Why Predictions Fail

The new year is upon us and with the changing of the calendar inevitably comes the raft of predictions for the upcoming year. These great soothsayers opine on everything from economic data, stocks, consumers trends or the latest fashions. Most of these are riskless predictions — except perhaps for one’s reputation — as being wrong brings no financial cost.

Businesses, however, must make predictions all the time, and being wrong can bring real financial cost from a failed project or product launch. So, why do predictions fail?  The primary cause is ignoring available information, leading to overconfidence.

Making Bad Predictions

In his book, Thinking Fast and Slow, Daniel Kahneman describes how when teams are charged with making predictions they often rely on what he refers to as an “inside view,” where the team bases their estimates solely on their “particular circumstances.” Curiously, even when individual members of team possess “outside” knowledge about failure rates of similar projects, they still tend to be overly optimistic about the current situation. In the book, he describes how a group of professors were collaborating on a book project and were asked at that the outset of the project how long they believed it would take to complete. They predicted the project would take around two years, and no one even contemplated the probability of failure. After these confidential predictions were tabulated, Kahneman asked one of the professors who had the most experience with these types of projects in the past, what he had learned. The professor pondered the question and responded that 40% failed entirely, and the successful ones took seven years to complete. As most teams do in these situations, they ignored this evidence and dove merrily into the book project — completing it eight years later.

The Planning Fallacy

What is it a play here, according to Kahneman, is the Planning Fallacy at work. This is a condition when plans or forecasts are:

·       Unrealistically close to the best-case scenario, and

·       Could be improved by consulting the statistics of similar cases

In the book project above, you could imagine a range of outcomes where completion in two years was likely among the best-case scenarios, but it’s what the team agreed was about right. What’s required, according to Kahneman, is an outside view that provides base rate information from similar situations.  Without this information, decision-makers will anchor their expectations on the initial forecast, making ill-informed and sometimes spectacularly wrong bets. Leading up to the collapse of the financial markets in 2008, data on mortgage default rates were readily available but promptly ignored, failure was not contemplated, except for the very few who spotted the trends (i.e. deviations from the base rate) and shorted the market.

Finding Your Base Rate

As you make your revenue and budget projections or contemplate funding new product development, you should begin by challenging your underlying assumptions about the range of potential outcomes. Your initial prediction will serve as the anchor for future adjustments, so if the initial predictions are wildly optimistic, your adjustments going forward will likely be frequent and substantial, creating financial as well as credibility issues with your team and stakeholders. So, look for similar “cases” that match your situation, and find that data that will help inform your estimates as to a range of outcomes. Using this data, note the best-case scenario and challenge yourself not to fall in love with it. Taking an “outside view”, compare your historical data (sales, profit margins, product cycles, etc.) to similar organizations, project types and look for deviations and patterns, not averages.

The point is not to squelch optimism or risk-taking, but that you need to be mindful of your decision-making methods. Take an outside view, avoid making the planning fallacy, and seek out your base rate data. 


Photo by Caleb Ekeroth on Unsplash