Data-driven estimation the power of the Monte Carlo Simulation

Nearly thirty years ago I worked in the Construction industry and one of my roles was to create estimates for how long a future novel piece of work would take in time/effort; resources and then attribute a risk profile and lastly the margin of return known as gross profit. At that point we have something like ten different teams. I say teams. We called them gangs. Whilst some of the gangs were relatively stable others were not given the fluid nature of the jobs in the pipeline team members needed to move for short periods of time. I would often seek to bring gangs together in response to the size, complexity and milestone (payment dates) as back then much of the work we did was earned value given key milestones.

How did we grow a successful business? There were many key factors that made the family business distinctive: branding; attitude to safety, strong leadership; performance-related pay and bonus schemes; highest standards of training; professional accreditation and many more. But for getting our estimates right? Monte Carlo was key. We were a deeply data-driven and data intelligent business.

Stated simply, the Monte Carlo simulation furnishes the decision-maker with a range of possible outcomes and the probabilities they will occur for any choice of action.. It shows the extreme possibilities—the outcomes of going for broke and for the most conservative decision—along with all possible consequences for middle-of-the-road decisions.

Stated briefly, Monte Carlo simulation provides a number of advantages over deterministic, or “single-point estimate” analysis including:

  • Probabilistic Results. Results show not only what could happen, but how likely each outcome is.
  • Graphical Results. Because of the data a Monte Carlo simulation generates, it’s easy to create graphs of different outcomes and their chances of occurrence.  This is important for communicating findings to other stakeholders.
  • Sensitivity Analysis. With just a few cases, deterministic analysis makes it difficult to see which variables impact the outcome the most.  In Monte Carlo simulation, it’s easy to see which inputs had the biggest effect on bottom-line results.
  • Scenario Analysis: see exactly which inputs had which values together when certain outcomes occurred.  This is invaluable for pursuing further analysis.
  • Correlation of Input model interdependent relationships between input variables.  It’s important for accuracy to represent how, in reality, when some factors goes up, others go up or down accordingly.

Of course, you do need a decent understanding of statistics. To this day I am still very grateful for my practical training back then, as well as completing advanced statistics such as multiple regression, analysis of variance and mixed analysis of variance when I completed my MSc in Organisational Psychology at Cardiff University.

One of the most exciting prospects of the current online Kanban tool is that we are starting a journey of moving from guess-estimates (data neutral based on faulty human reasoning and bias) to a data-driven science of estimation. And delight of delights? This tool has an in-build Monte Carlo simulation! Yes…dreams sometimes can come true…

Have a good weekend,

Jason