Algorithms Need Managers, Too summary
Companies today are turning more and more to computer algorithms, which perform step-by-step analytical operations at incredible speed and scale. Algorithms make predictions more accurate—but they also create risks of their own.
Algorithms have many shortcomings and the article provided many examples of instances where and an algorithm produces undesired results. Here is one of the better ones … “Another example comes from social media. Today many sites deploy algorithms to decide which ads and links to show users. When these algorithms focus too narrowly on maximizing user clickthroughs, sites become choked with low-quality “click-bait” articles. Click-through rates rise, but overall customer satisfaction may plummet.”
Sometimes algorithms are too literal. To get the most use out of an algorithm it must be used correctly and managed carefully. The article also refers to algorithms as “black boxes” that provide incomplete information with no explanation for their output. Example: “An algorithm can read through every New York Times article and tell you which is most likely to be shared on Twitter without necessarily explaining why people will be moved to tweet about it.”
Knowing that algorithms have their shortcomings allows us to use them more efficiently.
When using algorithms the manager must be explicitly clear. A manager can include “soft goals” in their algorithm to refine its results. Algorithms also suffer from myopia, or near sidedness. They seek to get their job done as fast as possible. If a manager is not careful, this might result in a dip in quality for the sake of speed.
When choosing the data inputs for an algorithm remember 2 things: the more data, the better and diversity matters.
Algorithms, when used properly, can improve a users experience. I discovered this first hand with the evolution of the “shuffle” option in iTunes.
I still use my first ipod (it has a wheel to navigate the library). For years I have been putting my ipod on “shuffle” and been amazed by the truly random mix of songs that plays. But things started to change when I began using my iphone as a music player; something seemed off. When I pressed shuffle on my iphone I no longer heard a random mix of smooth jazz and death metal. Instead, I noticed that my iphone was playing the music I wanted to hear, and that freaked me out. I would catch myself wondering, “How does my phone know I want to listen to Norah Jones?”
Well after a bit of research my fear of the Skynet takeover was put to rest. It turns out the “shuffle” option on newer apple products is no longer truly random. Instead, the ipods in our phones use algorithms to identify the type of music we have been listening to lately. Our phone then takes this information to create a “random playlist” that is not random at all. In fact, it is tailored specifically for us based on our listening trends. My phone knew I wanted to listen to Norah Jones because I have been listening to her all week.
Go figure. ¯\_(ツ)_/¯