On data, piña coladas and dating…


pina colada“Yes, I like piña coladas and getting caught in the rain,
I’m not much into health food, I am into champagne,
I’ve got to meet you by tomorrow noon and cut through all this red tape,
At a bar called O’Malley’s where we’ll plan our escape”
 Escape (The Piña Colada Song) Rupert Holmes

These days, we are more likely to join an online dating site, but Rupert Holmes’ 1979 hit single typifies an approach to dating by data.

The principles of data-driven dating are simple:

  • Each party completes two profiles.
  • In one, they describe their ideal mate. In the other, they describe themselves.
  • The dating system then performs some analyses and comes up with a short list of potential “ideal” matches.

Poor data quality may, of course, affect the results.

One or more of the parties may lie (or omit some facts) from their profile. I may not want to disclose a previous marriage, a failed business, or some other reality that I fear may chase potential partners away. Without full disclosure can the analysis be accurate?

I may be too picky – my ideal mate may not exist.

The dating analysis must choose which filters to apply, and which to discard. In analysis we often ignore outliers – those bits of data that seem to buck the trend. In dating, does this mean that we end up with an incompatible partner. In business, with a poor decision?

Finally, does data-driven dating remove emotion from the decision-making process?

In theory these matches are uniquely qualified – each result should be a good match. In practice, some parties may never find a match. Others may hit it off on the first date and move into a long-term relationship.

Statistics show that profiles with a picture get substantially more hits than profiles with no picture – irrespective of the content in the profile.

It gets more complex than that.

According to research from onlnie dating site OKCupid  (based on analysis of their 3.5 million active users):

  • Men who look away from the camera and don’t smile have a much higher chance of getting a response than those who look directly into the camera.
  • Unlike men, women do better when they look directly into the camera.
  • Women in their early 20s who show cleavage in their profiles do around 24% better, but that number rises to 79% by the time they hit age 32.
  • Similarly, men who show their abs do better—but men only show their abs when they have really nice abs.

The bottom line?

Most of us would ignore an otherwise “perfect” profile if we didn’t like the picture provided.

Emotion remains a key decision criteria in dating.

We cannot explain why two people who seem completely incompatible may hit it off, at two people with many interests in common may not be able to stand each other

Similarly, emotion plays a role in business decision-making.

Quality data can help to influence and inform decisions, but it will never be the only factor.

Information helps us to short list and focus our energies on a smaller set of choices. It will seldom make that choice for us.

Advertisements

One thought on “On data, piña coladas and dating…

  1. Hi Gary,

    This goes to reinforce the well known fact that decision support systems are a long way from reaching their nirvana. Guess that is why they are called decision SUPPORT (sorry for the caps – no other way for emphasis) system and not decision MAKING systems.

    Decision making systems, devoid of emotion, abound. Cue: driving aids and controls found in cars.

    But, still, the question remains. Can we ever create a system that can make “intelligent” human like decisions – with or without emotions? The answer resides in the field of AI.

    But before run-of-the-mill businesses start investing in full-blown AI systems, there is cognitive analytics that aims to bridge the gap. There is also a system that aims to eliminate discretionary (typically emotional) decision making. For example, what makes a cop (police officer) stop and question one person and not the other? Or a company employing one of two equally appealing candidates? Or awarding a contract to one and not the other company? Or decide to invest in this project and not the other project – given all else (data) being equal?

    Whichever way this journey pans out, data remains supreme and its quality matters.

    Bowie

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s