How Cocomelon EXPLOITS The YouTube Algorithm | Little Monster Media Co.

– Hey guys, what’s up, it’s Matt. Today we’re gonna be
taking a look at Cocomelon. Cocomelon is a kid’s
channel and if you are not a kid’s channel you might be
saying, well, this might not be super relevant to me. But I implore you to keep watching because I guarantee you, it is. Kid’s channels on YouTube
are some of the most intense and data-driven channels
in all of YouTube. I’ve spoken with hundreds
of different creators in the kids space and they are
some of the most insightful and thoughtful and data-driven people you can find on YouTube. These guys analyze every
single piece of data. They experiment, they test,
they look at those tests because it’s incredibly
powerful area of YouTube in terms of viewership. These channels are doing
billions of monthly views and there’s no better place to start with our kind of analysis
around that than with Cocomelon, one of the fastest growing
channels on YouTube. Currently the second most viewed
channel on all of YouTube. So there’s a lot that we
can take from their channel and apply broadly across YouTube. Maybe there’s, you know,
an idea in something that they’re doing that can
illuminate something for you. If you are a kid’s channel on YouTube then this is hyper-relevant. This is the number one
channel in your vertical and we’re gonna go and take
a look at their content, their channel, what they’re doing and see what we can learn from that and how we might be able to
apply it to our channels. So let’s dive in and take a look. So, as mentioned, right now
according to SocialBlade at the very least, Cocomelon
number two most viewed channel in the last 30 days, 2.7 billion views. What’s even crazier is the
rate at which they’ve grown. So if we take a look at their graph over the last, you know,
we’ll call it 15, 16 months they’ve essentially gone from
123 million monthly views to 2.5 billion monthly views. That’s a 2,000% gain
in a span of 15 months. That is a tremendous
amount of viewership gain. Now if we go back a little bit, we see right here in July and August viewership kind of plateaued a bit, right, they got stuck at about 1.8 billion and then, oddly enough, in December they dropped down a little bit. Now, January they’ve exploded, right. They’ve gone from 1.47 to
2.5 billion monthly views. They’ve added 700 million
monthly views in that time period which is just a massive,
massive amount of viewership. So what happened in
this time period, right? Well, we know two things
happened in July, August, and September and then again in December. So, first and foremost, there was a browse features algorithm shift that kind of started right in August and what they all the sudden were doing was featuring a lot more
content that was older, right. These were videos that were
four weeks, five weeks, four months, five months, five years old in browse features. And this was a time in which
there were some changes in the YouTube Kids app as well. We have very limited insight
into the YouTube Kids app so we don’t really know
what happened there but the browse features
changed for everyone. Now, they switched it back, right, because they realized new
videos weren’t getting a chance to shine and I’m
sure there were plenty of media companies and
creators that were like, “Hey, what the hell
are you doing YouTube?” And, of course, YouTube denies all of this but it is undoubtedly
true that this happened. We proved it with data, we
proved it across verticals, we proved it with many different channels. There’s no real debate there. And so, probably in this time Cocomelon had this same thing happen
to their new uploads. Then from September to about mid-December there was another shift and this shift was in suggested videos. And what we found with the
shift in suggested videos was that channels that were
high-frequency posters, right, we’re talkin’ about T-Series, we’re talkin’ about SET India, WWE, a whole host of channels including kid’s channels all the sudden saw all their viewership
dropping dramatically. Now, that’s as far as we can go from like a proof point universally. We have several theories
as to why this happened. Kind of first and foremost was that there was a greater importance
put on the click through rate of new videos in suggested videos. There’s a lot of evidence
that points to this both from analytics that we have access to but also more broadly that we know that high-frequency channels often have lower click through
rates on newer videos and this happens for a lot of reasons. Kind of first and foremost, when you’re a high-frequency poster,
and we’re talking about more than three times a week,
essentially what happens is your uploads begin
to step on each other and the new upload that comes out gets preference over the one
that came just before it. So if you’re uploading
on daily basis, right, your new upload is going
to get more impressions than your previous upload
which is going to stifle its view velocity. On top of that, if you
have multiple new uploads, right, multiple uploads within
a given 48 to 72 hour period, those videos are likely to
get thumbnail impressions very close to each other and
what ends up happening then is you get these kind of
empty calorie impressions where one of your viewers is forced to make a decision
between one of your videos in which to click on. So even if you do get a
click, which is fantastic, other videos still get that
impression without a click and on top of that those
videos are more likely to be recommended to the same person over and over and over
again because they are new driving down their click through rate. In addition, high-frequency
channels tend to post on a wide variety of topics, right. They’re not trying to hit
a particular audience, what they’re trying to do is just saturate the platform and take
up as many impressions as they possibly can by
having this massive library. What ends up happening is you end up serving no audience directly
or from a super serve, you’re just kind of serving as many people as humanly possible. So your click through
rate, again, goes down because you’re not centrally focused on one particular audience. So the suggested video algorithm shifted to being focused on click through rate as having a greater weighting essentially. Well then high volume
channels are less likely to get more suggested video impressions thus driving down their viewership. ‘Cause we know, by and
large, suggested videos drive the vast majority
of views on YouTube. At the same time, we saw
lower-frequency channels really actually grow, right. MrBeast is a great example of this. Cocomelon is a great example of this. There are several other examples, mainly in the kid space because we have a limited data set of the top 500 channels that when they were a low-frequency, we’re talking about three
or less uploads a week, their viewership was growing. Lower-frequency channels tend to have higher click through rates because they don’t
suffer from the drawbacks that big channels do. They’re able to focus on
one very specific audience. They don’t have to upload tons of videos on tons of different topics and therefore they’re able to super serve a single audience which
is more than likely to click on that and their
uploads are less likely to be stepping on top of each other and causing kind of these
duplicate impressions issues. But the second thing that
may have been changed in suggested videos was
a heavier concentration of newer uploads being
put into suggested videos. Now, this may have been
an extension of the tweak in browse features where browse went from heavily library focused
to heavily new content focused or at least kind of swung
back to where it was before. Maybe that same kind of
shift happened in suggested. And some of the evidence
that points to this is that because high-frequency channels tend to get most of their
viewership from their library. If they’re getting more than
10, 13, 15% of their viewership from new content, I would say that’s rare. However, lower-frequency
channels tend to get more of their viewership
from their newer content because they’re getting
high click through rates, they’re being shared more
broadly when they come out. They have a higher ability,
or at least chance, to be featured in many other
channel’s suggested videos. A third potential reason,
and one that I absolutely do not buy whatsoever, is seasonality. Never in the history of YouTube has viewership dropped
from September to December. We have ten years, and now, I know history is not necessarily always
going to predict the future but I highly, highly doubt YouTube shed double digit
percentages of viewership between September and December
without a major freak out. So we can pretty much rule out seasonality as that has never been the case in the history of the platform and makes zero sense from
an anecdotal standpoint. So here’s T-Series and
their viewership, right, had a great peak right here in August and then in September they fell. They fell by 400 million monthly views. That’s greater than 10%. By the time November came they were down to 2.4 billion monthly views. That’s a drop of 700 million monthly views in the span of three months
and a percentage drop of over, they had a drop of over 20%. Now, they rebounded up in here and this relates to the
kind of the shift back a little bit in suggested. If we look at, and they are
a very high volume channel. Right, we’ve got another high volume channel here, SET India. Let’s look at SET India. Now, they’ve got a little
bit of an error here. Looks like they probably had
a bunch of videos privated or something like that. But August 1.56, 1.52 in
September, and 1.38 in November. Right, so, not as drastic but that’s still 200 billion view, or
200 million view loss. Alright, so this is a
low-frequency channel, right, had a pretty good summer. Here’s August, they got
hit by some of the issues around the browse features change ’cause I think they’re a newer channel and I think whenever
there’s gonna be a change in browse you’re gonna
have some turbulence right. But they hit around
701, right, and they got a little low in October but look at this, you know, up through December
significant growth right. 200 million monthly views,
that’s nearly 33% growth right, in the span of three months,
two months, sixty days. That’s a massive amount of viewership for a low-frequency channel, right. This channel only has I
think 300 total videos on it. They don’t upload all that much. Let’s look at another high-frequency. Let’s look at Zee TV. 58,000 uploads, right, monthly viewership goin’ up, up, up, 2.1
billion views in August. By October 1.06 billion views. So they lost 50% of their
viewership in 60 days, right. Down to below a billion in November. Kept falling in December. Right now they bounced back up, gained about 100 million monthly views. That’s about 11%, that’s
not insignificant, right. That’s a pretty nice jump back up. WWE, it’s another high-frequency channel. Holy hell, right. Ow, look at that 1.2 billion monthly views everything’s lookin’ good. That’s a pretty steady
upward trend in August but then September comes and they lose 250 million monthly views, 20% of their viewership. October 871, November 748,
750 million monthly views. By December this high-frequency channel is down to 550 million monthly
views from 1.2 billion. That’s well over 50% of their
viewership lost in 120 days. So one of the reasons
for Cocomelon’s growth is they happen to have
a programming strategy that allowed them to flourish when the algorithm was
shifted up a little bit. That said, let’s take a
look at what they’re doing and see what we can learn from that and apply to our channels. So, the first thing that
jumps out to me with Cocomelon is how infrequently they post content. A lot of kid’s channels blew
up by posting a ton of content. One, two uploads every day. And Cocomelon didn’t take that tact. What they’re doing is
two uploads every week. And they’re very, very
consistent with that. Given their recent uploads it looks like they upload every Tuesday and Friday which is actually a really good cadence for a two day a week, especially
in this particular vertical and is really good in
pretty much everything other than very specific verticals like education or very specific audiences like an audience that’s 24 and older. If your audience is younger than 24, you tend to get most of your
viewership on the weekends. Older than 24 tends to be
mostly during the work week. By spacing out they
insure that their uploads aren’t stepping on each other. On top of that uploading on Friday insures that video’s gonna be uploaded on the days when the most people are on the platform and has the most view velocity going into Saturday which
we know in the kid space is probably the biggest day
for viewership on the platform. An upload going on Tuesday, that’s probably a down day. However, by uploading on that Tuesday you are going to hit the subscribers and the viewers that are on the platform on that day and give
yourself enough spacing where that video has a chance to breathe. It increases your viewership. That viewership goes across all of your other library videos, kind of sets the stage for
a great upload on Friday. The second thing that jumps out at me about their programming strategy is that it’s actually
a little bit different than what we normally advise
and it’s kind of fascinating. They upload videos with
very different time ranges. So, they’ll upload a video
that’s four minutes long and a compilation that’s 35 minutes long. Let’s look at the four
minute long videos first. So many of their four minute long videos are racking up millions
and millions of views. In fact, more so than
their compilations, right. This video here, 11 million
views in the span of four days, three minutes twenty six seconds. Compilation, 35 minutes, 7 million views. Three minutes eleven
seconds, 19 million views. 36 minutes, 7.1 million views. Three minutes, 18 million views. In fact, it’s pretty safe to say that most of their shorter videos are actually doing better
than their compilations. Now, we would have to do
some serious data research to know exactly what the averages are but, you know, it does appear
that the shorter content is doing pretty well. Having different video
lengths on your channel is a really interesting,
and what I would say is a fairly new approach to programming. Now, by fairly new I just
mean that in the sense of kind of what the
general best practice is. The general best practice
is to post videos that get between six and eight minutes of average view duration. We’re seeing more and more that rule is kind of
falling apart a little bit and what that leads me to is that YouTube is doing a better and better job of programming to the individual. Individuals are more or less likely to click on a longer video. For example, in general what we see is that the longer a video is the lower the click through
rate is going to be. Essentially, there are
fewer and fewer people as videos get longer and longer and longer that would just be willing
to invest that amount of time into a particular video and this is gonna change across audiences and across verticals but,
again, generally true. So, in that same rational, shorter videos are likely to have higher
click through rates. And with what we just discussed in terms of suggested potentially
being more heavily weighted towards click through
rate, this kind of strategy might make a lot of sense where you 35 minute long
videos might be getting that really solid average view duration but at a lower click through rate versus your shorter videos
that obviously won’t have nearly as long of
an average view duration but might have a higher
click through rate. This might also point to something that there’s not a lot of data or evidence for right now, which is
a increasing weighting of percentage viewed. YouTube’s big talking point
is follow the audience that they’re now optimizing
not for watch time but for viewer satisfaction. Well, how do you measure
viewer satisfaction on a platform that doesn’t
have a “I am satisfied by this video” feedback function. Now, you could say likes and dislikes, well, in the kid space
that’s incredibly difficult mainly because most kids don’t have, like, accounts accounts and don’t interact or engage with content and in the Kids app they literally can’t. So, my theory, but again, there’s limited data supporting this, is that YouTube is going to be putting a greater, greater weighting potentially onto percentage viewed
or, at the very least, in certain instances,
look at percentage viewed as an indicator of
satisfaction with the video because, think about it,
why do you click on a video? You click on a video
either to get information or to be entertained, basically. And so, if you’re a viewer
who is watching a video to be entertained and it’s
a 35 minute long video and you only watched 10% of that well, you were only entertained for about three minutes and
thirty seconds of that video and then you click away. Well, that might not be a very good video to satisfy people that
are potentially looking for 35 minutes of entertainment. That keep people watching for an extended percentage of time. Regardless, this shows that
it is absolutely possible for channels to be wildly successful in uploading content
that is under 10 minutes. My thinking would be that this largely just points to high click through rate. Any channel that is the
second most viewed channel on all of YouTube, generally speaking, is probably going to have
a high click through rate and I think it’s very safe to say that if you only upload two videos a week that you have a high click through rate if you’re the second most viewed channel. And we know that click through rate is largely influenced by thumbnails. Let’s take a look at
some of their thumbnails. Alright, so, what is very, very clear with Cocomelon thumbnails is that they use extremely bright colors. Now, the very first
color that pops out to me in just about every thumbnail
is the color yellow, right. Yellow is in their logo, sure, but look at all the videos that also contain other
yellow elements, right. Hair there, and these buttons here, the background, the shirt, and I believe this bunny ear here, this
square here is yellow. There’s yellow element here,
I’m not sure what that is but up here, yellow, yellow,
yellow, yellow, yellow, yellow right, yellow, yellow, yellow, right, yellow, yellow, yellow,
orange that’s pretty close, we’ll count it, right, again yellow hair, wood, very light, mostly
yellow, pants yellow, right. Yellow shape right here,
yellow, yellow, yellow, yellow, yellow, yellow, right. 89 million views, their
most successful video over the last two months besides
this one here, fair enough, big giant yellow school bus. We got some yellow in here, right, and you could argue that this kind of like beige-ish background is yellow. But yellow is all over their thumbnails and we know for a fact that
yellow stands out to humans and the reason why is
that we have three cones in our eyes that perceive light. We have a blue cone, a
green cone, and red cone. Well, when yellow light comes at our eyes the green and red cones are activated at a very high wavelength. So it’s hitting two thirds of our cones and it’s doing it at a high wavelength. This makes it stand out
to us in a pretty big way. On top of that, think
about it, school buses, warning signs, other things
that want to draw attention use this bright yellow color. So when all else fails, use yellow. Next in their thumbnails
is the close up on the face and often times it’s
difficult for cartoons to use the YouTube face but they’re kind of main baby character does it in almost every single thumbnail. Let’s take a look, right. So here, open mouth,
whites of the eyes, right, making eye contact with another character. Same thing here, open mouths. In fact, they all have open mouths, they’re all talking at the same time or maybe they’re crying ’cause
they’re all babies, right. But whites of the eyes,
whites of the eyes, whites of the eyes,
whites of the eyes, right. Whites of the eyes, strong
expression, open mouth, strong emotional expression, almost making eye contact with the viewer but a close up on the face. This one’s not fantastic
but also look at it, right, it’s not fantastic
for their channel in terms of viewership. Maybe it’s because this thumbnail isn’t quite as optimized
as some of their other ones from a click through rate perspective. Lots of close ups on faces,
lots of strong emotions, they don’t really make eye
contact with the viewer but that’s a little bit less important especially for this demographic. The one thing I would say
is that their thumbnails tend to be very, very busy which in the kid’s space
doesn’t matter as much. Like, they’re highly
optimized for the Kids app. And the Kids app is often
using very big thumbnails. There’s not a lot of small thumbnail usage in the Kids app and so if
they’re getting more and more and greater and greater click through rate within the Kids app that means they’re gonna be featured
more and more in the Kids app and the one thing that we’ve seen kind of across the board is that
YouTube uses the Kids app to program the YouTube site at large in a very significant way. Their viewership tends to
track very closely together and so if you’re a kid’s channel and your thumbnails are
a little bit cleaner, a little bit more structured
and easy to understand, you might want to try going with these type of thumbnails that
are a little bit more action. So there’s stuff going on
in these thumbnails, right. And if you look at a lot of
the kind of the thumbnails in the kid’s space that
historically have done well, they’re very static,
they’re very still, right. It’s like a close up on
a single image, right. And it’s interesting, if
you look at Ryan Toys Review a lot of his thumbnails
insinuate action as well and so the kids or their
parents might be more drawn to these kind of action based thumbnails than some of the more static stuff that we see used to be popular. The third thing that we
see in every single one of their thumbnails is
the same character, right. The same character over and
over and over again, right. It’s a baby. Now, if you have a one to
four or five, six year old, like I do, right, they’re
fascinated by babies most of the time. They’re gonna feel strongly
one way or another about it and they understand
that and so by focusing on that baby character,
Cocomelon might be tapping into something that is
very inherent to kids which is being drawn to babies. You know, like, what they see every day. Especially like most people
have kids kind of close in age to each other, right,
somewhere between, you know, one, two, three, four, five years apart and so they’re very familiar with babies and they might have a baby in their life and click on it because it’s
something that’s recognizable, something that’s identifiable to them. But, more importantly,
it’s the same character over and over and over again. Again, if we look at other
kind of kid’s channels in the space, especially
ones that focus on things like nursery rhymes and that sort of thing we see they have a lot
of different characters in their thumbnails, right. They’re kind of putting
a lot of stuff out there whereas Cocomelon said, “No, this works, we’re doubling down on this.” And every single thumbnail
features that baby. The fourth thing I’ve noticed
about Cocomelon’s thumbnails is there’s zero text, right. There aren’t even those like little blurbs that you sometimes see in the kid’s space that are like, “36 Minutes.” Right, and I know some
of that might be geared towards parents and trying
to get the parents to click, but I think we’re seeing more and more of the kids driving the action. Especially as YouTube
gets better and better about kind of managing the Kids app. Parents are feeling safer and safer about allowing their
kids to use the Kids app and make those choices for themselves. I believe they announced something maybe a little while
ago that they’re hiring 10,000 people to like review content and make sure it’s safe
and that every video that appears in the Kids app is a video that is safe for kids to watch. So, in that regard I think
we’re seeing kids drive the viewing behavior far more
than adults are right now. Also in thumbnails, let’s
take a look at their titles. So when we look at their titles, right, it’s always leading with a one
specific value proposition. A specific song or specific type of video that is going to lead that video whether it is a short video or whether it is a compilation video. So if we look at it, right, they have “This is the Way and
More” or “Hello Song,” “The Shape Song and More,”
“First Day of School and More.” Right, now, they are doing something a little bit interesting with their titles which is adding nursery
rhymes and kid songs and then their channel name. So for the last two years or so I’ve seen a lot of people on the fence as to whether or not to
include your channel name in the title and I think about it kind of in the same way that
I think about tags, right. Like, if YouTube doesn’t
care about our meta data, right, our title, our
description, our tags, well, then, it doesn’t matter. If they do care about
it then we should abide by the best practices
that we either observe other popular channels doing or that, from our own testing, we know works. And one of the kind of main
talking points around that is whether or not to include
the channel name in the title. Well, if the second largest channel in terms of monthly viewership
is including it in the title then I would say it’s
probably a good bet to do it because if it doesn’t matter,
then it doesn’t matter. If it does matter, well, then YouTube is going to key on that channel name and say, “Wow, all these videos, well, they all have this
channel name in the title. It’s a topic. Kids must care about it. Let’s suggest more of
these videos to that kid.” Same thing with your tags. Let’s look at their tags. Right, so if we look at their tags theoretically if VidIQ is right, this is also the order of their tags. And so we have Cocomelon, ABC Kid TV which was the name before
it became Cocomelon, nursery rhymes, children’s
songs, baby songs, kid’s songs, kindergarten songs, toddler songs, kid’s song, education, children learning, sing a
long, sing a long songs, kid’s songs, and on and on and on. You know what’s not there? The Shape Songs. So what is Cocomelon optimizing for in this particular video? Not showing up for The Shape Song, it’s showing up for all of
these kind of bigger concepts, these broader concepts. Now, again, if tags don’t matter in terms of what a video is about
and YouTube understanding what it’s about then it doesn’t matter. If it does, well, then,
they’ve done a very good job of optimizing this video
that has a tremendous amount of watch time, probably a
pretty good click through rate for people searching for those things or for other videos
labeled that thing, right. I guarantee you, almost all of those have a greater search
volume or at least audiences with greater interest
in watching more content about all of those terms,
than The Shape Song. If we look at their
descriptions, it’s interesting. They’re doing some things that are very classic, best practices. First and foremost, they lead
with a subscription link. Generally speaking
subscribers don’t matter, people say it’s a vanity metric, but someone subscribing on a channel is still a signal to YouTube that in their last session,
assuming this is their next one, it still sends a signal to YouTube saying, “I want to see more from this creator.” Now it might not be weighted super high but it does increase the chances of it. Next we see here in the description that they’re putting
in specific time stamps to each one of these songs, right. That’s intriguing from an average
view duration perspective. Next they’re linking to a playlist and then a couple of very
specific videos, right. And then finally, all the way down here, which probably less than .1%
are actually making it to there they’re putting these
links to their Facebook and their Twitter, right, which generally speaking you
don’t want to send people off the YouTube platform. That should be broadly
understood at this point. It creates an end to a session, right. YouTube generally demotes your videos. Having them buried here in the description isn’t gonna hurt anyone. It’s still good if someone’s looking for more information on you, right. If you’re a parent and you’re going, “Well, I don’t know about
this Cocomelon thing. I want to find out more about it.” These would be some of
the places you might look. Now, down here at the bottom
is a little interesting, right. So, they write out
nursery rhymes in English and a ton of different languages. Now, whether or not this is
affecting their viewership whatsoever could be a matter for debate. If descriptions don’t matter,
again, doesn’t matter. If descriptions do matter, well, then, they could be picking up
traffic either from search or from suggested for
videos that are like this. So what does it hurt? So, the last thing I want
to look at for Cocomelon is what happened to their
channel many, many months ago. So if we look at Cocomelon’s stats, right, and changes in stats always fascinate me because it indicates a
change on the platform or a change in your programming and if you make that change
and viewership goes up, that’s a great place to look for ideas on what to do from a
programming standpoint to increase viewership. So, there’s a very clear time
when Cocomelon jumped up. Right, now they were ABC
Kids TV for a very long time and then they changed
their name to Cocomelon. I’m not sure exactly
when this change happened and I don’t think that’s
necessarily the cause of the change here. If anything it potentially
could have hurt them in the algorithm. But looking at their traffic, right, they were a pretty good channel, right? Like you go back into 2016, right, they’re doing 40 million,
50 million monthly views this steady kind of increase but then something happens. June 2017, 54. July 2017, 72. August 2017, 70. September, 90. October, 123. And it just starts to explode. So they went from, in July of 2017, right, or here, we’ll start
here, in August of 2017, 70 million monthly views. By January, right, they were
at 400 million monthly views. That is a massive increase, right? That’s like a 500% jump. So March 2017 for the very first time, a video was kind of their new animation and new characters appeared, right. It did 4.3 million views, right. Or maybe this is it back here, right. This one, this is the two minute version, this is the hour long version. This thumbnail’s definitely better, right? But this is basically where
it started to happen for them. Right, and look at this character, right? The characters changed too. But regardless, right, we see a shift. Let’s look at their programming right. Programming and thumbnails. So, you got a bunch of animal characters, bunch of random characters, it looks like some of this is in 2-D animation versus 3-D. Let’s look, oh yeah, hardcore 2-D. Look at this, my goodness. Right, so, first of all 2-D to 3-D right. If you’re in the kid space and you’re not doing 3-D
right now on YouTube, like, re-evaluate your priorities. But they start bringing
in some 3-D, right? And then they find this
character, the baby, right. First appearance, second
appearance, right? Oh, well, that’s workin’. Third appearance, right? And we start getting a
little bit better, right, art design I think. But regardless, they
start showin’ up right? Now, we go back to some
animals in here, right. Maybe it did okay but,
again, we got the baby, poppin’ up the baby, poppin’ up, right. Baby, baby, baby, baby, right? We go back, right, now some
of these do really well. 166, that’s pretty phenomenal, right. But we’re fully in 3-D now, right. There’s no more 2-D animation
happening on this channel and we’re right about the point of August. So we look at this thumbnail, right, this isn’t actually a
fairly good thumbnail so we gotta imagine the
average view duration in here is pretty good, right. And let’s look at some of the programming that was kind of around it, right. Five and a half minutes,
six and a half minutes, seven minutes, three and a half minutes, eight minutes, three and a half minutes, seven and a half minutes,
right, hour long, right. Well, if these videos
are doin’ pretty good, and this was doin’ 52, 36,
right, 65, 42, 80, right, now some of this might be influenced by this video taking off,
but “No, No Bedtime Song” at seven and a half
minutes might just have a wonderful combination of
CTR and average view duration causing it to really explode. Or maybe other channels that were doing a bath time song, this
video for some reason was put into suggested
and it was the way to go. But this video changed the
course of their channel forever. I don’t know if there’s anything
that we can particularly take specifically from this, but look at their meta data. It’s really clean, it’s
really general, right. The thumbnail is still showing action. This is one of their first kind of like big action thumbnails, right. All this other stuff, this
is a little static, right? This is a little static,
this is very static, static, this one’s got some action in it, static, static, right, and all 2-D right? Lots of static stuff. This is kind of a static image, right? Maybe it’s implying some action but I would say this one
has a lot more action to it, even though it has less views. But if we look at this right? This girl is clearly,
right, trying to engage with the baby and the baby’s
clearly responding to it. Here they look like they’re posing. Look like they’re posing. This one’s eh, but kind of
look like they’re posing. Right, this one’s a little bit static but kind of feels like action. This one feels like posing, right? So think the shift in programming allowed their channel to take off and then they kind of began
to experiment with things. Again, right, because if
we look in here, right, they were kind of like
“Well, do we do individuals? Do we do a couple together? Do we loop it?” Right, and they’re at eight and five and then they pop out a one hour one and that one hour one takes off, right? At the time this video
probably out-performs most of the other videos
coming out, right? And they were like, “Well, wait a minute, we just stopped doing
one hour compilations. What’s the deal?” They did a five and a
one hour and a one hour and then an eight and
then a one hour, right? Now a thirty minute, then a eight minute, then a one hour, right? And then a five, a four,
a forty one, right? Four, four, seventeen, four,
forty two, three, right. Forty, two and a half,
forty, two and a half. Thirty three, two and
a half, twenty minute, eighteen minute, four
minutes, thirty five, three, two, thirty five, right. Two, thirty eight, four,
thirty nine, three, thirty nine, and then we find our pattern. That’s when we start seein’ it. You’re at an individual video and a forty minute
compilation twice a week. So you go through this
kind of period of change and with this change probably
their stats went nuts. And they were trying to figure
out, what is the pattern that we need to get into for our audience that is going to allow YouTube to grow our audience most efficiently. And I think they’ve found it. I think they’ve found it with two a week, one’s a single, one’s a compilation. The single is gonna be somewhere between kind of like two and
four, maybe five minutes on the long end and the
compilation’s gonna be somewhere between thirty
and forty minutes, right? And they’re not beholden
to one specific time where it has to be that way. Now, that said, generally speaking, I still think it’s probably best and who am I to contradict the number two most
viewed channel on YouTube, but I think it’s still probably best for kind of your normal channel that’s not doing 2.8 billion
or however many views a month to find the optimal length
for one of your videos because what that then allows you to do is to observe changes and
understand those changes with fewer variables
versus they might not know if it’s click through rate
or average view duration that’s driving a particular video and it might make it a little bit harder for them to make future choices
about their programming. Alright, so, that’s a little
behind the scenes look at how we might analyze Cocomelon and take a look at what was going on on the channel that kind
of led to the growth and what’s facilitating
some of the growth now. Thank you guys so much for watching. If there’s a channel you
want me to take a look at or a particular vertical you
want me to take a look at, write it down in the comments. Otherwise, please take a
look at one of these videos. Don’t forget to subscribe and
I will see you guys next week.

Tags: , , , , , , , , , , , , , , , , , , , , , , , ,


  • FunForKidsTV - Nursery Rhymes & Baby Songs says:

    Great info

  • Rosh Sillars says:

    I'm always impressed by your analysis. Thank you

  • Juan Felipe Campos says:

    Oooof. He did it again. Another awesome breakdown.

  • The Infographics Show says:

    Fantastic stuff. Great analysis!

  • Top 5 Best says:

    Simply incredible video!

  • happyfamily1004 says:

    Excellent Video!

  • LET UP! says:

    accurate analysis, but the big deal with this channel is:
    1- different length uploads, they made the audience retention works for them, by confusing the algorithm recommendation progress
    2- timestamp & viral titles and links of their viral videos, they made a chain of recommendation into suggested videos traffic

  • Skip the Tutorial says:

    Your points on the thumbnails here really stuck out to me – good stuff! Have you folks ever considered looking into David Dobrik and his whole ensemble and how they grew so massively? Or, for a different angle, the changes that Ross Bollinger took with Pencilmation to see so much success after the channel’s time on the platform would be a fascinating deep dive.

    Keep up the solid work! Looking forward to learning more

  • Evan Carmichael says:

    Pure fire value as always! Thanks Matt 🙂 #Believe

  • Jens Larsen says:

    Fantastic videos! Thank you, Matt!

  • Tom Nash says:

    Hey Matt this is brilliant. Well done.

  • Dear Gamer says:

    This was fantastic, I saw you live at Vidcon two years ago. Keep up the great work man!

  • JeffNotes says:

    There's an abundant youtube growth knowledge here, I am immediately subscribed! I hope you work on strategies for success in animation channels later on! It's something specific for my own channel that I'd like to learn more from =)

  • Kate Walker says:

    It's not just "guys" working on YouTube channels you know 😉

  • Nathan Hummel says:

    Loved the video Matt! Excellent analysis on Cocomelon! Any chance you could look at a channel in the tech review space, such as, Unbox Therapy? I am interested in hearing your thoughts on his channel.

  • Michelle Osorio's Dare To Dream School says:

    You are one of my favorite new channels. Subscribed with notifications. So glad y'all are making videos now.

  • Financial Freedom says:

    I would love to see Dan Lok channel

  • RuiN Reserve says:

    Great video 👌

  • Kiddopedia says:

    This is a great analysis, Matt. I have been reading your blog posts and I listened to you several times on podcasts. I am very happy to see that you are on YouTube now.

    There are excellent tips in this video like usage of color yellow and other elements on thumbnails, frequency of posting, etc. I create educational content for kids where I usually post 2-3 video per week. I have been planning to increase the number of videos that I publish. I am reconsidering it now, it seems that keeping the number and increasing video quality/length is a better plan.

  • Andrew Parsons says:

    Thank you for captioning the video!

  • Хапче В Душата says:

    So healthy stuff! Thanks

  • Хапче В Душата says:

    I understand that sending people off the platform to FB or other social media might demote your video. However the point and purpose of growing your audience in all social media, is so that at the end of the day, you can share youtube videos with your audience there, so they can help each other and bring more people back on YouTube. Like those audiences that are more active on Facebook than YouTube, but because they like your content they come back or share your content.

  • Binki Kids - Nursery Rhymes says:

    Great Info
    Also one thing to consider about Cocomelon is the channel age
    just like super simple songs these channels are more than 10 years old

  • KELimagination says:

    Great! This is relevant to me. Thank you!

Leave a Reply

Your email address will not be published. Required fields are marked *