Think about your five closest friends. What were your first impressions? Were they strikingly different from what you think of them today?

I first met my neighbor during freshman year of college through cacophony of sounds through the wall we shared, from visiting family speaking loudly in a foreign tongue to testing new ring tones and alarms on a new phone. I shot over a glare when I finally saw her in person, “so you are the one making all the noise!” We clearly got off on the wrong foot, but we became the best of friends.

Then there is the opposite phenomenon. Some people that I meet that impressed me as being confident and charismatic then after more interactions, I realized were at best provocative, at worst arrogant and disrespectful. In my eyes, the same qualities that enable someone to stand out from a crowd then would not police them when they go too far.

When I remember back to my first impressions of people I now know well, I often feel a jolt. Why the disparity? Why do I misjudge people?

This can be explained by machine learning

Machine learning is a technique whereby a computer can discover relationships between variables among data points in order to make predictions. The more data you feed the algorithm, the more accurate the prediction. However, you might need hundreds of thousands of data points for the prediction to be useful. Computers are machines that can make predictions if you feed it many many data points to learn from.

On the other hand, humans can be considered as a different type of machine. When a child encounters something new, she has the ability to understand and learn new things at a lightening pace, even though she has had very limited number of relevant data points in her experience. Humans are machines that can make surprisingly accurate predictions with very few data points.

The downside is that we are inherently biased and very bad at recognizing those biases.

When we first meet someone or encounter a new thing, our brain is wired to make snap judgments about what someone is like, even though we have only had seconds of interaction. We are likely to be fairly certain of our perception as well, hence the clichés about first impressions and love at first sight. In machine learning terms, this judgment, or prediction, is based on our previous experiences, or priors. Since you and I have different priors, even if we meet the same person, we are likely to have different predictions of what they are like. Even though our predictions are inherently biased, the irony is that both of us would still believe vehemently in our predictions.

If a computer, trained on a large representative data set, were to meet the same person, a few seconds of data would allow it to predict very little. But whatever it could predict would be unbiased.

Why does this matter?

It matters because we are asked to make many snap judgments every day, and not being aware of the inherent bias of how we are wired can mean lost opportunities. At the Harvard Business School, we often have speakers come to our case discussions to share their thoughts on the cases for 10 minutes, which never fail to illicit strong opinions from the crowd. From “so inspirational” to the adamant “I would never work for him!”, you would scarcely believe that they are talking about the same person.

But we are often wrong with our first impressions; our minds often change with time. As my professor wisely advised, “It is too easy to be impressed by the style and packaging of someone.” While we as humans are very good at making judgments based on little data, we are very bad at recognizing that our judgment are biased and what those biases are. The catch-22 of our biases is that once we recognize that we have a particular bias, then we can consciously correct for it and it will no longer affect your judgment. It is the biases that we are yet unaware of that are truly dangerous.

How should we fight our biases?

First, with more data points. This is why we read hundreds of cases during business school, to better train our pattern recognition system so that when we encounter a situation in a future job, we can recognize it from similar situations at other companies in other circumstances.

Then, there is conscious will. If you meet someone who did not struck you as exactly your type, do not close your mind to the possibility of befriending them but give it a big more time. Your mind machine will likely change. Even though we will still be affected by the biases we do not recognize, at least the magnitude should be steadily trending down.

It is important to not implicitly endorse behaviors that require quick judgments, in work and in life.

Don’t give in to the popular opinion. In business in America, what typically works is to be assertive and direct, but those are not all the ingredients for success. Don’t write someone off if they have a different presentation style. Instead, realize what biases you have and work to counteract those.

Don’t ask people questions like “what do you think about so and so” after only brief interactions and hold them to it. Don’t have your review system set up for people to evaluate others that they have scarcely worked with.

Don’t make investment decisions based on brief interactions. You may think that you know the archetype of the successful founder, but unless you are one of the top investors who have seen many many founders and learned over time how they perform, you are likely biased.

Just like you would not agree to get married to someone you have met only an hour ago, don’t hire based on 30-minute interviews, as you’ll be making decisions based on canned responses. Reflect on your hiring practices, are they systematically biased? A person is more than his résumé and sound bites.

People hire those like them. People invest in those like them. People befriend those like them. Don’t fall into this trap.

Instead, meet people not like you. Meet people that make you uncomfortable. Meet people from different cultures, of different beliefs, who grew up in different environments, who have views that may seem irreconcilable with your own. Open your mind. Give them a chance. You might just discover your stubborn biases.

Are you writing your mental program
or letting it dictate you?