I joined PwC a year ago in a team we call ‘Disruptive Analytics’. Before that, I spent 5 years at a large mining company as it was transitioning out of the commodities boom; the mantra changed from “production at ANY cost” to “production at the RIGHT cost”. During this time, I worked in supply chain integrated planning, reliability engineering, strategy & planning and project management roles at a number of sites across the Pilbara region. Throughout the business there were a plethora of models owned by various planning teams, which resulted in a highly bureaucratic organization. These planning silos caused initiatives to be implemented that caused inefficiencies, waste and rework.
2. What were some of the first modeling and optimization tools you learned? (If you began on EO, please let us know what other planning tools you developed / used in the past.)
At University, there was EViews, Simulink, Matlab and Solidworks. In the mining company, there was an internal Excel-based supply plan optimizer, an internal simulation model in ARENA, and an external simulation model that was considered a black box.
The technology did not exist to quickly model questions of resource allocation for the world’s largest privately-owned rail network. Therefore, it was accepted practice to build a local assessment model and then make assumptions about flow on effects. Hence, I developed a tool in R that used predictive analytics to quantify the benefit of alleviating constraints around the network. This was before I came across River Logic’s prescriptive analytics capability.
3. What did you find most frustrating about working with those tools?
Due to the rapid expansion of the supply chain, the model owned by external consultants was constantly made obsolete, and therefore, the turn-around time for results was not sufficient for tactical level questions. Once the questions were answered, the changes to the supply chain rendered the results redundant.
To make things worse, the internal AREANA simulation model was only accessible by a very small team of analysts, therefore this created political issues, and resulted in a large backlog of work and numerous silo models developed.
Lastly, only one person understood how the supply chain optimization model worked; hence, the planning process of a $16 billion revenue business was reliant on one man.
4. When and how did you first come across River Logic / Enterprise Optimizer®?
While interviewing for PwC, my future director dropped hints about River Logic’s Platform. I contacted River Logic and was sent a trial with training materials. Once completed, to consolidate the newly acquired skills, I was able to model the supply chain of a mining giant in the weekend before joining PwC. This allowed me to ask EO the burning list of questions that had always been a matter of debate over the years. It was good to finally have the answers.
In a short period of time, the model allowed me to easily switch the objective function from maximizing revenue to maximizing profit. This reallocation of the network had the potential to identify tens of millions of dollars of unnecessary costs. Furthermore, due to the opportunity value functionality, I was able to remove conjecture by quantitatively assessing the global impact of local improvement initiatives.
5. How long did it take you to get value out of EO compared to other solutions you’d learned?
Other solutions required large investments in development, and even larger investments in change management. The user friendly, visual capability of EO addresses this issue by allowing stakeholders to see the problem in a logical combination of inventory, conversion, source, sink and financial objects. Every client has new problems to be solved and we have yet to come across one that could not be modelled in EO.
6. How is modeling in EO different from modeling in other solutions (i.e., what specifically makes it easier / faster)?
EO gives me the confidence to walk into any client in any industry with the capability to get a working model up and running within hours. Every business has objective functions, constraints, choice variables and non-negativity criteria. Hence, even if the platform is not open in front of you, EO trains users in a school of thought that allows a logical approach to any business problem.
7. How is a business uniquely represented versus other modeling solutions?
EO has a seemingly magical ability to bring stakeholders from all over a business to discuss and agree on how their business should be modelled. It is scary when marketing, finance and operations teams are unaware of each other’s accountabilities — but it still happens. The throughput KPIs of operations teams does not align with the ROI requirements for finance, hence it is impossible to maximize long term growth of a business. Because operational activities can be mapped to financial implications, EO allows the client to embark on a journey to generate solutions to problems that have a global optima.
8. What sort of background do you need to learn EO versus other programming language-based solutions?
Unlike the programming background requirement for the optimization tools I have seen, there is no Java scripting or PhD in Optimization needed to build an EO model. Testament to this is that we have a former nurse modelling an energy company, a former physio modelling a public transport network, a law graduate modelling a mining company and an accountant graduate modelling a fast-moving consumer goods business.
9. What are the primary use cases PwC is targeting with River Logic?
EO is a planning game-changer because it is one component in a suite of applications that allow a client to see the bigger picture. Coupled with IoT, machine learning, cloud software, visualization and collaborative planning tools, we’re able to provide an end-to-end solution that can be tailored to any clients forecasting, planning and scheduling pain points.
10. How many EO models have you authored or assisted on over the last year?
My EO-cyclopedia library has grown to contain 16 volumes covering the beverage, banking, mining, shipping, energy, airline, rail, public transport and defense industries. Lately, the demand for client work and development has grown to such an extent that I cannot be across all of these industries alone. Hence at PwC, the number of modelers has grown from two people a year ago to a pool of 34 across six locations. A hands-off approach allows me to be part of 5 model development initiatives at once.
