author of mirdata.report 
On Quality - the Zen of Internet Maintenance
When I talk to friends of my father who have read Robert Pirsig’s Zen and the Art of Motorcycle Maintenance, they all say it had a profound effect on them back in the 1970s. Looking around, sometimes it seems the early tech of the 20th century was built primarily by people influenced by Pirsig’s writing. But it has been more than 40 years since the book was published—where are we now?
Today
Our culture is so obsessed with broadening access and driving engagement that we can overlook the fact that much of the ”quality” in our lives arises from a tiny minority that decides for the rest of us.
When these elites and technocrats focus on quality, as a society we seem to do well. The United States Constitution comes to mind here. It was the amazing result of a few individuals getting together to decide with little or no feedback what the rest of us need.
However, with technology, time and again we disregard quality in favor of engagement. The latter seems to be a more intuitive concept for the average Excel analyst. I mean, how do you report on quality?
And so the philosophy of today’s technologists is such that if the Constitution were to undergo redrafting today, there would probably be some companies wanting to solicit everyone’s feedback (the movie The Circle comes to mind here). I personally would hate to see the outcome of such a process.
credits: https://techcrunch.com/2017/04/05/netflixs-new-thumbs-rating-system-goes-live/
Netflix Rating System
Here in Seattle, Netflix is a big deal for much of the winter season. With winter approaching, I find myself more and more concerned about Netflix’s decision to replace its star ratings with a thumbs-up (or down) system.
Engagement, they say, “will increase the amount of data” and lead to better predictions. The message is appealing and probably strikes a chord with many Americans. After all, we are quite used to the idea in many shapes and forms. Look at these common assumptions: * Increased voting participation leads to better outcomes. * More bloggers leads to better news coverage. * More friend updates on social media leads to better relationships.
But does anyone really know these things for a fact? I mean, look at the US today—where is the evidence that those assumptions are true?
Like many of you, I have been a subscriber to Netflix and plan to be for now despite the news. But I am worried. You see, the 5-star system to me signaled a desire to understand my preferences. When I felt emotionally divided on certain films, I would avoid ranking altogether. Intelligent yet disturbing films strike that edge for me. My typical thought there is: “I don’t appreciate this artist’s choice of graphics in this particular film, but I still want to be exposed to the richness of ideas from this and other similar artists.”
That nuanced system is now no more. I have no doubt Netflix data scientists did their job well and verified that thumbs up/down is going to increase engagement. I believe they might have even confirmed that more of such simplified engagement leads to better predictions for films that I will ultimately rank thumbs up. They might even have gone above and beyond to introduce some mechanism of random insertions, by which they will introduce me to a random selection of films. And yet, somehow I am not warming up to the idea.
If the 5-star system lacked engagement, then it likely kept a lot of viewers who are not focused on quality out of the system. Ratings came from true quality viewership—people who could compare and even withhold their voice. Steve Jobs once said that consumers often don’t know what they want themselves. Tech companies are quick to recognize this fact when building their products, but slow when it concerns the facets of content and data that consumers create. Hypocritical, no?
Never-ending cycle
Unfortunately, too often I have seen technologies that I trusted deteriorate. They deteriorate only to be quickly replaced by another brand. The story goes something like this:
- The latest email client is introduced and people love it (let’s call it “hot mail”).
- Hot mail is scaled.
- People abandon the hot mail service in droves for new alternatives.
And so we jump from one brand to another—from Altavista to Yahoo, from Yahoo to Google. All because time and time again tech companies compromise quality for scale. For that reason, it is my hope that important questions are actually asked before companies such as Netflix decide to yet again overhaul something that has become part of our culture.
What I personally like to see is complexity of choices and their richness. For me the “complexity” of Netflix’s 6 choices (a no-star rating is a choice too) provided a plurality of responses.
Perhaps we need more people like David Hansson from Basecamp or Craig Newmark from Craig’s List—people who basically know when to stop. To appreciate the moment when it is OK to say no to more funding or subscribers. Perhaps we just need more people to recognize quality when they have found it. And more importantly to recognize when to STOP.
The Internet is one “thing” we as a society won’t be able to buy back.