YouTube Algorithm 2026 - What Creators Need to Know

0:00this worry that like oh well if my subscribers don't like it then YouTube's going to decide for everyone that it's not a good video. So first, it's not only your subscribers who will see it, and two, the algorithm really wants to um know what's best for each viewer, not just based on an overall aggregate opinion. The fact that someone else doesn't like something that I like

0:24shouldn't prevent me from seeing it. Yes. I'm Rene Ritchie, your creator liaison, and I am talking today to Todd Bouprey, who is senior director of growth and discovery at YouTube. Todd, what does your team work on? So my team works on a variety of things, but most of it is about recommending content to viewers, uh whether that be on the homepage, on the watch experience, through notifications, and we're really just trying to get the right content to the right viewer at the

0:51right time. So I went my scour the internet, uh every nook and cranny, and I was trying to see the most common questions that creators had about like {quote} the algorithm. And one that comes up a lot is sort of this idea that I as a creator make a video and then I upload it to YouTube and then the recommendation system, the algorithm, pushes that out

1:14to varying degrees of success. Um how does it actually work? Yeah, so I hear a lot of of the language that creators use does kind of involve this pushing, the algorithm pushing content to viewers. While we do push notifications to people's devices if they're opted in, most of the recommendations are more pulled for each viewer. And so the way to think about it is when a viewer opens their app or turns on their TV, um that's when the recommendation system is activated for that viewer. And the decisions that the recommendation system makes are generated at that moment, considering the what we understand about the viewer and what they've watched

1:57before, what the current context is. Like, what time of day is it? What Which device are they watching from? And so, what the system does at that moment is it takes that information and understands and then brings in candidate videos that might be a good match for that viewer and then ranks them for that viewer. And so, each viewer gets a

2:21different ranking um for for each video. Um you're going to get a different ranking that I'm going to get. And so, I think the more that creators can think about the recommendation system as optimizing for each viewer as opposed to very content-centric Yeah. you know, we're not really like looking at a video and saying, "Let's go figure out who the audience is going to be." It's looking at each viewer and saying, "What are the best videos and posts and other formats um for this viewer at this moment um giving

2:55everything we understand." So, to your point, there might be a video that has an absolute metrics that are very high, but you might love it and I might not, and you have to take that into account. That's right. Yeah. So, another widely held belief is that as soon as a video is uploaded, you test it with our subscribers. And if our subscribers aren't as engaged, like the

3:16video is just cooked. Yeah, so when you think about uh ranking a video um and you know at any given time, there's, you know thousands, millions of people loading their home pages, loading loading watch pages that have recommendations on them. And if you think right after you upload a video who are the viewers that are probably going to see it at the top and most

3:42prominently? Yeah, the more history that a viewer has with a channel and the topic probably it's going going rank higher for those viewers than the viewers who've never watched the channel before. But, when we look at the data, it's not, you know, it's not like, "Oh, you're only getting your very, you know, core users who've watched your channel, you know, every upload right at the beginning." We see diverse audiences that actually end up getting your video uh visible, maybe because they're scrolling farther. And so, it's not just like it's your core audience in the first hour or even

4:21minutes. Um I've looked at channels and I see like that the majority of their audience can be non-subscribers. And even in the first hour um videos are reaching people who've never watched the channel before. So, it is a mix. And the other thing to remember, though, is like it doesn't only matter like uh who saw this video in this hour in terms of like how the algorithm understands it. It's not just looking at the overall opinion of everybody in that hour as one, you know, performance metric. It's learning for different audiences. So, it can learn that for subscribers, this video is not, you

5:07know, [clears throat] performing well. And for people who haven't watched the channel, it is. And so, you know, we're going to use the relevant data for each viewer. So, this worry that like, "Oh, well, if my subscribers don't like it, then YouTube's going to decide for everyone that it's not a good video." So, first, it's not only your subscribers who will see it. And two, the algorithm really wants to um know what's best for each viewer, not just based on an overall aggregate opinion. Like, um I always like to think about how the fact that someone else doesn't like something that I like shouldn't prevent

5:47me from seeing it. Yes. Yeah. I've also stopped looking at my metrics for the first couple hours because I realized if someone is on their phone waiting for coffee and a notif- notification comes in, they might click it immediately especially if it's a shorter video. But if it's a long video, like an hour-long documentary or like very like challenge video, they might wait until they're at their TV at night or on the weekend to watch with their family. And so, I think that whole idea of like instant results is is

6:14something of the past. Yeah, I think creators do tend to overfocus on, you know, the immediate data and know that from the recommendation system's perspective, we're not basing decisions just on those initial moments. Every day there's billions of new uh impressions that flow through the system that we're constantly learning from. And that's why you see that, you know, yes, a lot of videos have their peak moments in the first day, but a lot of videos also get the majority of their views, you know, a month or or later if it's evergreen content. We're not going to stop you know, trying to see if there's viewers that um

6:57would be interested in it. All of my evergreen videos are my biggest videos and they took several months to become my biggest videos. But the other thing is I will see a thumbnail and I'll be like, "No, Todd." And I'll see it again and I'll be like, "No, Todd." But maybe like the sixth or seventh or 10th time I click on it and then I really enjoy it. And I can't be

7:14alone in that. Yeah, we This is one of the things that we struggle with sometimes because we when we survey users about their experiences with recommendations, one of the things they tell us is that they often see the same video recommended too many times. And they're like, "Why do you keep showing this? I want something new." And so, we are sensitive to that and we've run all sorts of experiments where we say, "Well, what if we say uh you you wouldn't be recommended a video more than eight times. We do see that people do engage sometimes, uh, you know, it takes a few tries, uh, but

7:49we're we're trying to get smarter about it so that, you know, if if it's truly likely that you tend to save your hour-long videos for watching on the TV in the evenings, then, you know, we will probably try to do a better job of making it front and center when you you have the right context, and less so, um, when you're in a different context and it's just not really relevant to you, cuz we don't want to waste your time, uh, recommending things at the wrong

8:22time. YouTube nor creators want the wrong video recommended to the wrong person. That's right. I worked with one creator one time who's, uh, had had a viewer who was missing out on on recent uploads from that, and, uh, we worked with a viewer to get permission to go into the system and and see if we could find a bug in in what the system was doing, and what we actually found is that that viewer had been recommended, uh, the recent uploads 19 times, and they didn't notice. If if you have a viewer that typically watches from 50 different channels regularly, like you were saying earlier, some days you just don't have

9:03enough time to consume everything you're interested in. Some days you may not scroll far enough. Maybe you only scrolled through your top 30. What happens to me is I have an idealized Renee who wants to watch every video that I that has come out from every creator, every topic that I care about, but I have 10 minutes to watch. All other things being equal, I would love to have watched that video, but in

9:24the moment I chose something else. Yeah, we had these same challenges even when it was just the subscriptions feed, because, as you mentioned earlier, there's often just more content that people want going than they're going to have time to browse through and and watch. So the best thing you can do is just tell that viewer to go to your channel

9:44and watch a bunch of videos. Yeah, you could always go to the channel, you could always go to the subscriptions feed and catch up on the things you've asked for. So what then does really matter for the algorithm for the recommendation system? I know creators would love you to say like one magic number, um but is there something that you look for and does it depend on like phone versus TV and morning versus night and all those

10:07factors? starting at the broadest perspective, like what we really what matters is that we're delivering a a great valuable experience that makes users feel like they want to come back to YouTube, not just today and you know, spend some time today, but that they're going to feel good the next time they have a moment,

10:26uh what app are they going to turn to? We'd like them to feel like YouTube delivers the most value and that and that that sustains over years. Um so that's what our North Star is, is long-term user value. Um in terms of how we incorporate that into um a particular ranking decision, we have to look at at many different metrics because no metric on its own is a good indicator of value. You could look at like views, um click-through rate and that will only show you, you know, maybe how attractive the thumbnail and the idea

11:05were. a viral video, like click-through rate and average view duration will go down as the views go up. So it gets kind of confusing for me. Right, so we we do see that in in click-through rate that the more impressions something gets, the broader the audience is reached and that tends to make those aggregates um look lower than if you're only targeting uh just the core fans. Um so click-through rate is one metric, you know, watch time is another metric, but you know, it's not always that the time you spent is valued, which is why we also bring in the surveys. And, you know, even with the time, we look at

11:46both the absolute time you spend as well as the relative time because, um, you know, it may mean something different if you spent, um, 5 minutes watching a, you know, 2-hour-long video versus a 5-minute video. Um, and and maybe maybe there's a video that didn't deliver at the end of the that that you kind of hung on for a a long time hoping for a payoff at the end and it just didn't deliver. We don't want to just assume that all time is valuable, which is why we bring in those surveys to ask people, um, not just about their behavior, but to understand what was satisfying to

12:28them. Um, we look at likes, we look at dislikes, we look at comments. All these things, none of them by themselves are good enough indicator to just only look at that metric because it only tells you part of the story. Um, and so we do look at, you know, many many metrics. Yeah. And and one of the main things that's changed over the past few years is, you know, you might ask, well, which which of those is weighted the most? Which is the most important? And it, you know, the system doesn't even work that there's a a waiting for each of these that is the same for everyone. It actually doesn't

13:05even work such that it's the same for you at all times. The system now can learn that different different factors can have different, um, indicators of value, um, on different platforms, for example. Maybe on television, uh, the time you spend it might be more or less important than what what you click on, uh, and the click through rate versus on mobile maybe it's it's a whether you like or dislike because that's easier to do on mobile than it is

13:34on on a TV maybe. And so now the answer is there isn't an answer a single answer but all of this I think is focusing on the wrong thing because um if you're a creator and you're trying just as we are to understand value, I think that you need to look at a lot of different things yourself and those things should be coming from the

14:00audience. Yeah. And so you know, some of those are going to be your metrics that you look in in analytics. And I guess what I would say is look at things relative to each other. If something tends to have a lower, you know, percent watched or a lower absolute watch time relative to other videos that are similar length, then that's probably something you can learn from. Look at your retention graphs on your videos to see where people are

14:26relatively dropping off. Look at your like rates and see which which ones are doing better or or not. But then don't just look at numbers. You know, listen to your audiences as well. One of the ways I think you're very thoughtful about this is you don't want to punish channels for things. So like you said the algorithm will look at comments but creators can turn off

14:51comments. They can moderate comments. They don't have to worry that absolute number of comments or they can turn off ads and not have to worry about their video getting lower reaches cuz ads are off. That is something that you're very thoughtful about. Yeah, we're we're careful not to have the system kind of make a decision based on any one factor and we, you know, like when we started looking at integrating comments, it was very important to us that we didn't penalize people that for for whatever reasons they choose not to have comments on, we don't want that the system to understand that to mean like people just aren't engaging with this

15:28content. So, we do look out for those considerations so that we can um represent the true value and not just some artifact. Creators seem to worry a lot about uh kind of what the algorithm thinks of their channel and has the algorithm decided that their channel is like in one category versus another. going to punish me, Todd, if I make a

15:54mistake? Yeah, or like do I have the freedom to try something different from what my channel has shown before? And and this is where I'd go back to you know, the system isn't trying to find a single category for your content. Um and in fact, we are very video-centric. This might seem silly to you, but one of the biggest breakthroughs I had was

16:17realizing that puddles don't cause rain. Um because like you could look at a puddle and it's raining and think that the puddles caused it to rain, but really the rain caused the puddles. And with a lot of these things, um for example, people think that YouTube is pushing longer videos now, but people are watching on television and they want longer videos. And so, YouTube is recommending those longer

16:38videos. YouTube isn't pushing them. They're as you're saying like pulling them. These are audience sort of decisions that the algorithm is following. Yeah, we aim to not have the algorithm like have a point of view on you know, oh well, you know, right now we think that this is this is what audiences want. We want the data to tell us that answer. And so, my job is to listen to what viewers are are telling YouTube about their experiences at YouTube. And what we've heard over the past few years has been that viewers sometimes feel overwhelmed with all the choices that

17:18they have. And this happens both on mobile, and I think one of the reasons why short-form video has been successful, not just on YouTube, but across the market, has been it eases the burden of decision-making. You don't necessarily need Yes, you can search for content, and yes, you could look at thumbnails and decide on the home page which short form you want to watch, but you can also just say, "Just give me something different. Just give me something else." And and not have to put a lot of effort into choosing. When you're trying to be entertained, you

17:54don't want it to feel like work. So, we've seen that on the short form end. On televisions, we hear similar things. Like, viewers don't want to be picking up the remote every 5 to 10 minutes to make a new video decision. This is why series tend to work on television because viewers can think about it and decide a series they want to try, and then they don't have to decide again

18:19maybe for hours. Yeah. And and so, this sort of decision fatigue does drive us to make sure that we're delivering to those needs that that viewers have. And so, that can lead to more growth in longer videos and more growth in super short videos that are put in the the swipeable format. It's not like YouTube is saying, "Oh, well, we want long videos, so let's let's juice that." It's really being

18:49pulled from the audience. One of the questions we've been getting a lot recently as well is the redesign of the home page on desktop and the much bigger thumbnails. And creators asking, "Has that affected recommendations?" Yeah, specifically, I've heard people were concerned that it might drive less viewership on uh, long-form because there's fewer uh, long-form, you know, thumbnails uh, on the page. Um, and actually the, you know, this this change is probably now

19:25coming up on a couple years. Um, it actually drove more uh, attention and engagement around long-form. Um, and you know, one of the counterintuitive things that that we learned over the years is that often less is more. We used to have much smaller thumbnails on mobile. Yeah. Um, under the assumption that density and choice were, you know, we'd have the most success if we could show you as many things as possible so that you'll find one that that you like. But what what ends up happening with viewers is like I was alluding to earlier, um, sometimes choice is overwhelming and they just don't choose anything. And so

20:04we ran lots of experiments before we rolled that out. And uh, we had the same concern about, you know, what it might do to to viewership and we wouldn't have we wouldn't have launched that change if if it uh, reduced viewership. Um, and it helped um, grow long-form and short-form. And uh, it it seems to be making it easier for

20:30people to decide what to watch. One of the things that interests me too is like I think people don't understand how much you test. Like maybe they assume you just decide something one day and implement it. Uh, but you see often on social media people say, "Oh, oh, the the homepage changed." And it's really

20:44just a very small experiment running. Uh, is that how you learn and how you figure out which changes make sense? Yeah, there's hundreds of experiments um, that are running at any given time. Think of it as like we're trying to climb a mountain Yeah. to the to the peak of of potential of what YouTube can be. And often times when you climb a mountain, you can't just like look at the peak and go straight for it, although that would be

21:08nice. Yeah. Often what you're doing is you're you're going back and forth on switchbacks. And if you're paying too much attention to the direction that that maybe we're going as we climb this mountain back and forth, you might you might be just kind of thrown off. Like, "Oh, are they going this way? Oh, are they going this way?" Like, no, we're we're going to the peak and and that's you know, our eyes are on maximum value for viewers. So, I encourage you as a creator to join us on that mission and kind of recognize that um you know, we have the same goals, which are to satisfy audiences. But, know that

21:48we are and have a long track record of growing the pie for everyone. And that's how we reach the peak. To go back to the idea about having different topics, I think one of the things that's hard for creators is understanding that when you do switch topics, they could have very different audiences and very different view potentials. Uh because we see cases where a video blows up that's an outlier

22:13and gets way more views than average. And then the creator worries, "Oh, they're not going to know my audience anymore." Or a video just bombs and gets way fewer. I'm going to give like a personal example. When I did a big phone review, I could count on hundreds of thousands of views. If I did an accessibility video, I'd maybe get like a few thousand, 10,000 views because just the audience for that wasn't as big. Almost like you do the big blockbuster movie, you get way more box office than if you do the little arthouse movie. And I think we just have to set our expectations as part of that. We can do

22:41anything we want, we just can't get a billion views for everything we want. Yeah, I think it's it's unhealthy for creators to always just aim for the high water mark of of their their channels because different view different content does have different appeal. And you know, if you're experimenting in a new area, it may take a while to for both audiences to warm up to you in that area, for you to develop your skill in knowing how to be masterful in that

23:14content area. Um it's like starting if you're a if you're a restaurant and you're used to making sushi and then you want to branch out into pasta. Yeah. You know, you open the pasta restaurant, it may not be successful on day one. Just like your sushi restaurant wasn't necessarily successful on day one. I

23:32think it's important to have patience. Um and and give it a chance and and see how it works and this kind of goes into another question I hear about is like, well, should I do it on the same channel or a different channel? The advice that we give around channels is really to align channels with audiences. So, if you think that, you know, this to use the sushi restaurant and the pasta restaurant, if you think there's a connection between them and it's going to be the same diners in both, then, you know, the same viewers of both topics,

24:11then you can try it on the same channel. But if if you really suspect that, oh, well, the sushi is like an upscale, you know, omakase and the the pasta is going to be a quick quick, you know, grab and go type thing, then you probably want a different brand for that, right? Because you're really trying to give different value propositions and I encourage creators to think about channels similar to brands and make it easy for audiences to know whether they should expect the same thing or not. But also recognize that it's okay if you don't make the right choice between those because the system doesn't

24:52only work if you make the right choice. So, for example, um there are uh tactics that we use in the recommendation system to recommend things across channels. Yeah. So, even if you do create a separate channel, we can figure out that oh, people who watch this channel also watch this channel and direct them. Um and if you if you make a mistake and put them on the same channel and it isn't the consistent audience, we do try to even within a channel uh distinguish the videos. Yeah. So, if we can see that a particular viewer uh when they consume this channel, they tend to only consume the short content

25:37versus the long content, then we will try to, you know, focus focus the recommendations on the things that they actually want to watch and not on the things they don't want to watch. Or if they like this particular topic, like maybe they like your accessibility content and they don't like iPhone reviews. Um we're going to try to recognize, oh, well, this is a good video for this viewer because they like the topic. Um so, we're going to be smart about that and not just like, you know, well, there's only one way. I encourage uh creators not to have that kind of black and white thinking cuz

26:14that's not how it actually works. And that's similar, for example, doing long-form and shorts on the same channel, podcast, live streams. Yeah. To the extent that there's a viewer who's into all those things, we want to show all those things to that viewer. But as we get signals from a viewer that they only care about um one subset of that, it whether it's a format, a length, a topic, we're going to try to to match to what the viewer

26:41uh is most interested in. I think one of the things that's challenging is that there's no one answer for everything for every creator. A lot of it really does depend also on what the creator's goals are. There are some creators who just want maximum views and there are other creators who want maximum revenue. And there are a lot of creators who get way fewer views but a way more money. And there are other creators who just want to share their voice. They have a particular opinion on any like it could be on entertainment or on politics or on something and they just want to get that in front of the

27:11world or in front of the audience that they care about regardless of absolute views or or or money. And all of these things go into a a very personal decision for the creator about do I want to accrue everything on one channel even if it feels like some videos are getting less views than others or do I want to like will I manage multiple channels so I can feel like my podcasts are doing well on this channel versus my shorts on that channel. It all feels very

27:34personal. Yeah, and I encourage creators to be thoughtful about what their goals are and recognize that sometimes your goal with a video is not to maximize viewership. Sometimes it's to educate people about a product that you want to sell to a subset that are going to be really into that. And so this is why kind of going back to that high watermark recognize that people have varied interests. Your

28:06audience is not all the same. Um you have some core fans. You have some casual fans. You have some people watching you for the first time ever. And uh plan your content such that you address all those different needs and set your expectations um according to you know, you're going to have more views from things that are more broadly

28:30appealing. And I think that also confuses creators sometimes. We'll look at videos suggesting my video and they'll be from a wide range of topics and we'll go like oh once again, the algorithm doesn't understand my audience. Yeah, so people what you're talking about specifically is some people go into the traffic sources and suggested and then look at what videos are um driving views

28:56um the most. And assume that this represents kind of some understanding of the algorithm's targeting. And I think that yes, while while there there were impressions on those videos, um I think the important thing to recognize is that the reason that a video can appear next to another video is not simply just because it's related.

29:22Yeah. And we see that as we've moved the recommendations on the watch page from only being related videos to videos that can be relevant for a number of reasons, you know, most of which is like the viewers interested in watching the video. And so a lot of those results um you see are personalized results

29:43based on what the viewers interests are. In fact, we today still see um the number one place that people are watching videos from channels they're subscribed to is actually on the watch page in that suggested list. So, I know a lot of people think about the subscriptions feed as as one place people go uh to catch up on their subscriptions, but that's actually quite small relative to the people who just see something from a channel they're interested in while they're watching a different channel. And it's super common for users to want to change the channel. Um and so that's why

30:20you might see things that are unrelated. They're not related because it's the same topic, but they're connected because there was a viewer that that both that watched both videos. And uh we want to reach that viewer uh for you, not just when they're uh happen to be on the same topic. And and you end up getting a lot more views that way. So, I I wouldn't worry when you see that successful cross topic

30:46um traffic happening for you. One of the things you told me that was really impactful was think like a viewer. And I think like at a certain point my creator brain and my viewer brain separated because I would see that and wonder about it, but also as a viewer, I would go and watch a cooking video and then in the sidebar there might be a sports video and I would just click on that because I'm interested in that as well and then maybe from there I click on a tech video and to me it's

31:08like exactly that, changing the channel. But to the creator on the back end it's like, "Why is a cooking video recommending my sports video?" Yeah, you got to remember that many viewers watch hundreds of channels and they're not all related. People have very diverse interests. Last time we talked you mentioned about trying to help newer videos that may not have an obvious audience yet by giving them some visibility on the homepage. I know I'm still getting videos that have, you know, tens, dozens, hundreds of

31:38views. And most of the time they're very like they're very key to my interest. I don't watch all of them, but they are things that I would watch. Is that ongoing and how is that working? When a new channel publishes for the first time, there are zero viewers who have that channel in their history. Does that mean that they're not

31:55appealing to any viewer? No, and that's why we have parts of the system that don't depend on an individual viewer's history with that channel. And we're going to be using other signals that we have about the fit between that content and that viewer that aren't dependent on that viewer having history with the channel. We're still trying to make those as relevant as possible, so we'll look at things like the user's topical preferences and what we understand about what this video is about to try to make those matches when when we don't have the history and then we hope that we'll find some viewers who are

32:35interested and then build from there. And if we don't, then maybe that video just won't get more impressions. And you still and YouTube does care about creators making it. That's something that YouTube tracks. Yeah, we care about how many videos have a chance by being seen by viewers, and you know, we've talked before about how every day in recommendations, there's hundreds of millions of videos that get

33:06seen by someone. And we think that the recommendation system is better when it it understands more videos. And so, we do care about how many videos have a chance, and then we also care about whether those videos led to success for creators. And so, we have things that we track about how creators are growing through the ecosystem, and are they making a certain amount of revenue to be sustainable as a creator. Those are all things that we

33:36track and set goals to improve on. Sometimes the metrics are really hard. For example, we have creators who will turn off embeds or uncheck sub feed and notification because they worry that if retention is bad on an embed, it'll it'll hurt their video on the home page. Or if people like don't watch it in the sub feed, or if their subscribers don't like it, it'll hurt them on search, maybe. Can you explain a little bit about how you think about the traffic

34:02sources? Yeah, so for traffic sources, I think one of the main things to keep in mind is when we're ranking videos in the home page context, what's most valuable for us to look at from a data perspective is how well videos perform in that home page context. So, if people have a very different behavior when they come in through an embed or through an ad or through a link from a friend, if that isn't predictive of what they're going to do when they're shown a video on the home page, it doesn't make sense for us to weight that information very much. In terms of like not notifying

34:41subscribers, um some creators experiment with that, and some claim that it helps them, others not so much. I don't I wouldn't advise people to turn off uh notifying subscribers. Remember that we're not just using one audience's perspective to decide for all other audiences. We're trying to learn differently. How do subscribers react? How do non-subscribers react? How do people who've never watched this channel react? Those are those are different audiences with different expectations and different behaviors, and so you're not going to be We're not going to decide just based on the first hour or first 1,000 people who see a video

35:29whether it's worthy for everyone to see. And so, if you uncheck that box, I feel like you're you're not allowing the system to connect with a subset of viewers who are going to be interested. And this This whole subset thing I think is really important. If you go into your subs feed Mhm. traffic source under browse features, Cuz you can expand it and see each

35:55individual traffic source. I encourage you to spend time looking at that because it does give you one slice of metrics that is not uh primarily influenced by an algorithm. Yeah. And so, you can kind of see how one audience, your subscribers, is reacting to that to your to your upload, and you can look relative to other videos um at the click-through rate and the watch time and things like that to get a better sense of you know, the audience isn't really that different from video to video in your subscribers feed. It's always your subscribers. And so look at that to kind of get a sense you know, if you see that people are

36:38clicking less, it's probably something to do with the content and not who the video was targeted to because the targeting isn't changing. So, yes, you can you can learn that's kind of a clean way to isolate the content from the algorithm. Now, one thing you'll notice while you're in there is probably all your videos have about maybe a 10% or less click-through

37:02rate. Yeah. That means that even your best videos 90% of the time your subscribed audience isn't deciding to watch. The people who press subscribe and then went deliberately to the subscriber feed to look. Yeah. So, that says a few things. One is that the signal of subscribe isn't as strong as a lot of people think because it doesn't mean I'm going to watch it 100%. It

37:28might mean I'm going to watch it 10%. Um another thing it tells you is that that's normal. Yeah. It's normal for an audience that has been interested in your channel to not be interested in every upload. So, if you think that like your subscribers aren't aren't going to click on a video, well, your subscribers aren't going to click on

37:52most of your videos. And so don't worry so much. We already know how to deal with the fact that 90% plus of even a loyal previously loyal audience isn't going to resonate with every video. And so, we've built the system to find the subset of subscribers and fans who who are going to be interested in each video. So, don't over index on

38:20the aggregate stats. Yeah. I I love that we also have new casual and and regular viewers now because I think when you see, let's say 10,000 or 100,000 views per video, you just start to think it's the same 10,000 or the same 100,000 or the same million people watching every video, but when I looked at that stat, it turns out it was a lot of very different people watching each video. The number of people who actually did watch a high percentage of

38:43videos was very low. Yeah, I think creators also just underestimate how much other channels Yeah. their fans watch and that the portion of their audience that watches every video is so tiny. Todd, I could literally talk to you for the next 8 hours, but I know we have to get you back to work hand serving every video to every recommendation on

39:07YouTube. also know that creators need to get back to work. They don't have all day to listen to me talk and, you know, we say focus on your audience, uh, not the algorithm. So, hopefully this helped you get your algorithmic fix for now and, uh, you can go back to focusing on your audience. And you know the drill. Comments, questions, thoughts, drop them all below

39:30and keep it real.