Music Interview: YouTube Algorithm Causes Major Disruption to Jazz Video Guy, Bret Primack
By Steve Provizer
Bret Primack explains how YouTube has basically nuked the Jazz Video Guy channel. And the same thing is happening to other content creators.
This follows up an interview I did with Jazz Video Guy Bret Primack on April 24, 2019, in which he said: “Thankfully, YouTube saved jazz.” Things have changed. Check that conversation out to learn about Primack’s bio and his thoughts on the jazz video landscape.
To begin with, we need to look at YouTube algorithms — they’re a big part of the story.
The Oxford Dictionary defines an algorithm as “a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer.” I think of them as a way to implement market research at a microscopic level. Using algorithms is the ultimate form of number-crunching, and the internet is a number-crunchers dream: YouTube generates about 80 billion signals every day, according to this site.
Algorithms drive 70 percent of all views on YouTube. They apparently change every day, with the goal of keeping people on the site and exposing them to as many ads as possible. YouTube executives and engineers are elusive when it comes to the mysterious (proprietary) algorithms, which use data from viewers and channels to learn (and determine) who will watch, like, and share which kinds of videos.
There are three important places where the power of the algorithm shows up: 1) YouTube’s homepage, showing you rows and rows of video suggestions; 2) powering YouTube’s search function; 3) selecting the “Suggested Videos” that show up in a column on the right edge of the page. In “auto-play,” the first of these suggested videos is also “up next.” The videos’ end screens show even more suggestions. These are the ways YouTube says it is “surfacing” a video.
Using the data profile of everyone who uses the platform (this kind of data is what you trade on the Internet to receive free content), the algorithm generates video candidates, then ranks them by how well they do in general. It groups similar users together. A video watched to the end carries much more statistical weight than one abandoned after seconds. That helps to cut down on the profusion of click bait. YouTube honchos say some randomness is introduced that encourages watching a variety of subjects and introduces videos from less popular channels. I find it hard to figure out the degree to which that fits into the overall picture. Thanks to this site for a lot of this info.
AF: Bret, what have been the average numbers of viewers for your videos? When did you first notice the decline in viewership and what were your first thoughts?
Bret Primack: Up until three months ago, nearly all my new videos averaged 1500 views after 10 days. Then I began to notice that was no longer happening. Most of my new videos didn’t get more than 750 views, even those from more popular artists including Sonny Rollins and Freddie Hubbard. That soon translated into a big drop in overall viewership. From a daily average of 8,000 views to 2,500 or 3,000 per day.
AF: What’s the average length of your videos, and how has that affected viewership.
Primack: My videos average about seven minutes. Since day one, 17 years ago, only about 20 percent of the audience watches an entire video. I lose around 50 percent of viewers after the first minute. My average view duration is about 3:30, or about one third of the video. Why don’t people watch an entire video? Declining attention spans, multitasking, and a need to keep finding something that might be more interesting. The suggested videos that appear on the right side of the screen don’t help. When people are watching a video — if their attention lags for one second — they are likely to just click on one of the suggested videos.
In the past year, with the growing popularity of TikTok, many people no longer watch longer videos. They’re looking for immediate gratification, something that comes and goes very quickly. Now, to compete with Facebook, TikTok, and Instagram, YouTube is now totally focused on Shorts, videos that are one minute in length. And, because YouTube Shorts have become so popular so quickly, the YouTube algorithm is reflecting the change: at this point, creators of shorter videos are being more prominently featured in Suggested Videos and search engine results.
Up until three months ago, the YouTube algorithm favored views per video. Today it leans more towards the length of video views — and whether or not a user stays on YouTube after leaving a video. It’s all about keeping viewers on YouTube for as long as possible, so that they will eventually click on the ads. Remember, YouTube is an ad-driven site, like broadcast television.
With shorter videos, once a viewer watches one short, it’s very easy to keep scrolling for others. In seven minutes someone may have watched 14 different videos, but the viewer is still there. When you wake up from this hypnotic experience, you think, where have I been the past 45 minutes?
AF: You mentioned that you heard from other YouTube entrepreneurs whose numbers have fallen off. What can you say about the content areas of their videos?
Primack: Many creators have experienced this sort of drop-off in viewership, no matter what their content. David Pakman has a popular political channel whose video “YouTube Nuked My Channel” details his story. Rick Beato, a popular creator in the music space, has also expressed his frustration with the direction YouTube has taken. I’m not in touch with any other jazz content creators so I can’t speak about their stats.
YouTube emphasizes the importance of titles and thumbnails as a way to attract viewer attention/interest. Previously, I could just put a photo of the artist in a thumbnail and a title like “Rollins Plays Misty.” That just doesn’t work anymore. The video I’m posting today is titled “Is Sonny Rollins the Greatest Saxophonist of All Time.” I’m sure that will garner plenty of views. It’s almost like YouTube has gone tabloid. Sensationalism sells. I don’t like it, but if I want viewers for my videos I have to play the game according to YouTube’s rules. Several people have accused me of focusing on clickbait. I tell them that now I have to.
AF: When you went to YouTube looking for some guidance, what happened?
Primack: It’s hard to reach and communicate with an actual human being at YouTube. There’s a support chat feature every creator has access to but the people who respond in that area seem to know nothing. Trying to get their help is a total waste of time. But, after weeks after writing emails, etc., I finally got hooked up with someone who took the time to do an analysis of my channel. It turns out that one stat stood out: my click-through rate. That’s the percentage of people who see the thumbnail and then click on it. I had been averaging 5 percent but, for some reason in the past few months, that decreased to 2.5 percent. And that has a ripple effect, so over time, my overall views are down 66 percent.
I have nearly 115,000 subscribers. Why is it that I can post a video and garner only 250 viewers? That’s because YouTube isn’t recommending my videos anymore. I posted a Burt Bacharach video last week, after his passing. An hour after I posted, I went to the Search Engine to see if my posting was included. I searched for Burt Bacharach videos posted in the last two hours and my video didn’t show up on the result. When I asked YouTube what the problem was, they had no explanation. “We’ll look into it” was their answer. And I never heard back from them.
What I found interesting is that I’ve been posting the same content for 17 years. Why, now, do fewer people click on the videos? YouTube has suggested better titles and thumbnails and posting Shorts. I’ve started doing that. After a month, I will revisit the stats and see if there’s any change.
AF: You’ve said you’ll continue as a filmmaker and the 2,500 videos you’ve posted will remain on YouTube. Will you continue to post on YouTube or look for an alternative?
Primack: I’m not quitting YouTube. I’ve developed an action plan and I’m busy implementing that. But after a month, if I don’t see any increase in my views, I will most likely post only one video a week, instead of three or four. Sadly, YouTube is really the only video platform where I can reach large numbers of people.
A lot of writers are using Substack to develop a paid audience (see my conversation with author Ted Gioia about Substack here). I don’t think that’s possible with jazz videos. Like so much of the content on the web, people get it for free. Once they get it for free, it’s nearly impossible to get them to pay for it. I have Patreon attached to my channel, which many people use to generate money. I average about $12 a month from Patreon, even though I suggest to my viewers it’s a way of supporting what I’m doing.
I’ve also asked my viewers to like, comment, and subscribe to my videos. 90 percent of my viewers aren’t subscribers. If I had a greater percentage of subscribers, that would help me in the YouTube algorithm. But, sadly, very few people will actually do that. It’s much easier to sit and watch and do nothing.
As long as social media outlets are controlled by major corporations, quality won’t matter. It’s about numbers, quantity. If it were possible, YouTube would post executions, because they would draw the largest audiences. That’s why Trump will soon be back on Facebook. He draws a large audience. Why should it matter to Facebook if they’re destroying democracy? YouTube is like a tabloid newspaper, the more outrageous the headline, the title, the more viewers. That is all that matters.
AF: Any advice for content creators?
Primack: At this point, I have no advice to creators of any sort of content: except to say that if you create work with the lowest common denominator, if you successfully prey on people’s fears, you’ll probably get a lot more viewers than if you express any kind of truth.
Note: Primack added this in follow-up emails:
I’m implementing part of my Save Jazz Video Guy Action plan. Shorts. Posting twice a day now to YouTube shorts, Facebook reels, and Tiktok. But not Twitter. Not a Musk fan.
One can go deeply into the video weeds to try and make sense of algorithms and all the rest of it, but it boils down to this: YouTube is a business — and a lucrative one. In 2006, Google bought YouTube for $1.6 billion. In 2022, its revenue was $29.24 billion.
In 2023, the jazz demographic is aging, and the attention of younger people is increasingly drawn to very short videos. There may be talk from YouTube of introducing randomness to encourage watching a variety of subjects and videos from less popular channels, but this will have minuscule impact on the fortunes of a shrinking musical genre that calls for comparatively long form videos. For 17 years, Primack was able to draw enough eyeballs to make some money on YouTube. Chicken feed, of course, given that the only sure way to get a good piece of that action is to know what the social media trends are and tailor postings accordingly.
This conversation has a familiar ring to it. Every time a feature film about jazz is released, no matter how bad the movie might be, there’s a debate about whether it’s still a good thing, because it might get people to listen to jazz (c.f. Whiplash, La La Land, Miles Ahead, etc.). A similar question can be asked about the YouTube realm: rather than present an entire concert or composition, can a jazz video be tailored to draw YouTube eyeballs and still deliver the jazz “message”? Will the short-form videos that the Jazz Video Guy is beginning to produce draw enough attention for YouTube to “surface” his videos, to the point that more people will sample the 2,500 other videos he’s produced? We’ll check back with Primack — and Google Analytics — in a couple of months and find out.
Steve Provizer writes on a range of subjects, most often the arts. He is a musician and blogs about jazz here.