Using
results only within the precision of the estimates, making decisions which are within the
author’s recommendations, and which are realistically supported by data and not
stretched to make someone’s point
This model will give analysts a more accurate basis for making end user computing
technology decisions.
a transition costing model for EUC
Following are “management level” activity based
categories which can be used to measure the cost of each transition in a time sequence.
These include direct and indirect costs:
Long Term Planning covers all activities which identify decision points and action steps
for end user computing over the long term. This planning is part of the overall
organizational strategic planning and the IS department’s planning.
Decision Making includes all activities leading up to an actual technology acquisition
decision. This includes investigating best process practices, defining needs, researching
the installed base, researching the market, developing decision criteria, selection and
testing of alternatives, selection of configuration, and selection of vendor(s).
Planning/logistics activities are those which result in a tactical implementation plan.
This includes determining to-do’s, timing and responsibilities for the execution of
all the following transition activities.
Procurement refers to the actual acquisition of hardware, software and services. This
involves negotiating details of configuration, services, pricing and delivery, shipping
assurance, and processing order, invoicing, payment and reconciliation documentation.
Hardware covers all EUC equipment costs. This includes direct costs of end user computing
and networking infrastructure hardware and upgrades over the life of the study.
Software is similar to hardware. Direct costs of initial purchase and upgrade are
included. Costs of infrastructure software should again be part of overall software costs.
Deployment covers the tasks required to get the right equipment to users. For new
purchases, this means shepherding the equipment and software from the receiving dock to
the desktop. For transitions, deployment means storing and moving equipment and software
from place to place as it migrates from “pioneers” to “foot draggers.”
Installation is a significant category under “deployment.” These activities
include the initial installation of new hardware and software systems, continuing
installation of upgrades, and installation of moved systems.
Application conversion addresses the one-time activities in converting users to new
software titles or to new versions of existing titles. (Some organizations choose to place
software upgrade acquisition, deployment and installation costs here.) End user and
support staff software training is typically required.
File conversion accompanies application conversion. The quality of translation
capabilities varies greatly, with users having to modify translated files with the new
program or having to completely re-enter data and formulas or formatting.
User productivity/disruption occurs whenever changes are made. Communications and
scheduling takes time. EUC equipment is unavailable during the installation. There is a
period of relearning required for new system components and software. And there is a
reduction in productivity until the new capabilities are understood and used efficiently.
Support is a continuous need. This includes centralized support from the Information
Center, distributed support by department, purchased support from outside vendors, support
documentation, and informal support from co-workers.
Training - support can be a separate support category. This includes all tasks and
resources required to give support personnel the proper training required for the full
range of systems and software in place. All aspects of locating, evaluating, acquiring,
providing and paying for training are included. The physical assets and allocated costs of
equipment and space are included. Training costs must also include the cost of fill-in
labor or the opportunity costs of lost productivity.
Training end-users is another category which can be separated if desired. Similar to
support department training, this includes all tasks and resources required to give end
users the proper training required for the full range of systems and software in place.
Trainer costs cover the salary and overhead costs of trainer-employees and the contract
costs of outside training.
PR/communications includes the activities required in communicating with all individuals
involved in the management, administration and use of EUC. Any change causes a chain of
communications to flow up and down and back and forth throughout an organization.
Documentation is required for all transition activities. These activities include
information such as physical EUC inventories and location, change communications,
procedures and policies, workflow processes, scheduling and so on.
These categories focus on internal costs. Organizations can utilize either internal or
external costs as appropriate. The key is to not exclude any of the activities from the
analysis.
This section is not intended to be an exhaustive list of activity based cost categories.
The categories indicate how an activity based approach will help generate a thorough
estimate of costs in order to help the organization make the best EUC decisions.
recommendations
In this new technology environment, the old planning and
analysis tools, developed in an era of homogenous and relatively stable architectures, are
obsolete in an era of constant transition and a wide mix of technologies.
Information systems professionals and users must update their analysis tools and decision
processes. A plan must be put in place to develop a fact-based transition model:
1. Commit to upgrading the cost analysis philosophy and methodology.
Make the decision to invest in a new, more accurate costing approach.
2. Assemble a process improvement team. Form a cross functional
end user computing costing team. Include all the internal stakeholders: users, finance,
management … not just IS. Consider including external stakeholders (when practical.)
3. Perform a “best practices” review. Research the internal
best practices of the organization (such as plant floor equipment studies.) Review past
costing/justification studies to see if they delivered to projections. Work with the
accounting department to devise methods of determining “hidden” costs of EUC or
“indirect” costs. Challenge the methods and assumptions of industry data from
consultants. Verify their applicability to the current situation.
4. Determine accurate direct and indirect cost estimates (decisions based upon
data). Make decisions based upon objective evidence whenever possible. (In ISO
9000 parlance, “objective evidence is a factual statement which can be
verified.”) Challenge decisions made by individual fiat or because “everyone
knows …”
When feasible, do one-time studies to determine costs, such as estimating the number of
end users per support person by platform, or the cycle of help calls after a new
installation. Don’t allow estimated category costs to become precise totals.
Don’t allow industry cost estimates to become internal constants without validation.
5. Develop a transition based methodology. Create a projected scenario
for anticipated acquisitions and transitions over a time period (3 years). Cost out the
various alternatives and utilize the results in the acquisition decision making process.
Author Lawrence J. Peter of The Peter Principle fame has written, “If you don’t
know where you are going, you will get there.” He went on to observe, “And if
you don’t know where you are going, any road will do.” Implementing a transition
based costing model is a necessity for organizations that plan to “get where they are
going.”
top |
