The discussion clarifies misconceptions about YouTube's algorithm, emphasizing its evolution from simple metrics like views to a sophisticated system focused on viewer satisfaction and retention. It uses billions of signals to personalize recommendations, considering factors such as watch time, engagement, and contextual viewing habits. The algorithm is indifferent to creators, aiming solely to keep viewers engaged by serving content they enjoy. A bad video alone won't harm a channel, but consistent underperformance may reduce visibility as the algorithm prioritizes satisfying content. Key recommendation factors include personalization, video performance (e.g., click-through rates, retention), and external elements like topic trends. The conversation also highlights resources like Creator Insider for deeper insights and debunks myths about algorithmic "hate," stressing that success hinges on understanding and serving audience preferences.
Hey, I'm at and I'm at and this is creative generation All right, welcome back to creative generation this week it is me and with Fred Rico. Hey, go on Fred Yeah, good. Hey, buddy. I'm good. We're gonna bring the energy We're gonna not do a Q&A Roles because Long time listeners know that I break all the rules that I make Mm-hmm because I make them I get to break them. I've thrown the wheel out I Have taken some questions that we get a lot of and it's about the Big bad scary thing called the YouTube algorithm There's multiple of them. I've got a series of questions. I've kind of Sort of them out based on what people ask a lot of but also what I think people should know and Fred I'm gonna grill you on them and by the end of this Everyone's gonna feel super comfortable understanding what the YouTube algorithms are doing and why they don't hate you and We can all go and make awesome content and be super happy. How's that sound? That sounds good. Am I over promising you are over promising Let's see if we can over deliver as well. All right cool. All right Why did I say algorithms? What are they? What's going on like can give me the overview like I think that's the thing We built this thing up. Let's break it down. Well, I mean I guess the best place to start Well, we usually start a lot of our vent we always talk about the foundational aspects of YouTube and understanding the platform Once you understand the platform I think you can better understand what the algorithms are trying to do so a lot of people still think of of YouTube like the you know old school YouTube the YouTube from the olden days right look at all the views and likes and and then it decided hey You know based on this video that had so many views and this video that had so many likes that had more than this video that was prioritized and Pushed forward. It's really moved on. Hang on. Let's give an example of that because some people might not know it that is yeah Let's just like look up Charlie bit my finger What's that one where the kids coming back from the dentist those have things one off? There's one off in a short viral videos when there were neither platforms that were on like this where all the viral videos went So that's what a lot of people still think it is that that's the direction It's going but it's not it's a different type of platform. That was a very old YouTube. It's grown up a lot It's a teenager now. Well, how old is it now if dang if teen yeah the teenager now so it's it's advanced So what basically happened over time was it went from just looking at things like views and likes and this idea of our Alities that are looking at more important facts that realize that people were looking for more they wanted to be entertained You know, they're not just there to see one off videos. They were there to watch Multiple videos so started to progress and then started looking at things like watch time and watch time became a really important aspect Of the system and people can really obsess with that like trying to get as much watch time as possible and then you to realize you know what It's not just watch time like you know not all watch time is the same. You know someone might watch and Half an hour of news and not you know because they they have to Rather than because that's something they enjoy and want to so it's not gonna keep defeating them You know things like music and again So I started looking more metrics like what really satisfies but what what really satisfies a person what are the really after what will keep them watching again and again Keep coming back to the platform and really enjoying all the content kept capture their attention So as these things got more sophisticated the algorithms like you mentioned before got more sophisticated understanding what people liked You know understanding where they were watching it how they were watching it and what kind of kind of they wanted to watch So they started to develop this very Smart response to everyone who logs in looking at something like 8 billion different signals to understand how to customize the experience for that You use our viewer to keep them watching as long as possible so over time this what's happened these algorithms have developed Then grown as YouTube is grown to give the viewer the best experience possible and keep them watching as long as possible so Short answer of what does it want it wants? People just it wants to keep people's attention in this day and age Every creator the currency or operating as attention you have to keep attention not just from other platforms from everything else everyone else Everyone is doing right so that's what it's trying to do is trying to help you know get someone's attention and keep them watching Quick tangent. Maybe it's not even a tangent. It's a related question. Does the this the algorithm like care about individual Channels like and people does it hate hate certain people? People really get in touch with the fact that they think the algorithm hates them I have some incredible I don't know what's going on that hates me something's going on here But no, I don't think so. I think the the algorithm is really in different the algorithm You know looks at the viewer it chases the viewer follows the viewer. It's designed to see what does the viewer want and how can I give them more of that? How can I keep their attention? So you know, I think trying to answer this in touch with YouTube said this as well as like You know, when you think the would algorithm replace that with viewer and you're going to pretty good approximation of what it's trying to do It's trying to replicate those human aspects they keep people watching so End of the day if you think the algorithm hate is probably the viewers Like what you're after so That's what I do like the algorithm doesn't hate you. It's indifferent to basically cut across all the trying to do is try to find the best Content possible put it in front of someone and keep the watching as long as possible. Do you create a the algorithm doesn't care about you? Care is about the viewer. It will say you should care about care. You should care about if you care about your audience Anyway, I think It's it's that anything doesn't care what I think the way it you should look at it is it doesn't particularly You know put any emphasis on on any particular person and any emotional kind of way right? It's it's emotional This is basically how I describe it. Yeah. No, I'm I'm I'm taking to an extreme having fun with it. Okay, so um and I think I think a really interesting point there to point out I think this stats still correct correct me if I'm wrong but like 70% or more videos are found through home and suggested so that's because Said algorithms puts it in front of the viewer. Yeah, so yeah, that's how powerful this thing is Yeah, it's how important it's pretty high So basically that that feeds are when you first log onto YouTube with our new desktop or your mobile device and then once you watch a video and You know gives you suggestions. That's the yeah algorithm in sort of full tilt It's basically trying to figure out you know when you first get there. Wow. What could you possibly want to watch and then once you're watching something How can I keep you watching so that's what it's basically I putting Oh, that just reminds me of a creator we had in a um an accelerator once who got really annoyed that's their roommate kept Boring their laptop and was watching all different stuff and they're like man you stuff and with my algorithm So their home feed was totally yeah skew with stuff. They didn't really watch because their roommate was boring their laptop I love it. All right, so this all seeing all knowing cold In different algorithm that just wants to give the viewer satisfaction. Yeah, it's making a whole bunch of recommendations like it makes Recommendations for the viewer that a personalized per viewer right yeah, that's kind of where we're at Each viewer gets their own recommendation. Yeah, what's it making those recommendations based on So basically the system recommends content based on on three key aspects and look if you want to a full summary Actually, YouTube did a really good version of this on creator insiders. You can go watch those videos there I think actually made it a two-part it so they really got into it But in a nutshell it looks at three key areas first one being personalization That's based on the viewer themselves like what gives them enjoyment. What are they enjoy watching right? That's key and foremost That then looks things like you know based on all these factors like what's new? What's interesting what's coming up? What's different? You know what haven't they seen before and it looks at aspects around that And then it looks at their overall interest in a topic like how interested is that person in particular's topic or series of topics and how are they all connected So it's looking at those personalization aspects The next area it looks at is you know, we call performance and that's based on the actual videos themselves that going up It looks at the videos and said okay, what is the click-through rate like on this video for example? Right? What is the overall engagement like how is a retention being on this video? Is it got those satisfaction markers we talked about before right and the last one we call peripherals But it's really the external factors and that's things like the topic generally and um, you know things with like the competition The topics and not all topics are the same. I think the example they give is like Different sports like soccer might be a lot larger than you know horse riding for example um, so not all topics are equal and It sort of balances out the serving based on that but also things like seasonality like you know football for example might be More popular during football season, but not as popular in the office season So it's looking at all those factors. So basically that personization aspect around the viewer themselves performance aspect based around the videos And then that peripheral or external aspects like the external factors How do they affect the overall distribution of a video and in a nutshell? That's the three areas we can probably do like a full hour in each of them And like I said if you want to check it out a bit more detail check out the videos on crater insider the two partners They did talking about these three aspects Boom and a good resource crater insider if you're not watching that already You want to know it out on what's going on on youtube and how it all works. You should definitely check it out Okay, cool. So it's done that It's put the videos in front of me like how does it know that people like Liking a video like Obviously, you know, we talked about earth back in the day. It was like views and likes and then now like there's the all powerful watch time Like people like watch time retention So like is that it like how does it know? Yeah, what is it looking for? What's Satisfaction it's a it's a combination of of
a lot of things. Yes, one is looking at the overall retention on a video, but it's also looking at like this side of satisfaction, like not all watch time is equal. So what is actually satisfying the viewer? An example that I give is like I think I mentioned before is like you might watch a tutorial on like on something or what some news or you know I might need to know how to fix a door. So I watch like a half hour or an hour video on how to fix a door, but you know if it looks at it on the face level say hey why this person watched an hour long video on how to fix a door. I'm going to keep giving him content on that. That would be really boring and kind of intuitive. So it's not trying to do that. It's basically trying to look at it and say okay, why is this person watching the video? Are they watching you know other markers that help me understand that the person is actually enjoying it? And that's why it looks like you know like so over a watch time comments the surveys that actually asks you about after you watch the video. For example, you know what do you think of this video? You know how to make people all those things going to understanding what satisfies you, but you know watching the video and enjoying the video and coming back is obviously a key part of that, but the idea of satisfaction is becoming more and more important and as the system has become more intelligent there is they're more readily able to determine how satisfying a video is and what will keep people watching. And the fact that you know you're on these platforms for a really long time you know constantly consuming content does show that they are pretty good at finding satisfying content. Have you ever answered one of those surveys? I have. I do answer them because of the because I know that they help the system understand what I like, but I do under I do feel a little service. Do you think you're the only one? I'd love to know if anyone out there has filled out the survey please hit us up. Let us know on our socials or something. DM me. I want to know who are you? What's wrong with you? You need to say that they get thousands of responses a day so they must they must have people feeling I don't know I mean in the fact that they still do it means it's probably valuable. So there you go. I guess there's what like over a billion people or something logged into YouTube or whatever the study is you know a few there like of course there's going to be thousands of responses. I don't know sorry I'm going I'm going deep and dark on this one surveys they're the hardest thing to get people to fill out but they're very useful and I've never done it even though I know what it does. It's interesting what you you mentioned to you like about like how to fix a door like that context you're all understanding of what you want and when you want it like I think there was a time that if you did look at how to fix XYZ if they just got filled with more of how to fix XYZ but now it's it's clearly getting smarter what's going on there because yeah it seems to understand the context of that's right watching. Context is becoming more important too and you know there was a time where the system did not understand the context of what you're watching and sort of when you're watching it an example I used to have was when my dude was really young she used to watch videos of birds eating grain of seeds when she was eating dinner it's very calming so typically you put it on on the Google Home in the background and there'll be this like 10 hour video of birds eating seeds and they were designed for house cats all right like to keep them sane during the day you keep this video on in the background but my daughter really liked it when she was eating she used to watch the birds eating it was really great but what I'd found was like I was getting these videos fairly regularly all over the place but over time I found that it was only serving those videos or as a suggestion to that Google Home and not on my other devices so it's basically understanding the context and realize that hey you don't like these bird feeding videos all the time you might like them on your one device that understands that's the things like context do do matter and that becomes more apparent obviously the different ways you use your devices so it really understands like on different surfaces people are you know more interested in browsing versus on some surfaces where they want to be maybe a little bit more specific for example so yeah those kind of things do do make a difference now the birds the birds coming you can't reminds me of the example we use around people who love guinea pigs and there's that eight hour music video for guinea pigs natural stress and anxiety relief for guinea pigs that has had well over two and a half million views eight hours stress relief music for you guinea pigs that's where I just think YouTube's a weird and wonderful world my stress out guinea pig is so happy now all right now that makes like that's and that's good like you know that's that's why we don't end up with door fixing videos slammed in your face all the time when you only needed it to fix the door and you're done or the bird videos you know playing in the background right now for you on your google home on your TV rather than on your mobile okay changing gears just a little bit like we've got this all-seeing all-knowing algorithms that are then trying to you know find the right video for the viewer based on all these things ever and you know going back to the algorithm hates me or all like the impact of the algorithm on like a bad video like people like oh this video killed my like killed me in the algorithm and we hear that all the time still and it from creators they're like this is took one to many bad videos and like your content will stop being served forever like good old content new content so like is that true like is it like can you bad a bad video just tank you like just one bad video here and there that I don't know let's not get into like what a bad video we see like that's just a whole can of worms I mean yeah people get very focused on underperforming videos and then they worry that it will affect their entire channel at the end of the day I don't think YouTube gives a channel like an overall like a reputation score and say hey this is a bad channel don't get near it and list the channel like violates the the community policies I don't think that's an issue or yeah they hate that yeah so I mean what it looks at is the individual videos you put out and that relationship they have with the viewer what can sometimes happens that creators create a lot of videos that underperform and they're you know there's a lot of variety in the videos and they're never really and you know have a specific direction and YouTube is just not how to survive the videos so you know it's not gonna take every video you do and just put it out the entire planet that would be insane because there's so many videos being posted on the platform all the time so it's really trying to understand who your audience is who enjoys your videos who's satisfied with videos and try to give them more of them so it's basically creating this relationship with you know the individual viewers and the individual videos and your videos and the collection of your videos as a whole really and trying to put it out so no one bad video won't affect it but if you consistently make bad videos what's gonna happen is that the audience will lose interest in the channel or the videos and when you know it looks at the data says hey these videos aren't getting any watch time they're getting any any good retention like no one particularly seems interested it's just not gonna keep serving the videos again and again and again so keep that in mind but no one bad video won't affect the channel like it won't cause a massive downward trend basically on the flip side you know Julian O'Shea so Julian O'Shea is a creator and we've had him on the podcast check out his episodes it tells interesting stories about seemingly mundane things like sidewalk cross buttons and stuff like that but Julian had this idea that YouTube should have a button that each creator once a year can be like this is my video this is the video for the year this is the good one this is the best one they get to hit that button and bushgar it gets put in front of like as many people as possible that that YouTube like still with relevancy but just bushgar out out it goes it'd be kind of fun but when you think of how many people of creating videos on YouTube it would be a complete monster but what do you think about that one well the question without notice on that one I think that what that's an interesting I think it'll be misused greatly all right um something people like you know talk about a lot still and they're really like it's a lovely dare I say vanity metric um subscribers they like I've got a lot like people go I've got lots of subscribe subscribers why isn't YouTube putting that video in front of those subscribers they get really narkey that I've got all these subscribers and they're like they don't get the video served to them like YouTube is broken the algorithm's stuffed because it doesn't show my videos to my subscribers what's up but yeah that's tracks that should it shouldn't it well I guess the subscriber problem is based on the fact that um you know well it's problematic because it's such an old it's not an old fact it's been around forever um and so many people put their worth as a youtuber based on that number right so and I can even YouTube like they'll give you the play buttons based in certain milestones there's still a way to sort of roughly tell how bigger creator is but at the end of the day these algorithms are so sophisticated that you don't need to subscribe in order for it to give you the content you want in fact if you think about the content you watch a lot of on YouTube the stuff it serves you most of the time you're not subscribed to those channels right it just knows and understands that's what you like so it'll put it in front of you so you watch it um in fact I watch a channel the other day I was watching it I'm like hey I've been watching this for a year and I haven't subscribe but I always get it my feed.
because it knows I like and I enjoy that channel. So it's subscribe is the idea of having a being a subscriber isn't as important as it used to be. And if you think about it too, like how many channels have you subscribed over the years and how many of those channels have you subscribed to do you still watch, right? I may be watch five of the channel that subscribed out of the hundreds I've done over the years. So if it basically served you only a subscribe content, you'd be getting hundreds of videos from channels you know no care about. It's not like you unsubscribe when you stop liking a channel. You probably just leave it and let it go in the background. So that's why it's not as as as important a factor anymore. So I wouldn't focus on subs, especially as YouTube gets older, even in the last week or two as you'd see, like YouTube has changed the way the sub button works and the color of it and stuff like that is it sort of maybe moves away from the importance of it. Really the most important thing is you want to get people to watch your content, watch as much of it as possible, really enjoy it, really be satisfied by it. Like that's the thing that's important. That's why things like retention is important. Watching your audience are they a returning audience? Are they coming back to your channel? Like those are the things I'd be focusing on. Definitely. I think people only unsubscribe from a channel if that creator is cancelled. Yeah. Or the half of me happens and they really dislike it. Yeah, like really. You just, he's got time for that. I don't even know how you do it. Okay, we've thrown a whole bunch of stuff. So I think you alluded to you kind of went there just then but like it's all it's all well and good to have all this information. What's the practical of it? Like what should a creator do about this? Like what's the takeaway on that? Like, it's nice to know. I mean, what do I do? This actually can't be down to something like I said, Todd, who's the head of the algorithm team when we were chatting at one of the Vidcon's a few years back said this and I thought it was actually, you know, it made a lot of sense and he actually said this from a point of view, a completely logical point of view. He goes, if you think about it, if you get somebody just to watch a bit of content on your channel and they stay and watch the whole bit of content and then they come back and watch another bit of content, you know, that's a really good signal to say, hey, there's something here that potentially is interesting to this person. Like, if you can basically stoke that behavior, getting people to watch one bit of content, they can be back and watching another bit of content at a base level. That's a pretty valuable thing. So if you've wanted to concentrate on anything, try and focus on that. And I think it's, I think it's really good advice because one, it's from the guys who run the algorithm team, two, it's something that I think over the years has stood up very well. And three at the end of the day, it's sort of the goal of YouTube, right? Like I said before, right at the start, they want to keep your attention and if you're watching all these videos and coming back and watching all the other videos, you've got their attention and that's what the system is trying to do. That's what the algorithms are trying to do. So if you want to focus anything, try to get somebody to watch a video on your channel, get them to watch as much as possible, then stay and watch another one. Come back and watch more. And remember, it's not about you, the creator. It's about your audience. So look Fred, we have and dear listeners, dear creators, we've you know, scraped the surface of the almighty YouTube algorithm. But you know, personally, it's about as deep as you probably really need to get to understanding it. I think, I really, like, we spent too much time in this world of creators and creator education, diving too deep into the understanding that mightnute machinations of this beast. But anyway, if you do have more questions, you want to go deeper, you want to get really nerdy with it, we can do that. We can do future episodes on that. Or you can come on over to the creator generation discord and we can get real deep on that too with us and the other community. So that's in the show notes below. So hit that up, you can get in touch with us and we can rock and roll and go nerdy. Sorry for throwing out the creator generation Q&A roulette wheel. I'll have to maybe resurrect it back from the dumpster. But thanks for hanging out in Fred Rieger. Thank you for your wisdom. No worries. See you guys soon. Cheers, bye.
Podcast Summary
Key Points:
YouTube's algorithm has evolved from prioritizing views and likes to focusing on viewer satisfaction and retention, using personalized signals to keep users engaged.
The algorithm is indifferent to individual creators; it serves viewers by recommending content based on personalization, video performance, and external factors like topic popularity.
A single underperforming video does not "tank" a channel, but consistently low-quality content can reduce recommendations as the algorithm prioritizes viewer satisfaction.
Context matters—the algorithm considers device usage and viewing habits to tailor recommendations, avoiding irrelevant content like repetitive tutorials or niche videos.
Summary:
The discussion clarifies misconceptions about YouTube's algorithm, emphasizing its evolution from simple metrics like views to a sophisticated system focused on viewer satisfaction and retention. It uses billions of signals to personalize recommendations, considering factors such as watch time, engagement, and contextual viewing habits. The algorithm is indifferent to creators, aiming solely to keep viewers engaged by serving content they enjoy.
A bad video alone won't harm a channel, but consistent underperformance may reduce visibility as the algorithm prioritizes satisfying content. , click-through rates, retention), and external elements like topic trends. The conversation also highlights resources like Creator Insider for deeper insights and debunks myths about algorithmic "hate," stressing that success hinges on understanding and serving audience preferences.
FAQs
The YouTube algorithm is a system that personalizes content recommendations for each viewer based on billions of signals. Its primary goal is to keep viewers engaged by suggesting videos that satisfy their interests and keep them watching as long as possible.
No, the algorithm is indifferent to individual channels and does not have personal preferences. It focuses on viewer satisfaction, recommending content based on what keeps each viewer engaged, rather than favoring or penalizing creators.
YouTube recommends videos based on three key areas: personalization (viewer preferences and enjoyment), performance (video metrics like click-through rate and retention), and external factors (topic popularity, seasonality, and competition).
YouTube looks beyond just watch time to gauge satisfaction, considering factors like retention, engagement (likes, comments), and viewer surveys. It aims to understand what content genuinely satisfies viewers, not just how long they watch.
No, a single bad video won't tank a channel. The algorithm evaluates each video individually based on its performance with viewers. However, consistently underperforming videos may lead to reduced recommendations as viewer interest declines.
YouTube considers context, such as the device used and viewing habits, to tailor recommendations. For example, it may suggest different content on a TV versus a mobile device, understanding that browsing preferences can vary by context.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.