The term “Unknown Unknowns” was made famous by the former Secretary of Defense Donald Rumsfeld in 2002.
Many people do not realize that this term was coined by the American psychologist Joseph Luft and is commonly used in NASA. It means the risk you have not planned for or know about is a “black swan event.”
Originally in the 2nd century CE, a “black swan” is metaphorical for a rare event. By the 16th century, the expression symbolized an impossible event until the Dutch explored and found real black swans in Australia. Since then, the term basically means an unpredicted event that was not planned for or an unknown unknown.
Unknown unknowns are one of four possible scenarios when planning or designing. Known knowns are what we are aware of and understand. Known unknowns are what we are aware of but do not understand. While unknown knowns are what we are unaware of but understand.
It might seem hopeless to try compensating for such events, but some methods can be used to make the system more robust in case an unknown event occurs. In his book “The Black Swan,” Nassim Nicholas Taleb considers making this robustness a necessity.
One method when planning or designing is to use a solution-neutral approach. This approach is based on “following the numbers” or “following the process” to elicit the solution. Normally, we have a preconceived solution in mind when we start solving a problem that opens us up to problems with making incorrect assumptions or using mediocre solutions which are normally outdated. By using a solution-neutral approach, we have a greater chance of accounting for as many possible events and making the system more robust and thus less susceptible to unknown unknowns.
Another way to compensate for encountering a black swan event is by using or building a more diverse team with different points of view and life experiences. Or by continuously designing for “good enough” and not for perfection and improving the design through iterations. Big data means using diverse data so that the AI or machine learning is not skewed to the point it is not robust enough to handle a black swan event.
However, for many, these methods could be difficult to implement. A simple method to increase your plan or design robustness is always to aim to have a final group of selected solutions using the 80/20 rule for their chances of success and then reevaluate them among themselves. This helps to reduce bias, improve the solution towards a solution-neutral approach, and, most importantly, increase the robustness of the final solution towards unknown unknowns.