The Problem with Predictions

The James Webb telescope lifted off from South America’s lush tropical Atlantic coast on December 25th, 2021. This was a historic moment for humanity, with James Webb being one of the most grossly over-budget and over-schedule projects ever completed. What was originally expected to cost $1 billion and launch in 2010 was finally put into operation a decade late, with $8 billion in additional costs.

?This project’s performance can hardly be called an exception. Its costs and duration are eye-raising, but the overrun ratios are, unfortunately, the norm. Governments seem to weather these overruns fairly well, but can your career or business? Take a moment to think about your own professional experiences. When was that merger to be completed? When was the new product scheduled to launch? What was the original estimated cost and schedule for that new information system implementation?

?

Why do we keep missing our forecasts, and can we do anything about it?

?Answering “Why do we keep missing?” requires a deep dive into human psychology, politics, and statistics. A complete exploration of the “why” requires much more than an article-length treatment. For those interested in learning more about the “why” question, I recommend starting with this HBR article and the book, “Thinking, Fast and Slow” by Daniel Kahneman.

?Can we do anything about it??Absolutely.?There are many things that executives, managers, and planners can do to immediately improve their forecasts. Here is the quick-start action plan that is accessible to everyone:?Be Skeptical, Use Tools, Communicate Visually.

?

Be Skeptical…mostly of yourself.

?“The first principle is that you must not fool yourself—and you are the easiest person to fool. So you have to be very careful about that.” – Richard Feynman

?A great starting point for building a productively skeptical mindset is to print the “Cone of Uncertainty” and hang it on the wall next to your computer. This graphic helps us most when we plot ourselves (honestly) on the chart. If we are relatively early in the planning process, then we will very likely be underestimating costs and schedules by a factor of four to ten. There is also a slight chance that we are overestimating, but this rarely happens. We should be very skeptical of early estimates and?use multipliers to improve our cost and schedule predictive models. Use the inverse of these multipliers when predicting benefits.

?

Use Tools.

?A risk register is one of the most effective tools for improving planning and forecasting. Risk analysis and risk workshop activities support the risk register. Here are five steps that anyone can use to integrate risks to improve their forecasts.

1. Document assumptions?– Assumptions are things the planner assumes will be true in the future. For example, suppose a car manufacturer plans to deliver a car order in November. In that case, the planner must assume that the factory will receive processing chips for those same cars in July. But the future is unknowable, so these assumptions are risks.
2. List assumptions as risks in the risk register?– Risk is the term that professionals use when describing any event that has a?probability?of occurring. If it occurs, then it will have a positive or negative impact on the forecast. Treating assumptions as certainties is a recipe for disaster. Don’t fool yourself. Be honest and explicit about assumptions and add them to the risk register.
3. Quantify the risks?– The planner must establish numerical values for how the risk could impact the cost and the schedule. There are some simple options for quantifying risks and some very complex ones. If you are new to quantifying risks, then simple methods are strongly recommended and still very effective.
4. Decide how to manage the risk?– Though you cannot completely control whether a future event happens, there are likely things you can do to make it more or less likely. The actions you take to nudge the risk in the direction you want it to go (before it happens) are your proactive plan. The action you will take if/when the risk happens is your reactive plan. It is critical to remember that creating a plan is not enough. A plan must be executed to be effective.
5. Add management and contingency reserves to the estimate?– Contingency reserves are for known unknowns. These values are derived from the risk register and can include potential cost and schedule impacts associated with known risks. Management reserves are for unknown unknowns. If you have any experience planning anything, then you understand that it is impossible to conceive of every possible event that could occur in the future. Still, you know that something unexpected will likely happen. Contingency reserves provide a buffer for those unexpected events.

?

Communicate Visually.

?Now what? Pull it all together and present it to your team! An excellent graphic to utilize is the cone of uncertainty. Many variations are available, and they can be easily tailored to meet your needs. Some supporting graphics that may further enhance the presentation include:

• risk matrix
• risk RACI charts
• project budget and schedule charts that depict reserves

?

Communicating visually?enables a diverse team to form a shared understanding of potential future outcomes and see how their actions will influence the actual results.

??

If you are interested in digging deeper, then here are some resources to explore:

• Planning Fallacy – Why do we underestimate how long it will take to complete a task?
• Optimism Bias – Why do we overestimate the probability of success?
• The Standard for Risk Management in Portfolios, Programs, and Projects
• Practice Standard for Scheduling