Newsletter December 2018

We hope you enjoy the December 2018 issue of “RiskIntegral” - an occasional newsletter about the analysis and management of risk, mainly in projects, by Risk Integration Management Pty Ltd (RIMPL). In this issue, we report on our development of a hybrid methodology for assessing time and cost contingencies and our experience using it on several projects.

We invite you to donate to support a young project controls engineer and his family as he fights cancer.

We end the newsletter with news of our activities during 2018.

Optimising and Realistically Quantifying Project Risk

Introduction

Since 2008, the RIMPL team has been using and developing its methodology to quantify project schedule and cost uncertainty to help clients assign duration and cost contingencies for their projects. The fully Integrated cost and schedule Risk Analysis (IRA) methodology developed by the RIMPL team based on the critical path method is founded on two facts that we have observed:

1. The equivalence of time and cost in projects - “time is money” – with cost consequences of delay driving most of the cost overruns in projects, particularly large and complex projects; and

2. The high proportion of project schedules that are poor predictors of project time outcomes, due to a combination of failure to represent adequately the way the project is to be executed and poor technical quality.

The RIMPL team has enjoyed success in helping clients optimise their schedules and in forecasting more realistic cost and particularly time outcomes for their projects. There have been numerous examples where the systematic IRA and SRA processes have enabled more realistic schedule forecasts to be provided than those predicted by the project plan. Schedule drivers have also been highlighted enabling logic changes to be made, saving time and money.

However, the goal of realistic spreads of cost being forecast has proven more elusive, with narrow distributions and the long “tail” to probabilistic cost distributions, characteristic of large and complex projects, typically missing.

This article describes how RIMPL has overcome this problem by developing and using a Hybrid Methodology combining Parametric modelling of Systemic Risk with Integrated cost & schedule Risk Analysis (P+IRA)

2017: First Attempt to combine Parametric & IRA Methodologies

Following publication of John Hollmann’s seminal book in 2016 “Project Risk Quantification”, subtitled “A Practitioner’s Guide to Realistic Cost and Schedule Risk Management”, summarised in last year’s newsletter and reviewed at length in our December 2016 issue of RiskIntegral, I presented a paper at the 2017 AACE DRM Methodology Conference which incorporated our first attempt at integrating Parametric assessment of Systemic Risk.

The paper, titled “Modelling Realistic Outcomes using Integrated Cost and Schedule Risk Analysis”. This paper reported on an IRA model of design and construction of a tollway. It included the use of parametric modelling of systemic risk in time and cost as included in Hollmann’s Project Risk Quantification book and modified by RIMPL. It took the full schedule and cost effects of systemic risk, converting the corresponding uncertainty distributions into risk factors and applied them to the IRA model. While this worked satisfactorily and overcame the problem of negative optimistic duration values that we identified in the hybrid methodology described in Appendix A of Hollmann’s book (combining David Hulett’s Integrated Cost-Schedule Risk Analysis methodology described in AACE RP 57-09 with Hollmann’s Parametric modelling), it uncovered a more serious problem: how to retain duration and cost ranging of inherent risk without “double-dipping”?

The IRA methodology uses duration and cost ranging to quantify inherent risk. In combination with mapping of treated risk events with duration impact, it enables the riskiest pathways through the project schedule to be identified and optimised, saving time and therefore time-dependent costs. Hollmann’s Appendix A method strips out all such ranging because, he argues that it is already included in the Systemic risk quantified by the parametric modelling. While his “estimator’s solution” solves the double-dipping problem it also throws out the schedule and time-dependent cost optimising advantages of IRA methodology plus the significant advantages of quantifying and optimising trading treatment cost for probabilistic schedule (& time-dependent cost) benefit that is part of the IRA methodology. I felt that there had to be a way to combine Parametric and IRA methodologies that did not throw out “the (IRA) baby with the (double-dipping) bathwater”.

Early 2018: Mulling over the hybrid methodology problem

While we dabbled with adding Systemic Risk to IRA models in the first half of 2018, I was not convinced by Dr David Hulett’s assertion that systemic risk was like contingent risk events and could have a probability of less than 100%. This was in spite of my great admiration for and emulation of David Hulett’s work, which includes two highly influential books “Practical Schedule Risk Analysis” and “Integrated Cost-Schedule Risk Analysis (both published by Gower).

The research John Hollmann describes, stretching back 60 years, by John Hackney, Rand Corporation and IPA, is persuasive that systemic risk is an inherent consequence of the (primary) project delivery systems as well as other (secondary) systems in which the project is delivered such as the economic system, the market systems providing materials, equipment and services, etc. It is not possible to turn such systems on and off. They exist and the uncertainty they bring is in their variable impacts. Such impacts are quantifiable based on substantial investigative work and correlation of various possible drivers to find those with the strongest correlated effects. These are known to be led by scope definition, acceptance of which led to the progressive definition of scope by the stage gate system familiar to all who have worked in the project development environment. Other known drivers of systemic risk include the calibre of the project delivery organisation and the quality of the project controls.

Mid-2018: The Age of (Hybrid) Enlightenment

In June 2018, John Hollmann visited the east coast of Australia, as described under RIMPL News. John ran a three-day training course in Melbourne in Project Risk Quantification, based on his book. Along with my co-director colleague Peter Downie, I attended the training. Before I participated in the training, John invited me to give a brief presentation to the participants on our IRA methodology. In the course of preparing, I thought about addition and subtraction of probabilistic values and recalled that only Pmeans can be added and subtracted; P50s can be only if the distribution is symmetrical (i.e., P50 = Pmean), but other P values cannot.

Subsequently, I realised that one probability distribution can be subtracted from another by use of Monte Carlo Simulation.

The key to this is understanding the relationship between Systemic Risk as quantified by the Parametric Method and Inherent and Contingent Risk, as quantified by the IRA methodology. This is illustrated by the accompanying Venn Diagram.

Project Delivery Systemic Risk normally completely overlaps schedule and cost estimate uncertainty (inherent risk) and partially overlaps risk events in the project risk register (those risks that are systemic in nature), leaving only the major project specific risks. The systemic risks and minor project specific risks are covered by the Parametric assessment of Systemic Risk.

This led to our being able to incorporate Parametric assessment of systemic risk into our IRA methodology and to offer it to our clients.

Second Half of 2018: Providing P+IRA assessments to clients

RIMPL has so far provided hybrid Parametric + Integrated cost & schedule Risk Analysis services three times in 2018. The clients, sectors and scale of the three projects cover a wide range:

Client A is a major Australian pharmaceutical company building a substantial new facility on a site on which it has built a series of successful expansions in the last few years. Their board has mandated the use of QRA to assess contingency and this aspect is new to the project team, unlike most other aspects of the project. Consequently, they have embarked on a thorough exploration of what a QRA assessment of schedule and cost contingency entails to be well prepared when the actual QRA is performed, including conducting close to a full dress-rehearsal in November/December. As might be expected from a pharmaceutical company, they demonstrate a high degree of professionalism and competence in the areas in which they have prior experience and are conducting the QRA at a later stage of design completion (~60%) than is usual in the chemical process industries (~30%). The P+IRA process has proceeded reasonably well from RIMPL’s perspective, with project controls being the weakest link in the chain. Although not strictly a megaproject (where capex > $1bn), the complexity of the project and the pharmaceutical environment more than qualify it for megaproject status.

Client B is building a mineral processing plant in south west Western Australia. This is a megaproject with an EPCM contractor to the project owner. The owner is a world scale producer but this is their first megaproject and their experience elsewhere makes them concerned about the time and cost outcomes of this project. The schedule used for the modelling was large (>4,400 normal activities) and extensive use was made of cost and duration risk factors to enable the ranging of the schedule and estimate to be completed.

Client C is the owner of a small natural gas processing plant in southwest Victoria carrying out an expansion of the compression capacity of the plant, along with parallel sustaining capital works. The project is driven by commercial negotiations with a customer and the negotiations have had a significant impact on the scope of the project. This being a brownfield project in a well-understood site, the combining of the Parametric and IRA methodologies, led to the unusual situation of the Inherent Cost risk in the base IRA model due to ranging of the cost inputs exceeding the Parametric assessment of Systemic Cost uncertainty. Referring to the Venn Diagram, this implies that the solid orange Inherent risk area is larger than the translucent Systemic Risk area, at least for cost. The reverse was true for the schedule uncertainty, as is normally expected. We therefore assumed that time-independent cost uncertainty was fully covered by the IRA base model cost distribution so no Systemic Cost Risk Factor was required. However, the Parametric schedule uncertainty distribution exceeding the IRA base model schedule uncertainty distribution meant that we had to subtract the two distributions by Monte Carlo Simulation and assign a Net Systemic Duration Risk Factor to the schedule activities capable of driving the key schedule milestones. This also meant that time dependent costs were also driven by net systemic risk. The P+IRA modelling provided a basis for quantifying both cost and schedule contingencies and the process also helped the client optimise the schedule for determining the timing of a shutdown to enable subsequent installation and commissioning of additional compression capacity.

General Comments

For Clients A & B, Net Systemic Risk was the dominant cost risk. Unsurprisingly perhaps, that was not the case for Client C, where weather was the biggest cost driver.

Net Systemic Risk was the biggest driver of schedule delay for all three projects.

In general, Net Systemic Risk provided greater spreads, making the models both more optimistic and more pessimistic. This provided more credibility to the analyses which otherwise would have had narrower time and especially cost distributions.

In closing this article, it is pleasing to us that John Hollmann’s preferred hybrid methodology - Parametric + Expected Value (P+EV), which requires no project schedule, is proving to be of increasing interest to clients in Australia, particularly for Sustaining Capital and Maintenance projects. Wherever a shorter duration, lower cost contingency assessment is attractive to clients, P+EV methodology is appealing.

Raising Funds for Claudio Olivieri

Claudio Olivieri, a very active and positive member of the project controls community and the Australian Cost Engineering Society (ACES), has been diagnosed with Acute Lymphoblastic Leukaemia (ALL).

Claudio has a lovely wife (Romina) and two cherished daughters. The goal of this campaign is to raise funds that can help support Romina and the girls through this very tough time. Funds may be used to assist with additional child care, babysitting, petrol, parking, and any one of the many costs that come with being sick.

Claudio is one of the most positive people you could meet, but he and his family are facing a difficult time, made worse by the fact that migrated to Australia relatively recently and do not have the support network of family that most of us can call on.

Claudio needs the support of colleagues and friends. Please use the following crowd-funding website to help Claudio and his family: https://www.gofundme.com/fight-it-claudio

RIMPL NEWS

Quantitative Risk Analyses

Since our last RIMPL Newsletter, the following activities in 2018 have kept us active in our core business:

• An IRA was performed for a marine infrastructure project through Argonautica in the first half of 2018.

• The three P+IRA assessments described in the feature article were performed in the second half of 2018, with further services for two of these clients expected to occur into Q1 2019.

Project Controls & Risk Management Services

We provided project planning, cost control and risk management services for infrastructure projects:

• Provision of risk management services to the Victorian Government for the High Capacity Metro Trains project through 2018 via service provider Indec Ltd.

• Provision of planning, cost control, contract and risk management services for a marine infrastructure project being managed by Argonautica Pty Ltd for a major logistics company up to the middle of 2018.

RIMPL Support of visit of John K Hollmann to Australia in June 2018

In June 2018, John Hollmann visited the east coast of Australia and spoke at Australian Cost Engineering Society (ACES) / AACE Australian Chapter meetings in Melbourne, Sydney and Brisbane and addressed gatherings of senior project executives in those three cities. We were involved in supporting the visit. Managing director Colin Cropley accompanied John Hollmann through much of his visits to the three cities and participated in the training course conducted by John in Melbourne, along with co-director Peter Downie.

Responsibility for organising and funding the trip was taken by the ACES National Committee, of which Colin Cropley is a member. Bouquets to the ACES volunteers in each city who put in sustained efforts to arrange meetings for John with senior executives of various companies and government organisations. Special praise for John Paterson, Victorian ACES Chapter Chairman, who organised the training venue at Engineers Australia Victoria Head Office and who organised a breakfast meeting for John at EA Victoria and a meeting with the most senior executives of the Victorian Government’s infrastructure authorities.

Papers, Presentations and Conferences

• In late April/early May at the Conference on Railway Excellence (CORE 2018) held at the Sydney International Conference Centre in Darling Harbour, Colin Cropley presented a paper written by Peter Downie and Colin on the topic “The need for significantly improved accuracy in forecasting rail project time and cost outcomes”

• On 22nd November, RIMPL exhibited at the inaugural Project Controls Expo at the Melbourne Cricket Ground (MCG) as a Silver Sponsor. Manning the stand were the three RIMPL directors, Matt Dodds, Peter Downie and Colin Cropley. The stand, featuring a company banner with the caption “What is Systemic Risk and why does it matter?”, was visited by a good number of the more than 500 delegates who attended the Expo, well-organised by Project Controls Online as an extension of their UK-based PC Expos, which are attended annually by more than 1,000 delegates.

• By invitation of the organisers, Colin Cropley presented to a Megaprojects forum on the subject “The problem of forecasting realistic time and cost contingencies for megaprojects”.

CONTACT US

Risk Integration Management Pty Ltd
18 Cottesloe Court
Doncaster East VIC 3109
AUSTRALIA
+61 412 031 161
[email protected]