YouTube's Algorithm Explained with Mr. Beast

[Applause] what's up fake gun [Music] [Applause] [Music] [Applause] is this on can you hear me everybody oh we i guess we can sit real quick all right i was getting people in the back i was expecting this room to be so big how's it going hey everybody we all right we don't have to do that but i appreciate it i appreciate it so wow this is crazy i think that's probably the most people i've ever talked to in real life or at least since cover it started yeah it's cool well thank you everybody for coming out uh real quick i just want to introduce todd because some of you might not know who he is uh yes this is todd so uh i think i think this talk could be very valuable to a lot of you i assume uh you all make videos and on youtube and stuff and so todd is actually the head of discovery at youtube and honestly yeah i would i would take as far as saying like there's not a single person on the planet that knows the youtube algorithm better than todd because he he's worked on it for you know a huge part of his life and so it's always very interesting to hear his response to the questions that i hear creators asking so what we're going to do is i'm going to ask him a couple questions and then he'll ask me questions and we'll kind of go back and forth because i think that's like an interesting format um and just one last thing and then you can uh jump in we we basically source questions from like discord because a lot of these questions i kind of feel like i know what he's gonna say but i think these are things you guys would ask him and vice versa so try to pick things that would give you guys the most value and should be fun yeah and i realized yeah why is this louder hey there's clearly a lot of mr beast fans in the audience um just want to remind everybody we're here on the creator track and you know it's really great for jimmy to come out and help uh other creators succeed on youtube so let's uh keep it focused on i think they're here for you yeah i know i'm just a support character [Applause] all right so we'll jump into it uh you should take notes or whatever i feel like this will be a pretty informationally dense uh uh chat so uh just to start things off uh mr todd the first thing i want to hit you with is uh what is your job at youtube because i feel like they'd love to know that yeah so i i leave the discovery product team and you know i go to a lot of meetings all day but what i'm really doing is trying to help make youtube products better for viewers so i'm really focused on helping each viewer open up the youtube app and find a video that they're just going to love watching and make them want to keep coming back to youtube the way i do that is i work with a bunch of like product designers to figure out how can we make the interface more fun to use uh make it easier to find videos work with a lot of machine learning engineers who are working on like pretty high tech these people have phds that are in machine learning that can process data and uh but ultimately it comes down to recommending the right video to the right viewer which is perfect because uh let's start with the fun one uh how does the recommendation algorithm work all right well here everyone get comfortable it's probably gonna be a long answer the secret is there's actually a bunch of monkeys in this room yeah is that why a monkey pops up when it's it's the four-wheeler yeah that's the that's the working monkeys whenever there's an error yeah they've been notified all right unacceptable answer tell us all right so there's a lot of complexity behind the algorithm and but to try to break it down for everyone there's kind of two key parts that we focus on one of them being personalization and that's really about what do we understand about each viewer and then the other part is performance and what are we learning about how a particular video is performing more generally as opposed to an individual person um the first thing that a lot of creators kind of get wrong is they think about okay i'm going to upload a video to youtube and then youtube's going to look at my video and it's going to send it out while that does happen for notifications when when you're talking about the recommendations on the home page or on the you know when you're looking at up next while you're watching a video those recommendations get calculated in real time when somebody opens up the app and hits one of those pages we gotta figure out what is the best video to show this person in less than a second and so the first thing we do is we start with personalization which is like what do we know about this viewer we know okay well most viewers on youtube it's not their first time so they have some history of videos they've watched before and channels they've watched before and so the first thing we're gonna do is lean into that and say okay well we know this person's watched mr beast gaming and uh so let's let's see if mr beastgaming has any new videos okay and then um so we look at the channels that they're subscribed to the channels they watch we also look at individual videos so you know whether somebody watched a particular mr beast gaming video versus a different one might might give you more minecraft videos yeah maybe they're more into minecraft and so one of the things that we do is look at like okay let's look at this mr beast gaming minecraft video what were all the other videos that were watched by everyone who watched that video and so we for each video we could kind of come up with a ranked list of the best other videos that audiences tend to watch and so we can go through you know many of the videos in your history and for each one kind of pull in okay here's some good videos for this video here's some good videos for that video we get hundreds of maybe thousands of examples of videos that we think this user might like well and i think what was interesting when we were talking before we came up it's like you're you know saying that it's not when you upload it's not like it just hands the video to people and all everyone on youtube is just magically on at all times it's more when people come on it finds videos that people want to watch which is kind of a different perspective than a lot of creators look at it which makes sense yeah so then um we also have some more sophisticated techniques that we're not going to do like a phd machine learning class here but i mean that would be fun it would be but um we use uh deep learning is uh what it's called and it's it actually models kind of like how the human brain works and learns where it forms kind of neurons and connections between them and you know the way that that the machine learns is we feed it a bunch of data about videos that people watch and it can find patterns and kind of make connections between videos and so we we throw everything we know about about a user and try to map them out and push it through the brain and figure out like what are the closest videos to that that person yeah so um so that's that's a lot on the personalization side and the other side i mentioned was performance so um we're collecting data in real time right so as we start showing a video to you know maybe we start with somebody who watches the channel regularly and then we look at what their history is and then we start distributing the video into the list of all the things that person's watched before and we look at like well what did what did how do similar people react how do people who like minecraft who watch mr beast gaming how did they react when we showed it to them did they click on it did they watch how long did they watch do they guys should be writing stuff down right now yeah that's good did they dislike it did they click not interested in their feed did they scroll past it did they uh you know another thing that we've been really trying to do over the past couple years is what we call satisfaction and what that's about is it's it's more than just watch time because not all watch time is equally valuable to the viewer how many of you have like gone and watched a video and spent like 10 minutes and when you're done you're just kind of like meh it was okay and then another video you may have watched you're like oh my god that was amazing it changed my life i'm subscribed to this channel i'm like watching this every day we really want to reward the videos that make people like really inspired and want to come back to youtube and the way that we try to figure that out is not just by looking at how much time people watched but by running surveys and we actually get millions of survey responses a day where we you know while somebody's scrolling through their homepage will say hey what did you think of this video you watched yesterday and um that helps us differentiate that stuff we feed those millions of responses into that brain and it kind of sorts through and finds out what's more satisfying and that performance you know is used to ultimately predict for the viewer that's showing up a bunch of different factors so it's going to predict if we show this to the to the person are they going to click on it it's going to predict you know and based on what it's learned from other users how long will they watch it what the probability is they'll like it and what would they answer if we if they did see a survey and we asked them how was the video i love it yeah and then we're able to rank um all those videos that we we came up with in the beginning uh and then there's some more stuff that happens after that because sometimes if you just rank the top 10 videos in order based on that score you might end up with like only one channel's videos at the top and viewers have told us like when they open up youtube they want to see a variety of stuff they just don't want like one channel at the top unless it's mr beast and then it's okay so uh so yeah we have some more stuff in there to diversify the results but ultimately all this stuff is trying to do is figure out what's the video you're going to be most interested in watching and that you're going to love and it's going to make you want to come back to youtube so one of the things we're trying to figure out more of is like not only can we predict what you're going to do with this video but what are you gonna do you know tomorrow on youtube you know how is this video gonna make you you know fall more in love with youtube because it's a great video and so uh yeah i love it i feel like we should give him a round of applause for that honestly that was good i don't i honestly don't know how it gets much more valuable than that um well i mean you can hit me with one of your questions if you want now i i feel bad because i don't all right give us much value so you were a small creator once yeah um there's a probably a lot of creators out there that don't have 97 million subscribers something like that yeah um what would you say to those folks in the audience that are maybe they're even just starting on youtube and they don't you know are not getting basically what would i do if i was a small creator yeah how would you get traction yeah this is a fun one um well the big thing is when when i was a smaller channel uh just to be honest the videos i was making sucked but i thought they were great at the time but you know i watch them now and they're they're pretty bad um and so you know i thought they should have gotten views and they deserve to get views but looking back like the views i got i was lucky to get you know those they're pretty terrible so for most of the time when i meet smaller creators my advice is usually just like make 100 videos and just improve one thing with each video because like if you're you know if you're not if unless you like made tons of uh films and stuff like that coming to youtube usually you got to spend a couple years some of us it takes a decade like me to refine your skills and get it where you're actually a good storyteller and good at making videos and good at entertaining people before it even makes sense for you to start getting views you know what i mean um because sometimes sometimes we forget like the numbers on the screen are real people you know what i mean and we're also there's infinite amounts of videos out there so until you hone your craft and get really good at it you shouldn't shouldn't even really worry about the algorithm too much which is funny following that up but uh because you know step one is just getting it where you're making content people want to watch and then you know then getting it where you're making content people have to watch you know what i mean um and going from there so ideally i mean you know if you're doing engineering videos like mark rover you can't do 100 videos so take it with a grain of salt based on you know the type of content you do but ideally you just make a ton of videos you improve something each time and you just do that for as long as possible and like you'll notice if you do that for 100 videos the 100th video will literally be like four times better than your first video if you actually improve something each time um it's pretty hard for it not to be you know and so now that's so that's basically what i did from like 13 to 20. i just made a thousand videos and the thousandth one was just like oh it's finally good people finally want to watch it um and so that's that's kind of my advice and then once you get to the point where you're like really confident and the videos are really good then i would start to worry a little bit more about thumbnails or other things in optimization but at first core foundation is actually making great videos and that's just a skill that takes time cool i i i'd add to that some things that i i've seen talking to some creators is um i'd encourage you to watch youtube be a viewer uh put yourself in the shoes of somebody who just has you know millions of choices and you know if you you may love minecraft um go to go to youtube and search for minecraft and look at the videos that are showing up at the top these most of these videos are from creators who have who have you know trained for a marathon so to speak and you know over many years honing the craft of video skills and they a lot of them you know dream for example found a way to innovate you know they weren't just doing oh i'm just making my standard minecraft house or whatever they found a way to introduce storytelling or you know find something that other people weren't already doing uh because if there's already you know ten thousand creators covering something exactly you're gonna have to do something different to make it interesting to people and make them want to click your video instead of somebody else's yeah and it'll build from there agreed all right so let's talk about ideas for content you've had a journey where you know you started out doing gaming content you went into like counting to a hundred thousand reading the dictionary you know tell us about your journey on on deciding what content to make and you know how is it today like you know how has it changed over time uh yeah uh that's actually a perfect one to follow up what you just asked me essentially just every video of my entire life i've always tried to just get better with each video and so yeah i mean that's how it started with just terrible videos with a terrible laptop and then we worked up to you know eventually recreating squid game um i don't know i just it's like it's kind of hard because i'd have to take you through the entire history of my life but just if every video you're improving something and you just you know probably now it's been 11 or 12 years those like little improvements just add up and over time it just you know it looks like um some big leap happened but it was like just very gradual over time you know which you probably already know that but uh yeah i mean i guess that's that's really the answer for me it's just like small improvements over time is like most people don't just go from like doing a minecraft video to then making squid game the next month you know i mean like there's usually like a four or five year gap in between to like get up to there um and then also for me i've just always reinvested all the money i made so when i got my first brand deal i gave it away and up to the state we just spend the money on content so that kind of helps so if you're always improving and you're just reinvesting the money eventually it just keeps growing do you feel like there's a do you feel like you're leading the audience in terms of where you're going in a content perspective or you feel like they're leading you like how do you feel like that push and pull happen yeah i would i guess uh i'm leading the audience in a way because i you know a lot of people you know think they know what they want to watch and you know they'll be like oh i want to see these and then you do something different that's bigger and like oh actually i like that you know um and so i try not to do like a series for too long right like i could like back in the day i used to donate to twitch streamers to see what their reactions were we just donate 10 grand and see what they would say and people love those videos we probably did like 15 of them but i can't do that forever you know that's not just going to be my entire youtube career so you do at some point you know at least for us we you know sometimes have to like make a decision like even though the series is doing well we need to evolve we can't just keep doing the same thing forever like if i made 150 of those i would just become the guy who donates twitch streamers you know what i mean and like that's obviously not what i want i want to keep growing and innovating and adapting so yeah i guess it's just kind of trying to kind of trying to predict the future instead of like running something into the ground and then pivoting kind of just be like okay it's probably about to run into the ground let's just pivot before it does and like find the next series and adapt yeah i often uh hear the analogy i think maybe henry ford or someone said that if you ask people like at the beginning of the 1900s what they wanted for transportation they'd save faster horses when really they wanted cars um and so you have to be careful you don't you know if you want to lead uh in innovation you gotta be a creative creator it's called a creator for a reason because you gotta be creative yeah and you know not every video did well so sometimes when you are innovating and you're trying all these different things you're going to have videos that flop and that's just part of it you know um so as long as you're you know one little 10 out of 10 doesn't scare you that's uh you just keep adapting and innovating and eventually you'll hit it um talking about small creators um just want to emphasize to all the creators out there that are wondering if youtube cares about like small creators or or uh just like big ones like mr beast and you know we do care about small creators we actually have teams that are dedicated to making sure that small creators can still break through on youtube um and we're trying to figure out you know it's a harder it's a harder problem whenever you have a creator that you know nothing about they upload a video maybe it is about minecraft it's like just knowing that this is a video about minecraft doesn't really tell you much about who's going to be interested in watching it so there's only so much you can get from just the the title and the description and really the magic of i think the recommendation system is being able to collect that feedback from the audience uh and figure out well how do people actually react to it that's the best way of knowing who's going to like it is by seeing who already likes it yeah and so these teams that are focused on small creators are looking at things like well how many creators today made it to like 300 subscribers for the first time there's thousands of channels that are doing that every day every day yeah every day really and so um and then they go on and we care about them moving up and so uh it's something we care a lot about and uh it's a hard problem but uh we're trying to find the balance so we call it uh the explore exploit problem so exploit is whenever you already have a lot of information about a video or about a a person a viewer and you're just able to kind of use that like once we know which channels you like and and which videos are are just being clicked by everyone that's an easy thing to recommend whereas explore is like okay we have all these videos over here no one's really watched them maybe there's a diamond in there that is the next mr beast how can we like find uh find out who's going to watch them and we're really trying to to uh find ways to introduce them to viewers um more intelligently yeah and even more elaborate on that i i know some people who just you know they they love youtube and they're just like patty gallios and people like that that just like study like crazy and i have a friend who like had uh helped a youtuber with like a hundred subscribers and we helped him like optimize video and it ended up getting like a million and a half views without us pushing it so i know it does work if a video is clicked and watched a high percentage and the viewers enjoy it it'll just get like a hundred views then the next day a thousand and ten thousand and just keep climbing um so it's funny to see it work in real time because you can kind of almost think like what it's thinking there it's like oh these ten thousand people liked it let's give it to 50 000. it's like oh they liked it and it keeps growing um which is what's interesting so you talked about um you know making better videos and just like getting better and better so when it comes to analytics and metrics as you're kind of learning what people like what what did you look at um when i was growing well when i was growing you can see ctr i don't know if a lot of you are new youtubers like five years ago they didn't show click-through rate um so it was a lot more brutal of a climb for us because we would like look at all the videos and all the data that was in the studio and we couldn't really ever figure out why certain videos with 70 retention and high evd did well and other ones with the same stats did bad uh and so that was like i was very perplexed for many years and then i whatever it was like four years ago you guys showed uh started showing ctr so we could see the click-through rate on videos and then it kind of made sense i was like oh thumbnails do matter um and so that that was great um but obviously you guys have it so for me the only thing i don't even really look at that as much anymore well after i upload a video i just kind of look at the retention chart and just try to see like where did i mess up you know why why are there divs how could i have made it better um and that's really i mean there's there's infinite amounts of data and like because we do channels different languages sometimes i look at the countries and there's like you use other stuff but 99 the time i just upload a video and then just see you know did i lose retention where at and like that's it's really all you need and then you take notes and then you you know fact check it against previous videos to see okay well i sneezed here and lost viewers that happened the last times i sneezed in videos or or is it just like a coincidence so this is just the boring part of the video um and then you just kind of take those and you gather enough notes over time where you kind of just start to see like oh well their attention isn't as great when the rooms are dimly lit you know people prefer a little bit brighter rooms or or whatever probably because sometimes it might be hard to see um and you just kind of gather all these notes and that's really it at this point um yeah because it's pretty simple if people watch the videos usually they do pretty well how often do you uh change a thumbnail on a on a video is that something do you usually just stick with one or you try well so i think we get better at making thumbnails every year so it's more when i upload a thumbnail i usually think it's good and i don't change it most of the time but what happens is like three years later i look back and i'm like oh that's a horrible thumbnail and so it's usually like every two years we just go back and update a lot of our older thumbnails just because they're terrible um just with like the new knowledge we have and what we've learned and funny enough that actually usually does help quite a bit when we update the thumbnails on old videos that they usually do see a little bit of an uptick because they're just you know if we take an old thumbnail where it's like uh seeing if whatever like a thousand a hundred thousand magazines can stop a car and it just looks terrible and then with the new stuff i know we zoom in a little bit so it's a little more visible and make things look better then magically it starts doing a little bit better because now people actually understand what it is because i'm not as much of an idiot anymore um so yeah i don't remember what you asked but yeah um do you usually have a vision for the thumbnail in mind whenever you start a video yeah i wish uh you know it would be fun i wish there was like a little projector could uh throw things up we usually sketch out the thumbnails actually chucky's somewhere in here he does it for me uh we sketch out the thumbnails beforehand to kind of see if we we like it like one of our upcoming videos we're gonna have people put their hand on a jet and then last take it off keeps it and so it's like funny doing the little sketches of it like because we're trying to figure out how do you show a jet but also like show people because people are small and you got to be able to see them so we just get it where it's like okay this is good because a lot of people do the inverse they film the video and then they realize oh i don't like the thumbnail but you're in too deep you already filmed the video so you kind of just you know have to make do or i'd rather have that existential crisis before i filmed the video that way i can make changes are there any metrics that you think creators might focus too much on or not enough on or uh metrics um that's a good one i think i think just in general for the most part people just don't focus enough on like i would love to hear more people which i think could you ask me back there you're gonna ask me in a minute but i'd love to hear more people talk about like how to become a better storyteller how to like improve ways of like making better videos and less about like how to game the algorithm or things like that because you know the thing i've noticed and i think most people noticed like the people who just make great videos usually do pretty well you know what i mean people who are doing things different are interesting so i'd love to hear probably just a little less talk about the algorithm because algorithm is literally just the audience you know you're just giving the viewers what they want to watch you're giving people what they're going to click and what they're going to watch and so it's kind of irrelevant that it's what more matters is like how do you make a good video that people want to watch and i would love to just see more people talk about that if that makes any sense yeah usually when when people come to me with algorithm questions they're like what time should i upload for the algorithm or like how often should i upload for the algorithm i encourage them to replace the word algorithm in their question with audience so like you really need to know your audience and say what time should if you're writing for yourself that is what i would write down like any literally anytime you use the word algorithm just replace it with audience like it because it really is true i hear i'd actually do the same thing because people feel like should i should i upload at midnight will the algorithm punish me and like it's literally simple it's like are your viewers on at midnight it's like it's just a simple question but you know and i'm guilty of doing that too and it's kind of funny once you just start interchanging the words things start clicking you're like oh well yeah a lot of them are probably asleep i guess i'll just wait you know yeah i'll put in a plug in terms of metrics that i think creators don't spend enough time looking at in that help them understand the audience and that would be the audience tab in analytics um there's a lot of metrics inside the uh analytics studio but you kind of have to click on the tabs to get in there the audience tab will tell you things like what are other videos that your audience watched this week yeah which that is interesting yes you can get a better flavor for like the interest of your audience that way it also can show you things like what videos attracted a new audience to your channel that came back to watch more and that's you know because as you remember youtube is trying to figure out what videos to recommend people so they keep coming back to youtube well if somebody watches one video from your channel and then they never come back that's kind of a sign that it wasn't that great but if they start watching a particular video and then they keep coming back to your channel that's a good signal we're going to use to to try to show that better they watch one video and they watch 20 on the channel then i mean yeah i couldn't think of a better signal like oh this guy clearly likes this channel yeah yeah so i would i would recommend the audience tab in analytics and then also there's another hidden feature that i think a lot of people don't know about it's it's a comparison feature where you can compare all your videos um against each other controlling for time like what what videos got my most uh views in like the first week or in the first 24 days yeah because you can group videos and compare to other groups which that is an interesting one you know for the most part on a video by video basis i usually just look at the retention chart to see if it was successful but obviously if we're just analyzing the channel as whole the grouping feature is interesting because you can like if you have multiple series on your channel you can uh group them together and see like which ones drive more subscribers or which ones have longer watch time or things like that like just a weird thing it's like back in the day i gave my 3 million subscribers 3 million pennies my 4 million subscribers 4 million pieces of popcorn like i would do these uh things every time i hit a subscriber milestone and so we grouped those up and then we compared it to a group of like normal videos i mean it's obvious but those videos like converting subscribers like crazy because obviously the theme of the video is subscribing so like people are just like oh i guess i'll subscribe um but you can get like cool insights like that when you group videos and compare to other groups yeah and then the the comparison feature i think you go into advanced mode and then like comparison um and you can look at like uh compare videos like based in the first 24 hours yeah i feel like i haven't asked you a question a while let's see uh um this one looks like a fun one um so can a video's potential be stopped if uh the channel subscribers decide not to interact with the new video yeah this is a question we hear a lot of creators they're concerned that maybe they're trying to experiment with a new format or a new type of content and they're worried that oh if my existing audience doesn't like this you know and so the the algorithm is going to recommend it to them and then they're not going to click on or they're going to click on it and then not not like it basically if your subscribers don't like the video is it just dead in the water and never going to live yeah so you know again we want to get the video in front of the viewer who wants to watch it and you know if you actually look at your analytics you'll find that it's pretty rare that your video is only shown to subscribers um in fact most most channels i know most of their their views come from 90 most reviews are from non-subscribers yeah and so uh we're not going to only show it to people who've watched your channel before um because the subset of subscribers that do watch it and some of them will and like it we're gonna see what else those people like and then show it to people like them and not people like um the ones who don't watch it and your subscribers generally are actually pretty diverse you like you know humans want to simplify things down and kind of assume oh every subscriber like they just want to watch all my videos it's not that simple people have different interests they want to watch different things and we try to find the patterns that work and you know we don't just look at oh well a few subscribers didn't like it so no one's gonna like it because that's not true yeah but still if you pivot too hard and you know you're making videos about painting bowling balls and then one day you just start doing videos about massages then it's gonna you're gonna basically alienate a lot of your subscribers so it still will probably eventually find the audience that likes videos about learning how to do massages or whatever but ideally when you do pivot you do it slow over time uh and not so just harsh because that is one thing uh just because you inspired me when you were talking about that a lot of people when they like try new content they're like oh the algorithm hates me for you know trying something new or the algorithm hates this but again replace it with audience is like imagine you're watching spongebob and you watch 100 episodes of spongebob and the 101st is just adventure time you'd be like what is this this is confusing you know um and that's kind of how your audience is too if you make a pivot too quickly so ideally it's like just a little bit over time and you slowly switch over to new content it's not super abrupt yeah i'd also say be patient with uh with the with the traffic too because if you if you are pivoting to something new and you're you're kind of used to if you're being consistent and your core audience is watching something and then you switch it is going to take the algorithm a bit more time i literally had that happen with someone like two days ago who's freaking out because he made this video it was around like the uh trial that was going on with the johnny depp stuff and uh it was it was a pretty interesting one breaking it down and i i thought i told him it would do well and then he uploaded it and it didn't do as hot and then i was like just wait a week and he's like freaking out every day like well let's do all the stool and then like after like six days he's like oh it's doing phenomenal and you see it pick up and exactly especially like the more further it is from what you normally do you just gotta give it a little time to get picked up yeah don't obsess too much on the real time stats i know that when you upload you're you're pretty excited to see you know was this a good video um sometimes it can take a few days and so just have patience and uh don't sweat it too much love it let's see um okay here here's a good one so what if you were us like what what would you do with shorts or or what are how are you integrating shorts with long form or or whatever i just want to kind of like inspire you tell us about shorts and how do you think as long form creators should approach them yeah i think the first thing i'll say about shorts is that uh we're kind of still in the first innings of shorts uh it's been amazing to see the growth of it i think we just announced it like we have now 1.5 billion uh people a month that are watching shorts and uh over 30 billion views a day um and so there's a lot of viewer attention on shorts and we're actually seeing a variety of different ways that creators are using him so of course we have creators that are already doing long-form content experimenting with shorts some of them you know i think there's different strategies and my advice to you is to try to figure out like what is your goal with shorts are you trying to deepen engagement with your current fans and um that might have one content strategy or you could be saying hey i want to like experiment in a new direction um you know maybe maybe i could use this as like an experiment platform to see what are some other things i could do without kind of uh you know taking risks in my in my main long form section yeah um how about you do you think that you're do you think about shorts as serving your current audience are you trying to acquire new new viewers yeah well i think you're hitting on the head it's still a little new so we're all still figuring it out i i will say i've i noticed a lot of people have started doing them like ryan trojan arac and just a lot of my friends in general are doing them and it's it seems to be a huge snap positive i mean they're getting more subscribers it doesn't seem to be hurting the long-form viewership so it uh it is cool to see how it's developed actually one of my friends dylan's here he was just uh showing me a couple days ago he's started going really hard on shorts and we're just looking at how it's actually helped his long-form content a little bit like some of the people are watching the shorts actually going and watching his longer videos um which is cool because he didn't upload a long form video then he uploaded a short that went viral um and then his revenue like skyrocketed even those shorts don't make money because just his uh long form video just started getting more views um and it was driving a little bit traffic over so yeah i know a lot of creators including you have been concerned about like oh well if i upload shorts and you know they don't monetize the same way as long form is that gonna like hurt my channel and so we did some data analysis of that we have some data scientists that looked at channels that were doing long form only and channels that were doing long form and short and we did see that the ones that were doing shorts as well were actually growing faster and had more success and so it's not going to like reduce well i also i think it's clear that's kind of the direction things are going like short form is here to stay and only going to get bigger and bigger so you know i think eventually even if it's not ideally perfect right now eventually you guys are going to figure out a way to fully integrate it in and so it's just a thing like if you're not really doing it now like you should be yeah one one story to give you a sense of like how things work in in my day-to-day when we first launched shorts uh one thing that happened with the algorithm in the in the long form is that a viewer would go and maybe they go into shorts and watch like 200 short videos and then they'd they'd go back outside of shorts to just kind of like the main youtube and the algorithm actually wasn't used to seeing viewers watching a bunch of short videos and then going back to watching long videos and so when when the algorithm saw a viewer oh look the last 200 videos this viewer watched were all under a minute this viewer is like they only want to watch short videos and so it it basically stopped recommending those viewers long-form content in the main youtube and and viewers you know started complaining to us and that one of my jobs is to listen to what viewers are saying and and so what we did at that time was say okay the algorithm doesn't know how to handle shorts and and treat that the right way so we're going to split off uh the history for the viewers so when a viewer goes into the main youtube we're gonna not pay attention to what shorts they watch because we don't want the algorithm to get the wrong idea but but what that means is since we did that that there isn't much of a bridge from short to long form so if if a viewer does discover a channel in shorts and then goes back to long form right now that the system doesn't actually know that they what they discovered that channel because we disconnected it and so one of the things the team is working on right now is building that bridge back in and making sure that the system can actually understand oh let's give the system the context that they watch these sorts of videos in shorts so they were in like a shorts mode and then it can learn oh okay well when the user's in shorts mode i'll give them shorts and when they're not you know i can i can understand that context that's that's the types things that we add to the system so they can do the right thing so much so much information just real quick is there like a certain amount of time or can we just keep uh blabbering i still have more questions we got about ten more are we ten are we allowed to go over i'm having fun um i don't know i'll go until they make me get off the stage um so let me see we kind of have like we ramble a little bit so a lot of questions i was going to ask you we kind of already hit so i just want to like let me let me ask you one about um so you you have a a big user base now but i also know you want to grow um do you do you feel tension between serving people who've been like following your channel for like five years versus like trying to bring in somebody new who yeah like one of the things that i remember talking about a vid summit i was looking at your channel and one thing it was before squid game came out um so you hadn't had that mega banger 200 and are you 200 and some million view video yet and uh we were looking at the returning viewer and new viewer uh data and what what surprised me is like a about a third of your viewers that month had not watched your channel at all in the previous year yeah and that was like amazing to me that such an established creator was still getting you know a third of its audience from people just like just watching for the first time is anything you want to say about like new viewers and yeah uh i i uh that one's definitely uh one where i don't there are different approaches that work like you know there are people who have like super tight-knit communities are killing it like hypothetically moist critical or ludwig and then you know people who make a little bit more broad videos like me so i want to be careful because i don't want to like tell like imply that there's only one way of doing it but for us we like to make videos that anyone could just jump into any video and basically you know start watching and understand it so because you know if you want to get as many views as possible that's like kind of one of the things you got to be able to appeal to people you know for the most part like if someone's 40 wants to watch it or someone's 15. you know what i mean and it just so happens that my humor and the way i act just kind of lines up where we can kind of do it so yeah to answer it i think we try to make videos that newer viewer viewers would enjoy as well as people who've been watching for five years but um you know other people you know have a lot of inside jokes and a lot of just you know um less you know bigger buried entry but once you're in there you're kind of just like a lifelong fan and so kind of the any approach works you know what i mean um so yeah it's one of those things where i don't want to like push people one direction because there's just infinite amounts of ways people go about it and they they're all successful you know let maybe let's go back to storytelling so um yeah at vidsummit last year uh hayden hillier smith got up and he's a top editor for a bunch of uh top youtubers yeah and he's started editing some of our videos he's great he kills it yeah yeah so he he got up on stage and basically said stop looking at analytics uh people are too obsessed with analytics and they've forgotten how to tell a story and so i know that that's you know it's kind of funny hearing you quote that yeah well i mean jokes i know yeah no i i think that that when you do tell a story it affects the analytics right and so it's a tool for how have you thought about how does storytelling work like how has that been different for you and maybe some of your early videos like versus your your videos that are more focused on storytelling how do you tell a story yeah that's where for me i've always struggled with it because you know since we get so many views of videos sometimes it's like hard to tell who which people you know to give depth to when we have people interview like if we have 50 people put their hand on the lamborghini you know which ones like give depth in and how much depth do you show for a person because you know a lot of storytelling is getting people attached to the audience attached to someone so then like the payoff at the end speaker or whatever um so that's something honestly i'm still trying to figure out because a lot of times when uh people are explaining things i feel like it's pretty boring and i usually cut it out but then other people are like oh my gosh that's so interesting i would love to know more about them um but in general there's a book called uh save the cat which hayden it's funny we keep talking about him but he's obsessed with storytelling that's all he ever talks about so it's it's fun because every i talk to him like a couple times a year and every time i do he just brain dumps me on like a million things he's learned but he recommended a book called save the cat which is kind of a book about that i won't go in depth but if you want to know more about storytelling i thought that was pretty cool um yeah i thought your your most recent video with the like uh solitary confinement uh yeah the guy that was in there for many days yeah um was there any like um back and forth on that one in terms of like retention and storytelling or yeah i mean that one was a little bit easier because it was one guy we basically put a person in a room and every day he stayed in the room we got ten thousand dollars um so since we he went i think 23 days we had plenty of content to like build the story around and kind of play with so that was kind of just cheat codes because you know we have like 300 hours worth of footage so it's pretty pretty hard not to like make a pretty good story out of it you know what i mean um here let me see if i got anything for you let me flip through this um what um what interesting trends are you seeing on the platform um there's a lot a lot of things happening so when the pandemic hit one of the surprises for me was the you know i think everybody kind of expected there was a viewership spike right because people were like stuck in their houses they didn't have much to do they watched youtube but what surprised me was like the creation spike because i was kind of thinking well it'd be harder to make videos because you can't kind of get together with people but we saw a surge in creation and then you know especially in live live has really been taking off on the platform interesting and so uh i think you know people are using live as a way to just feel more connected to other people because it's more of a simultaneous experience they can chat with each other things like that um you know also during the pandemic a lot of young people like had to learn how to learn online right instead of going to school so learning has been growing um shorts obviously is a is a huge thing yeah um let's see let's uh kids just got out of school or whatever you know so like how does that or like what are other seasonal things you'd love to yeah seasonal things typically happen like what you'd expect when school gets out is that um rather than having a lot of like spiky traffic on weekends uh you get more consistent traffic throughout the week okay and so so like on uh during the summer like or during when people are in school and busy during the day you know obviously traffic spikes and weeks that's why i upload on saturdays um and that is something i've noticed as well but during the summer it's way more consistent like but we basically say every day is a saturday you know like two months of the year yeah or it might spread out a little bit more so if you're used to like uploading on a spiky saturday and now that school's out you know people aren't necessarily just waiting until saturday they're watching throughout the week you may see fewer views on saturday but by the end of the week it's the same or more so yeah that's why you shouldn't focus too much on like the real time because you know it could just be people are well eventually doing other things if someone watches your videos eventually when they log on it will pop up there because like you talked before when people open up youtube it brings them videos not when you upload it finds people for the videos so right yeah another big uh thing we're seeing is is the growth in watching on on living room uh like gaming consoles or smart tvs uh that's the fastest growing platform at youtube mobile is still the number one way people watch but um you know we're seeing huge growth in tv so that's something for you to think about what's going to look good on a big screen or a small screen yeah i wanted to see if i can see do you can you see it on your phone your tv viewership so i agree i mean most youtube i watch is on a tv now now i'm curious if i can just say what it is oh well can't find it um but uh yeah that's you know obviously i i have never used cable in my life and so and now i've mostly just watched videos on my tv every morning i just wake up and watch wind over production probably matter and stuff like that on the tv um what what's something that so you're always trying to make better videos now you have a team you know they don't may not have as much experience making videos as as you um what are you all kind of finding a need to remind folks to not forget about doing when you're making a trying to make a better video is there anything that comes up that's like yeah the people that work with me um that's a specific question what do i usually need to remind them of i mean i think a lot of people that i work with usually forget like they try to find like what's the easiest way to do something or you know what i mean like if i need a submarine to spend 50 hours in or or you know if i would have put a basketball court in the cloud and play basketball in the sky or something like but instead as one does yes yes uh but instead of looking at like how do we make the best content possible even if it's harder usually they look at it what's the easiest way to accomplish it which usually you know there's a harder way that makes better content and i would love to you know i like to be able to make the decision on the trade-offs you know like ah it's not worth the effort but this is um so for people working with me that's usually the biggest thing it's like adjusting your head to like it doesn't really matter how hard it is we're just trying to make the best video possible which is like obviously in traditional film like kind of the opposite of how they do it is there anything that like you forget or they forget to um remember to film or incorporate into the video um um let's see well one thing it's funny how i like how specific they're getting but um we i like to try to end the videos with like a joke or like some type of emotional response like like if you watch a video and you laugh at the end i just feel like you're more likely to watch another video because it's just like oh it makes you think the video was better than it is um so i usually try to like end the video with something cool but we forget a lot like you know because if we film for 50 hours and at the end you know i'm in the desert i usually just like i don't care i want to go take a nap um but ideally yeah that's since i think you asked like what are things we forget sometimes it's that usually trying to end the video on an emotional note because we're usually so worn out by the end we're just like oh oh well uh let's see i wish um you know i wish i had some more juicy stuff to ask you but we we kind of we kind of hit all the questions i really had for you i mean like i have a couple other ones but when we were just rambling yeah guys are we covering them no i mean let me think um mike what are you excited about with the what the future or whatever on youtube um i'm excited about like how everybody's kind of woken up to the fact that the creator economy is a thing and it's like i remember in 2020 when the pandemic hit and we were seeing a surge in activity like news outlets started focusing on oh there's actually like a creator economy where some creators can like make money on the internet yeah and we were like what welcome to the party we've been doing this for a bit and uh you know i'm excited i i've been at youtube over eight years and and i stay there because i really appreciate the values of the company like really trying to do the right thing by creators like 15 years ago we started sharing revenue with creators uh on their videos 15 years yeah since the partner program started wow yep and yeah i think the next era that we're moving into is you know a lot of people are talking about multiple formats and you know whether it's short form long form whether you want to do image posts with memes or live streams or you know live shopping seems to be something that's coming up as a trend i heard uh in october there was a creator in china that sold almost 2 billion worth of uh of lipstick or makeup whatever yeah in one day yeah which is crazy if you haven't uh that's actually true if you want to google it and read about it um you know actually what i was thinking if they don't kick us off if any of my friends out there if you want to just text me some questions ask him or text me things you want me to talk about that would be fun like what are what are some things i'm missing um and i don't know if you want to hit me with another one i mean here you've been waving your hand for forever let's do it just go ahead come on up come come up to the stage every time i look at todd he's been waving in the background so this whole time thanks very much my name is joey so i have over two million views under my belt been doing this for that's good uh actually since i went to my first vidcon my daughter dragged me to it nine years ago i said wow you can make money doing this okay um here's my question is that we know that ctr and abd are the two most important things for the algorithm yeah audience yeah and and that's click-through rate and average view duration um and and um lately youtube's done two things i think to impact avd one is the introduction of chapters which basically means that hey tell us the chapters like in a book and then the viewers can skip to the good parts and it's like well that seems like that would lower abd the other thing is that now we see this this much-watched wave you hover a video and you can see oh these are the best parts i feel like that kind of impacts how long they watch because i can just skip the boring parts and get to the good parts so here's my question and and by the way i saw an interview with you on creator insider channel which everybody should watch if you're in the business of making money on youtube you'd be remiss not to watch this channel it's fantastic it's usually on your dashboard it just pops up whenever they have yeah yeah i watched it before it was on the dashboard it's really good and it keeps you informed on everything latest but somebody they asked you a question once and what's more important um um the the amount of view time absolute minutes or the percentage watched and the question was something along the lines of what's better 10 minutes of a 20-minute video or seven minutes of a 10-minute video and use your answer was well more is better like okay more minutes or more time we've never you know i've been doing this for eight years and i still don't know the answer to this question i i mean that one i can hit you with uh on the i there was a couple months but i'll start with the last one um the the answer from everything i've seen uh no let them hang out with us up here oh we're kicking them off um so yeah uh it's fine it's uh so ideally you know in a perfect world the longer the video is that you can still hypothetically hold 70 retention the better yeah because you got to look at it in a sense of like you're also kind of competing with every other video and stuff like that so typically the harder it is to do something the better it will do you know if you can get 70 retention on a 30 minute video i don't know i've never met a single creator that could do that at scale so of course it's going to go more viral than if you got 70 on a 10 minute video well obviously while people still rate high end surveys and like like the video so i like to look at more of a frame of mind as like what is more difficult and typically what's more difficult is done less so we'll be rewarded more because the algorithm's like oh this is this is insane we've never seen this before audience you know oh this is insane i've never watched a 30 minute video and actually loved it before this is crazy you know what i mean uh so that's kind of the stance i like to take uh make i try to aim for 70 retention because like i just whenever i look at other people's analytics i usually see 50 or 60 so i i feel like 70 is just so crazy or if you can get 10 million people to watch on average 70 of a long video it's like that's just you know it's kind of like it just is a good signal that people enjoyed it and it's a good sign um and so i try to aim for 70 and then you know at the longer the video is the while still maintaining it typically the better in my opinion i don't know if you want to jump in yeah i mean the the message is it makes the make the video as long as it needs to be agreed while you keep the audience's retention and if you see the retention dropping then you're probably making it too long yeah because there's no way you're going to get a 70 retention out of you that's just not phenomenal and then back to your point about like chapters and the graph of like stuff people are replaying um you know one of the reasons that we we do those things is because we get feedback from viewers who say like you know creators are making videos that are too long and they're not getting to the point and i just want to get in there and see the thing that i came for and because i would never say anything that makes the viewing experience better is negative because that's ultimately how the platform keeps growing and how we keep our jobs in 10 years you know i mean so if people really are like oh i'm fed up i need to just find the highlight of the thing i want and go to it they probably were going to do that anyways it's kind of my gut reaction yeah if you're seeing your retention go down because people are jumping around it probably means you're you're putting too much filler in there and uh you know we have to give the viewer the tools to be able to like find what they want because if they don't like some other platforms are going to figure it out and people are just going to migrate because you know creators there are have the right incentives to like get to the point and we're like oh no this one's only 13. we got to do something to make it longer but you don't you just do whatever makes the best feel possible we've gotten 100 million views on like a one minute video a four minute video an eight minute video a 12 minute video a 25 minute video like i we've got ultra viral in every length of every video i've nev i've never seen anything that says the videos have to be longer you just have to make videos people want to watch really um so that's what i would say thanks it was fun chatting thank you have one over here for you [Applause] can i ask the question uh people i got some texts okay if you want to uh hit some of these so this is from renee um are thumbnails less important in the age of mobile and autoplay i don't really know what that is but that was the first text i got yeah i think uh thumbnails are still important um if again like remember how i said you need to spend time using youtube and watching it and looking at through the eyes of the viewer if you do that you'll notice that on the home page of youtube we've been moving more towards previewing the video just starting playing it automatically and like last year we made it possible for people to actually unmute and just watch right there on their homepage so uh how do you unmute it when it's out of playing there's a little icon on top of the thumbnail and once you unmute it for one video then all the other videos will be unmuted as well forever or until you close the app uh until you close the app okay um and so what we were finding there is if we spend less time like just showing a static thumbnail and just get people right into the video you know they can make better decisions about which videos to watch and so in the homepage environment you're gonna want to be able to have a good intro to your video right yeah um what techniques do you use uh to kind of hook people in the beginning yeah well this is actually something we were talking about a little bit back there i have noticed that uh because agreed now a lot of people you know are on their phones and videos autoplay on home page which i think they should because sometimes you're like i don't know if i want to watch that and then when you get the glimpse of the first five seconds you're like oh okay i do want to watch it and then you click so it helps you like you know be more confident before you click or tap whatever so one thing i've definitely noticed is like that when we have it like not dark in the intro when it's like a little bit brighter like it tends to do better um which i notice you also use you know you sometimes show captions at the front so for the people who do have have it on mute they can learn more what the video is about things you can do just kind of assume that like i don't i mean you'd have to look if you really wanted to what percentage of views came from mobile and browse on mobile and then you could see and probably a lot of those people they're not really even tapping the thumbnail per se that's more they're tapping after they see the first couple seconds of video so when you keep that in mind you can kind of gear the first couple of seconds of your video like okay like he's saying someone's going to watch this on mute they're going to see this how does this reinforce what the title is and make them interested yeah i mean it's a small thing and it's not like what's going to make or break it but obviously all those small things add up i think we have time for one more question yes uh thank you a lot for your conversation i have a question for mr beast uh actually i know how to increase their their revenue on your videos into three times more and produce much cooler videos than the squid leaves we have an idea uh and we got investments from creative director of apple and also from founder of wine and we have a platform uh where you can sell your personal items from your video of uh and as a collectible items for example you got the question yeah just a second you got the video with the orbeez remember uh and you could listed this video on our platform for example 10 000 orbeez and sell them uh for example uh for ten dollars and you can can i speak with you uh after and and all right you're good uh yeah i mean i would probably just if i wanted to just list them on my merch site you know just like no it's not the marriage it's a collectible items that you can uh sell from your content because uh you can make a cool content yeah i mean i don't know if i wanted i would probably just sell it on my merch site though just to make it you know easier because like yeah but we can organize all of you everybody we appreciate you thank you jimmy thank you that was fun wait did did we go over i i am curious though are we will we just be able to talk here forever okay i have a question how many people here have already done the shorts drive through where the mr beast gun ball machine is hey let's go okay i i see that there are a few hands that are not up um i hear there's going to be some you know cool surprises there this afternoon so i would encourage you if you haven't checked it out go to the shorts drive through and uh check out the gumball machine thank you everybody for coming out it was fun i hope you guys learned a lot thank you all right take care go for it take a picture yeah just take it quick

Channel: VidCon

Views: 117,114

Keywords: VidCon, vidcon now

Id: fQPUH2Ylb88

Length: 61min 45sec (3705 seconds)

Published: Wed Aug 03 2022