After a slew of dehumanizing experiences with my cable provider I cut it off and went rogue. I purchased two devices for my home entertainment, Chromecast and a Roku player. Between the two devices I have access to games, live TV, and just about anything I want to watch. I am unstoppable!
Last Saturday night, I fell asleep in front of the TV watching the news. My Roku has the ‘continue to play’ feature turned on in Hulu. This means when one program ends it suggests and then auto-plays another show for you. Thinking nothing of my slumber, my Roku went to work choosing what it thought I wanted to watch. When I woke up, I was astounded by what was chosen for me! For some reason I felt violated when I saw what the robots seemed to think I wanted to watch. The bottom of the barrel was Mob Wives and Hot in Cleveland (although who doesn’t love Betty White?).
Note the middle column where it says ‘shows you watch’.
Maybe this selection doesn’t say much of anything about me. So why was this so upsetting?
The suggestions Hulu provided me with previously were actually pretty decent. I’d discovered some new and beloved shows from those recommendations. However, I was truly upset at the selection of shows and subsequent predictions that were made for what I like. What it comes down to is that most people do not like being categorized or being made to feel like a number. Predictive analysis will evolve and we will adapt. While all of these algorithms are designed to help customize to our specific tastes, it’s far from personal. This isn’t the first time suggestions were made for me. Netflix has a similar method for suggestions:
Oooh, I’m dark and exciting! Much better.
I believe that in modern society we have to learn to accept a few things:
1. Our notion of privacy is going to continue to change over time
A few years ago I saw the legendary futurist Ray Kurzweil speak on the topic of privacy and technology at SXSW. He was very clear that what we see as the “line” for privacy today will continue to evolve over time. For example, a picture of me drinking beer would’ve been horrifying to see on the web in the 90s when I attended high school. Nowadays social media is littered with frequent posts from teenagers drinking, carousing, and carrying on, because they think nothing of privacy. Teenagers of today record EVERYTHING and post it. We trust our privacy settings to actually handle our privacy for us. That notion was completely non-existent 20 years ago.
2. Predictive Analysis will continue to evolve the more we stream via services like Hulu and Netflix
Whether we realize it or not, every preference we express is carefully recorded and tabulated. If you watch 1/3 of a show before you abandon it, that statistic is collected. After much frustration when seeing the same car ad for two weeks, I discovered that it’s up to me what ads I see. You can actually specify your advertising preferences in Hulu and tell them what products you’re making purchasing decisions on in the next 6 to 12 months. Indicating preferences does not an algorithm make, but all of this data gets sliced and diced for the analysis to come.
3. The Outliers don’t matter
There is a statistical concept that’s unknown to many, which is 1s and 5s syndrome. Most people do not take the time to review something unless they’re really happy or mad about it. The trick is to remove the outliers so that patterns can be found. My data sample was skewed with all of this junk, so I’m not currently able to rely on the suggestions provided.
Although I kind of like to skew data for fun. When my agency first started managing advertising budgets for clients on Facebook a few years ago, it was the wild, wild west. The CPMs were amazing and beat the pants off of Google PPC at the time. However, the search algorithms were fairly crude. I often skewed my personal data just to see what ads I would see. When I changed my age to 85 I was horrified! Advertising to the elderly was very crude, almost always about embarrassing ailments with very large text.
When it comes down to it, perhaps I was angry at my Roku for running on and on, skewing the Hulu algorithm data, because as a serial tinkerer I didn’t want that taken away from me. Or perhaps it was a stark warning to me that data is not a play thing any longer, it is directly responsible for our entertainment.
What have I learned from all of this?
I’ve learned to not be offended by algorithms as it has nothing to do with me, personally. We all want to feel special and important. As dynamic as I am, I fit the profile of a young(ish) woman who likes to watch the occasional trashy show. I do not have children, and therefore I do not purchase diapers. The predictive algorithm is often the slave of its master and sometimes your data just gets skewed. This doesn’t mean you’re not special and unique! All preferences we express online are marketing fodder.
Next time you feel offended by a suggestion just relax and take a deep breath, then take a second to high-five a thousand angels in your mind. Being mad at an algorithm is like being mad at the wind.
Featured image via Shutterstock