Chemical Engineering December 2014 - 37

Cover Story
Feature Report
Improve Your Contingency
Estimates for More
Realistic Project Budgets
Reliable risk-analysis and contingency-estimation practices help to better
manage costs in CPI projects of all sizes
John K. Hollmann
Validation Estimating, LLC
T
he average large project in the
chemical process industries
(CPI) overruns its sanctioned
budget, including contingency,
by 21%. Even more alarming, 10%
of large projects overrun their budget
by more than 70%. These are
findings based on research by the
author and others. However, research
also tells us that small projects,
such as those managed at the
plant level or those with little individual
impact on company profit,
tend to underrun their sanctioned
amount, often significantly. This
raises many questions regarding
the techniques used to improve estimates
of the contingency required in
these situations.
The wide accuracy range of CPI
projects indicates that there is tremendous
uncertainty for owner
capital projects of all sizes - much
more than accepted rules-of-thumb
may indicate (for instance, assuming
that ±10% is considered " typical " ).
The overruns for large projects
indicate that owners do not understand
the risks and are not managing
them well, or are biased to underfunding
the risk because proper
contingency might kill the project.
Conversely, small project systems
understand that there is variability
in project costs, but find that the
most expeditious way to keep moving
forward is to have a cushion in
each project budget to avoid " imperial
entanglements " (punitive reaction
to overruns).
Because each project is unique,
it is often hard for an engineer or
project leader to understand appropriate
costs and to see cost biases
(often their own) at work. Therefore,
to improve the pricing of risks
in their estimates and budgets,
owners must address both technical
and behavioral (bias) components
of risk analysis and contingency
estimating. Engineers responsible
for estimating budgets or managing
projects must sharpen their
knowledge of risk and its quantification.
The first step is to fully understand
the concepts of risk, contingency
and accuracy as applied to
cost-estimation principles.
Risk, contingency and accuracy
For projects, risk is synonymous
with uncertainty. International
Organization for Standardization
(ISO; Geneva, Switzerland; www.
iso.org), Project Management Institute
(PMI; Newton Square, Pa.;
www.pmi.org) and Association for
the Advancement of Cost Engineering
(AACE) International (Morgantown,
W.Va.; www.aacei.org) standards
agree that anything (threats
or opportunities) that might cause
project outcomes to differ from plan
is considered to be a risk. Contingency
and reserves are how risks
are funded in project budgets. Es36
ChemiCal engineering www.Chemengonline.Com DeCemBer 2014
calation and currency exchange are
additional ways, but are outside the
scope of this article.
A cost estimate includes two
main parts: the " base " (essentially
the risk-free costs other than specific
allowances) and the " risks "
(contingency, reserves, escalation
and currency). Separate risk types
are necessary, because there are
different best practices for estimating
and managing them. However,
engineers should keep in mind that
these risk types are not always easy
to parse cleanly.
Accuracy, for cost-estimation purposes,
is a shorthand term to describe
estimates as predictions of
uncertain outcomes - the estimate
is a range of possible costs, not a
discrete " number. " In that sense, a
good estimator is one who knows,
and is honest about, what he or she
does not know. The person preparing
the estimate must communicate
the risks, the range of possible
outcomes and, most importantly,
how those two go together. Accuracy
ranges are expressed as low
and high percentages that bound
the expectation of how final actual
costs will differ from the estimate.
For example, a range of +30%/-10%
tells management that the final
costs may be as much as 30% more
or 10% less than the estimate after
taking into account all of the identified
risks. For a range statement
to be complete, one must also state
the confidence in this range; for in
http://www.iso.org http://www.pmi.org http://www.aacei.org http://www.Chemengonline.Com
Chemical Engineering December 2014

Table of Contents for the Digital Edition of Chemical Engineering December 2014
