The IT project took twice as long as planned.  The construction project to expand the plant went over budget.

Why does this always happen?

Too many decisions are made with too much self-confidence and too much optimism.  As a result, poor decisions are made and objectives are not realized.  The decision-making flaw in all these situations is that we do not consider the ‘base rate.’

The base rate represents the statistical average of what happened previously – what were the actual results for people who made similar decisions in the past.

To use the examples from the start of this blog post.  We may have known that similar IT projects take 12 months to complete, yet we scheduled it to be done in 4 months.  Similarly, all of our plant expansion projects have gone over budget by 30% or more.  But, this time was going to be different, and so we approved a tight budget that, yet again, we did not achieve.

The behavioral economist, Richard Thaler, describes well why we make such optimistic decisions and suggests a way out of this dilemma:

When the expert was thinking about the problem as a member of the project team, he was locked in the inside view – caught up in the optimism that comes with group endeavors – and did not bother thinking about what psychologists call “base rates,” that is, the average time for similar projects.  When he put on his expert hat, thereby taking the outside view, he naturally thought of all the other projects he’d known and made a more accurate guess.  If the outside view is fleshed out carefully and informed with appropriate baseline data, it will be far more reliable than the inside view.

To make better decisions and avoid our insider or optimism bias, we need to start out with the base rate.  This requires that we do some research to seek out evidence about similar decisions (as an added benefit, this allows us to learn or be reminded of the successes and failures of previous decisions).  Then, we need to soberly analyze our current decision in the context of the base rate.

If the base rate is X, what is different about this marketing campaign, new product development, project, etc. that gives us the honest confidence that it can be completed much more quickly and less expensively than X?

In conclusion, the science journalist David Robson says it well:

The simplest way to avoid bias was to start out with a ‘base rate’.

APPENDIX – The Base Rate in Medical Decision-Making

In medicine and psychology, the base rate is the percentage of a population that demonstrates some characteristic.  Unfortunately, ignoring the base rate is common in medicine today.  And it costs billions in wasted health care dollars and significant extra stress to patients.

Let’s give an example:

We go to our doctor for a physical.  We are feeling fine except for the usual aches and pains.  The doctor suggests that we take a test that can detect a certain type of cancer.  While only 1 person out of a 1,000 in our age, sex and general health category has this type of cancer (this is the base rate), the doctor recommends this test because it is 95% accurate (with only 5% of those tested receiving a false positive) and because it is better to be safe than sorry and catch the cancer early.

So, we take the test, and it comes back as positive.  Now, given this positive test, what is the probability that we have this cancer?

In one survey of Harvard University and Boston University doctors and medical students only 25% got the correct answer (see below).  In this survey and countless others, the vast majority of doctors will say that we have a 95% chance that we have this cancer.  So, we, as patients, have a huge fright as we begin the process of fighting this cancer that we nearly certainly do not have…

The correct answer is that even with the positive test result, we have only a 2% chance of having this cancer!!

Here’s how.  Let’s assume 1,000 people similar to us in age, sex, and general health take the test.  How many people would we expect to actually have this cancer?  In this case, according to the base rate, it would only be 1 person.  However, by having 1,000 people take the test and a 5% false positive ratio, we will have 50 out of the 1,000 people test positive (falsely).

In total, 1 person has the cancer; but 51 people test positive (the 1 person who has the cancer and the 50 false positives).   So, the chance that we are the one person with cancer out of the 51 people who tested positive is 1 / 51 or 1.96%.

The moral of this story is two-fold:

1. Always consider the base rate
2. Always get a second test or a second opinion