 This is a presentation that is brand new. I've been working on it for over a year, and there's some data that I have today that's gonna make it worth it. So, you know, I'm super excited that you're here. A little bit of background about me. I've been on YouTube since the beginning, 2005. I own a company that focuses primarily on audience development and growth, and we work on monetization strategies. We do a lot of paid strategies as well. We get that world. I'm pretty much certified in everything to be certified in on YouTube and Google. I generated 33 billion views, 21 gold play buttons. Got me 21st the other day. It was really fun, and 19 billion views on Facebook. Here's something that you don't know about the 21 gold play buttons. I have goals in my life, and it's not necessarily putting the gold play button on the wall or anything, but I wanted to get it in every different vertical that I possibly could on YouTube just to prove that it can be done in every vertical. I truly do believe it, and we're gonna talk about that today, but we're gonna start with misconceptions on YouTube. There's a lot of different misconceptions on YouTube, and I get a lot of people reaching out to me as YouTube support, and they're like, hey, Daryl, there's something's going on with my channel. And I hear the terms all the time. My channel, my channel's dying. My channel, my channel, my channel. Now, if there's one piece of information I want you to walk away with is this. It's not about the channel anymore. It's not about your YouTube channel. Has it been about your YouTube channel for a very long time? What you need to do is start really understanding what it's truly about. When you switch the mindset of what it actually is, and you start delivering what YouTube actually wants, then it changes it. This is the key right here, that channels are no longer the dominant as they must were, but it's all about something, and the key is this. It's all about video, the YouTube video, and the viewer. And more importantly, getting the right type of viewer to watch your video. And what we're gonna kind of talk about this is really going over the algorithm and what it's looking for and helping you. And I can truly show you what you can actually do to improve. And a big applause to YouTube. I know YouTube has gone through some difficult times, but some of these that they're introducing now are game changers for you as creators and also brands. And I'm gonna talk about that specifically in this presentation. So, let's kind of go off of this. The algorithm, I know a lot of people really get frustrated with the algorithm. There are multiple algorithms on the YouTube. What makes up the different things that are happening on YouTube is the YouTube AI. That's basically machine learning. The machine is trying to learn and predict behavior for achieving the goals of YouTube. So, what do they want to do? Well, they're gonna learn and grow, and so there's a lot of updates that happen. There's over 200 updates that occurred this last year. And 2,000 split testing experiments that they go out and test things and try to gather more data to get the results. So, they actually set goals, let the AI do its thing, comes back and they see the results. And based off of that, they'll pivot a little bit and before you know it, it becomes an update. These updates are pretty big and they can affect us as creators if we're focusing in on the wrong things. And I think that a lot of creators that are creating content on YouTube really focus in on the wrong things and they're really missing the mark. And I'm gonna show you some of the things you can even look at. Goal number one for the AI is to predict what the viewer will watch. It's all about the video and it's all about the viewer. Now, when you can start figuring out who your viewer type is and kind of understanding that target audience and that target demo and really understanding what's there, then YouTube will help promote your stuff. It'll promote your videos out to the right type of viewer. So the number one goal that the AI has is try to predict what people watch. Why is that? Well, it's to maximize watch time because if they can predict what people will watch, they'll keep the viewers on the platform longer. Now, this is the key, is creators that understand this. Really just understand, look, I understand my target niche. I'm very localized on my channel. I understand what's going on and you're giving content that resonates really well with an audience and you get them to come back to YouTube day after day, then you are going to be blessed by the YouTube gods and you will literally start getting views. But today I'm gonna explain to you that it's changed a little bit and it's good for content creators, really, really good because understanding these goals will help you achieve your overall objective. Now, there's a paper that was released a couple of years ago by some engineers really talking about the algorithm, really going in depth on Google Brain, which is the engine for the machine learning on YouTube. There's a paper you can find online, really, really in depth. I'm still, my head's still spinning from some of the stuff that they're actually talking about in some of the equations. But ultimately it's the goal of predicting what viewers want. So I want to kind of explain an example of how the AI interacts. Now, let's just say that we have a viewer. This viewer loves to watch YouTube on their mobile device. Okay, so when they go on their mobile device, YouTube's starting to gather data on them specifically and they're starting to figure out their viewing behaviors and patterns. So they actually have viewing behavior and patterns for all of us and they're starting to identify what we're going to watch. What is the higher probability for us to watch next? So that viewer goes on the mobile device, looks at their phone and says, okay, these are the results that I get. I got video A, video C, and video B that's being suggested to me because I'm on the mobile device. Well, they get home, they turn on their Apple TV and it's get different results. Why did they get different results? Is because the viewing behavior of that viewer. Even though they're on a different device, they know, oh, when they're on mobile, they actually only do shorter videos and they have a lower tolerance for these types of videos. So let's suggest these types of videos but on Apple TV, the app, let's go ahead and give them different results because he actually did this person, this viewer has a longer view duration and will tolerate longer videos. And so what are we gonna do? We're gonna suggest them longer videos. So it's really, really intelligent of what it is. This is just an example of how the AI will work and think is really trying to predict what's gonna be best for the viewer. So it's gonna recommend shorter videos for this specific viewer in a specific case because of how they view on mobile and also they're gonna send longer videos for them on the YouTube app because of how they view YouTube on the YouTube app. So I hope that makes sense because what they're trying to do is predict that behavior and I can't reiterate this enough. It's always about the video and the viewer and if you can really start understanding and getting in the heads of your target audience and really knowing who they are, that will give you a better idea of how to actually use YouTube to start suggesting your videos. So what is the rule and traffic source on YouTube? Okay. What is the fastest way to go on YouTube? Suggestive videos. Seriously, it is suggested videos. When YouTube starts promoting your video, you will grow and I'm here to tell you that individual videos that take off will grow your subscriber base. But I don't want you to get so fixated on subscribers. I know that seems really odd because everyone has this and it's, I think it's just like the social status. I go like four million subscribers or whatever it is. But it's always about the engagement of the viewer and when you understand how to trigger that, then all the success will come with that as well as long as you understand the metrics and the triggers that will help YouTube decide and predict for your ideal target viewer. Okay. And so suggested videos is the fastest way to grow. In fact, I just saw a friend of mine started a YouTube channel and did it for years, but he had a breakout hit that picked up and he got 400,000 subscribers in a matter of a few weeks just off of one video. And so it's real. So there is a huge shift right now happening how YouTube suggests videos. And this is something I've been studying for years and I gotta put a disclaimer out. We manage a lot of channels and we only have data that we understand. And so I needed more data. I needed more data to draw the conclusions of what I was actually looking for. And I've seen a huge shift of how YouTube actually suggests videos. And I know a lot of you have reached out to YouTube and you've been into your frustration on Twitter because of some of the changes that occurred and you don't really understand what's going on. And so today my goal is to show you how YouTube is shifted and how to leverage it. Does that sound good? Okay. Big thanks to VidIQ Rob Sandy. Literally saved my bacon because I needed a whole bunch of information. And so he collectively gave me a huge sample size that I could deal with. So it wasn't just my small little 50 channel sample size. We had a huge sample size of information. Then let me show you what YouTube is looking at last year for suggestions. So on this there's a graph. The blue is suggested video. The red is subscriber. Now last year in April, the YouTube was suggesting a lot of specific videos in the first seven days. In fact, I was sitting on the stage or presenting on the stage talking about a seven day deterioration rate from your suggestion. You can see exactly what was happening in seven days. It was deteriorating. But things changed. And I wanna show you how things have flipped around and it's gonna help you change your strategy from it. In June, it goes up and this is on all the channels that were under 100,000. We'll show you some other results here in a second. You can see that where it changed here in September. There's a lot of people that are like, oh man, my subscribers are being notified and there's a lot of, my subscribers aren't seeing my videos. Look at that, they literally dropped. And then come, the big one was actually right around Thanksgiving. There's a huge change to happen in December. Notice that the blue line now is below, literally below and the subscriber is up higher. And then finally, April, here is the data. Now, this is what's interesting. If you were looking to the chart before in that first 24 hours, suggested a year ago, YouTube would suggest your video, 66% of those views would come in the first 24 hours. And then you can see the percentage just go down. However, in April, this year, look, it's only 28, 28%, but then you have 50, they're getting it out more to your subscribers. And so from doing this, I wanna show you the different things. We don't have a lot of time here, but it does change just a little bit for your size of the channel, but it doesn't change that much. And the big thing about it is YouTube is now normalizing this and really getting it in a way to suggest it in a way to get us all more views if you understand how to actually make that trigger. So that being said, this is probably the most fascinating one is the million subscriber one. Notice suggested videos was 66.7% in that first 24 hours last April. This April, it's 16%. But the subscribers numbers is 60%. So the more subscribers are actually viewing the videos, we're not gonna get into the whole notification debacle. Okay, so here's the conclusions to this. Views are being suggested more after 14 to 120 days. What does this mean for you as a creator? This means that you have a really long-tail strategy that you can get a lot more views after 14 days than you can before 14 days. You can actually start taking breaks now, you don't have to kill yourself trying to upload every day because a lot of the views are gonna be on the tail end. And if you do the strategy right, you can get a ton more views in less uploads. Also, this is probably the biggest thing for me is data gathering is now delayed in minutes instead of days. Now before, what you do is when you actually upload a video, YouTube would look at your metadata, they'd look at your title, your description, your tags, your closed captioning and then literally start to suggest that out to your own library and get suggested views off of that. And last year, it was really heavily pushed in that first seven days. Well now it shifted where they're able to gather the data that you need, disregard, I would say all the meta, but most of the meta and look at the next data points that they're looking at. And let me kind of talk about these data triggers because once you understand the data triggers that helps you understand how to shift your strategy and how you need to release your videos and such. So they're looking at the metadata in the first few hours. They're looking heavily, heavily, heavily, heavily on the viewer history data. And so if you have a viewer that's notoriously like religiously looking at your back catalog on a certain topic of videos and maybe watch two or three times the same video, well there, if you release something similar to that, of course that's going to be heavily suggested in that first seven day period to that subscriber because they're going to watch it. And it's really going to follow them around until they watch it because they're highly engaged with that. They're also going to look at people that are similar to that and help grow that and expand that, but that's more on the tail end of the suggestions. We're going to talk about impressions versus click through. There's only a couple things that ever got me really, really excited about the things that YouTube released and this is one of them. I'm here to tell you this right here is literally a game changer for everyone sitting in the audience. I don't care if you're a brand, I don't care if you're a creator, this data is so critical of starting the triggers that are needed and really help you to grow on YouTube. And so it's really, really important to understand this. And seriously, I don't know if there's anybody from YouTube here, but honestly my hats off, this makes everything all worth it because this data is what I've been just craving for because we never knew the impression versus click through rate. We could only guess and now we can actually make strategic data-driven decisions to make better content to get it more suggested. Now there's a couple of things that we're going to look at. We're going to go in depth on this and really break it down. But I want to show you the results. This is one of the channels that we had. The only thing that we did was have the data of understanding the split testing. Like we could only guess, oh maybe this thumbnail is not performing well, because all we were doing is guessing before. But all we did was change the thumbnail. That's it. There was no other strategy. We didn't change our content. We didn't do anything like that. We just changed the thumbnail and we had three times the amount of views just based off of that. That is huge. We didn't do anything to the content. All we did is spend a little bit more time on our back catalog, our underperforming thumbnails, made them better thumbnails and YouTube started to recommend them more. So much so that we got three times the traffic. The views on this is ridiculous too. So what is a good click-through rate? I know a lot of people says, hey I got this data down now. What is a good click-through rate? Is 5% good? Is 12% good? What's a good click-through rate? Now for me, I'm going to answer it a little bit differently. To understand YouTube, you've got to understand that everything's individualized and very niche-specific with video. So it's going to look at similar videos that are out there and see the click-through rate and see what their average is. And if you're higher than that average, then that's a good click-through rate. But for me, I always want to get a higher click-through rate because if I get a higher click-through rate, that means I'm going to have a higher average and YouTube's going to give me a higher percentage of recommending my content. And so what we're going to talk about is the long-tail approach. And this is something that... I was just talking with my friend at WWE. Like, this is a strategy that bigger brands and stuff actually do is the long-tail approach. They're not all putting out new content, but they're going to go through their back library and figure out how to maximize it because YouTube is now promoting older content really, really heavily. And if you can optimize that, you can get more views without even increasing your workload. So that being said, there's a couple methods, but after 28 days, I found that that's a good time to do A.B. testing on your thumbnails. Now, YouTube used to have a beta a long time ago on thumbnail A.B. testing. I don't know if they currently have that. So you might want to reach out to your partner manager and ask them if that beta is still open or where that's at. YouTube, if you're listening, if there's anything that would ever last the community of YouTube, it would be this, that we could actually do A.B. testing on our thumbnails. You provided the data for the impression you clicked through, but if we can pick winners of what thumbnails, then we can be more like Netflix if we could change the thumbnail for one demographic to the other. I would love you forever. And I wouldn't go to Instagram. Okay, going back to this, but the 28 days is really, really important because that's kind of what you get a ton of views and you can actually do some good split tests. Here is a split test that we get internally. You can see our click through rate was really good. I think anything above 12 is really good, but for me, I'm never satisfied because we can always do better. So with this, we had 14.2%. You can see it equated to over 100,000 views and so on by just changing the thumbnail. The thumbnail, no title, no description, no tags, just the thumbnail, it started getting recommended more where we're getting 23.1 click through rate. And notice the graph that YouTube's preventing 99.8% out to potential people that would be interested in that. That's huge. And what I would encourage everyone to do is have a strategy and it's worth hiring someone to design your thumbnails, it's worth hiring someone to implement this if you don't have the time to go back and literally do some split testing through your older content that's 28 days to 45 days or even older. What I like to do is go in, this is a Ninja tactic, but go into your real-time analytics, see what your top performing videos you can sort by last 60 minutes, 48 hours, and see which one's the 48 hours and which the videos that's on the rise that's a little bit older, it could be a year old or two years old or whatever. Just do this there and go up through the click through rate on that and then I literally will change that. I'll do an AB test on that one and I've seen it just go through the roof. Like literally get tripled the amount of views that I would have gotten otherwise. Peter Hollins, if you don't know who he is, he's on YouTube, he's pretty big. I really like him a lot. He sent me this graph. If you go to TubeBuddy, go to TubeBuddy.com. You can get this tool. It does split testing for you and it does it in a way that's okay. It's not the best way, but it's the only way that I found possible is to actually do good thumbnail testing, AB testing. Now what I do is I run these tests on my older content. I do not do it on the newer content and I'll kind of talk about that strategy of how you actually need to do that. So triggering data suggested video views. It's all about the metadata, your viewer history, your impressions versus click through rates. But then it's about audience retention and this is something that every content creator needs to pay really close together. Now, the new creator studio is amazing. I love it. It's like literally organizing in a way that common creators really understand how to optimize. And we need to really understand that and also true viewer engagement. I'm gonna talk about that here right now. I'm gonna give you a real example. I actually started a brand new YouTube channel and I have some partners in it. We got our first upload in May, May 30th, 2018. Zero views, zero subscribers. We didn't do any tactics. They aren't real YouTubers yet. Our target audience is very specific. Five year olds to 13 year old boys like we're really hyper focused in on that. We released some videos. This is one of our videos of the most powerful nerf gun mod that we did. It's two twin brothers that are my partners on it. Thumbnail is super, super engaging for sure. It's kind of telling the stories, very interactive. It's everything best practices that YouTube tells you to do. Well, here's the real thing is right here, we had 643 subscribers. And what we were able to do and what I really like to pay attention to is the average views per viewer is 2.7. And we'll talk about that because really understanding what that means can help you literally leverage it in a way that you need to. So here's their click through data. You're able to see that we're getting 50% recommended. It's a brand new channel and YouTube's recommending 50% of it. So when someone watches it, half of it's coming from YouTube suggestion right there. Our click through rate is decent. I wanna be in the 20s because I like to win. And then also our average view duration isn't as high as I like it. I like it around 50%. And so they're still learning and we're still growing from there. But the engagement is the key. Have you guys ever seen the audience retouching graph or are you looking at your data when you're releasing a video? What you wanna see is these little spikes that happen in the video. That is true engagement. It's not a like, it's not a share, but it's what the algorithm's looking at is a true engagement. That means they literally ring loud in the video and watch that part again. Okay, that is super powerful. Now, I've seen content creators get really creative with this to engage the community to as they're watching the videos that they catch certain things that might have something pop up after they wanted to do it. They get pretty creative with it. That's what you need to be looking at is how can we get these little pops to get higher than normal true engagement? Because when you get this, that's when you have massive growth. So really to understand how to actually get suggested views, you need to look at your impression versus click-through data about four days into it. If you do it earlier than that, it's going to be skewed the wrong way. You need to give it about four days and enough time to collect some data. And what I would do is if it's under the percentage of what you average, you get on an average, change the thumbnail immediately. When you're doing your thumbnails, I would design maybe two or three different versions of it, have it ready to go so that you can switch that out. I have literally seen channels that literally has stumped growth in that first four days but they changed out the thumbnail and the video took off. Does a couple of things. People re-watch it because they think it's a new video and they're able to do it. I'm not saying be deceptive, it's just saying let's get a higher click-through rate. Next, I would do an AB test. I would do an AB test at 28, 45 days. I would literally go back through a past catalog and literally pick videos that you're cherry picking to do AB tests. Don't go hog wild where you're doing your whole inventory. Like literally be strategic of like 10, 15 videos that you're always just testing and improving. I would go with some of your top performing videos. Do not, I reiterate, do not mess with the metadata. Just change the thumbnail. Has no problem. If you start changing the metadata then there's some other things that you don't wanna have happen. Focus in on audience retention. Try to get it in the Goldilocks zone. The Goldilocks zone is 50% of the people that actually started the video finishes the video and be very strategic with that. I found the drop-off rate, if you will just fix the last 30 seconds to 40 seconds of all your videos, you can get a higher retention rate. Some people draw it out way too long and they should just end it like within 10 seconds and be really, really engaging. There's even some content creators that just end and they're like, what, what, what did I just end? But it's like literally getting that retention to happen. And so the only way you can do that is look at the data, look at the videos that perform higher, look at the data that YouTube's providing us and seeing what's going on. You can literally play the video and see what's going on. You can make some assumptions, you can see what's being commented and so on. Really, really important with that. The next is try to trigger true engagement. Now, what I like to do, this new channel, like we're all about true engagement because I truly believe that's a way to get suggested more. You saw this new channel, it's got a ton of views. Ton of subscribers in a short amount of time based off of that and it's being more interactive with the audience and to engage. There were some ways to do that in the past that we can't necessarily do now but you can't be more connected organically with your viewer where they'll want to be more involved with your content. So in this, all this is, it's not new, it's best practices guys. Really at the end of the day, YouTube literally has given us the keys to the kingdom with giving our impression through click through and having that first data metric, that first trigger that we need to really make good data driven decisions so that we can actually say, oh, now we're getting a higher suggestion. Now, wait, we're getting some drop off. We need to work on our videos to actually get our videos to be a little bit better. It's a way that we can re-edit them or cut them in a different way. They'll be more engaging. And truly, I want everyone start thinking of, this is probably the biggest tip that I can give you, is think in video series where you have like four or five or six videos that all relate to each other. That is super important because what it's going to do, it's gonna feed off each other and start going from there. And if you can actually get it or similar channels that have a similar type of audience is doing the similar type of things, then you can start feeding each other suggestions as long as you have a lot of crossover viewer data. And so you're able to see a lot of channels. Like one of the ones that I really like is Steven Carter's share of Lizzie Sharer's channels. And that's what they're doing. They literally do videos that compliment each other and they're able to grow from there. And so realistically, at the end of the day, check your information. Don't overanalyze where you have paralysis but really be very strategic of looking at your click-through data. Whatever the percentage is, let's make it go up and then really look at your audience retention and go from there. Now, I've had a couple people that says, hey, I got 78% click-through rate. No one sees 78% click-through rate. I'm doing pretty good. I says, well, there's a little graph above that of how YouTube's actually promoting that. If you're not getting it to the 80%, 90%, 99% of YouTube's suggestion, then there's a problem with that. So yeah, your click-through rate's good but your other problem is in the other data that YouTube's actually looking at. So guys, thank you so much for coming. There's so much you can learn here. There's a lot of creators that can really do it. Now, what I'm gonna do is after the session I'm gonna go outside and I will answer your questions till I lose my voice. That's why I'm here at VidCon. I really wanna answer personally your questions and so on. Thank you so much for coming. I hope you have a great time.