- Greetings, people of the Internet (chuckles).
Derral Eves here, now there's one constant at YouTube that
always happens and it's change.
Now I know a lot of content creators, friends,
and subscribers have been really reaching out to me
over the last quite a few months
'cause they're really worried about the algorithm
and the changes that occurred, now just recently,
I was actually talking with the search and discovery team
at YouTube, these are the ones that their team
actually makes the algorithm (chuckles)
actually puts it together and I was actually talking to them
about how many changes occurred on YouTube
and get this, last year, there was over 200 updates
to YouTube, that's right (chuckles) 200 updates.
Now here's the crazy thing, YouTube likes to experiment.
They actually put these little experiments out there
that affects groups of channels or videos and so on
and so forth and there was 10 times that amount last year.
Now that's a lot of things, now what YouTube does
with these experiments is try to figure out
where they're actually going to tweak the algorithm next
and they just kinda broaden their experiment
a little bit further, and if it's positive results,
then it basically will change the algorithm.
Now there's a lot of changes going on.
A lot of people reached out to me really concerned
where the algorithm is and so on 'cause YouTube's
not really transparent what the algorithm is.
They're trying to be better, which is really good,
but ultimately, there's a lot of questions you might have
about the algorithm, now, I do a couple things
to really help people in this realm.
Number one, I love education, I wanna help share
this information with everyone 'cause that's one
of my core mission here in life.
And I put on a conference every year.
It's called the VidSummit and the VidSummit
is something that's a little bit different
than any other conference out there
because we really just focus 100% on the business side
of the industry, building an audience, making more money,
finding ways to diversity the income, legal issues,
whatever it may, working with brands and businesses
and so on, and we really focus
in on that, so if you're interested about this,
it's coming up in October and you can get there
in L.A. October 10th through the 13th and you go
to VidSummit.com for more information.
Now here's the great thing.
In this video that I'm actually presenting today
on the algorithm, there's different places
in there that will show you where to actually get
replays for free of the last VidSummit.
So this is a pretty big deal, it's like 32 hours
of the best and brightest minds
in the industry coming together and presenting
that information to you.
Now, this presentation that I'm about to give
is something I put a lot of time and effort
in collecting a lot of data and this data,
I wanted to be able to communicate that properly
to my subscribers and also the individuals
that didn't have the chance to see my presentation
at VidCon, so I was asked to speak at VidCon.
We filled every seat, we had people sitting on the floor.
I was glad there was no fire marshal that was there
(chuckles) but we literally packed the house
and they locked the doors and there was literally hundreds
of people that weren't able to see my presentation.
Luckily, I brought my camera because I'm a YouTuber
(laughs) and we were able to record it
so thank everyone at Patreon for making this possible.
Those that are in our community, and also the people
at VidCon, so you get one of my two presentations
I give here, and it's about to start right now.
(metallic chiming music)
(zapping)
- [Announcer] DerralEves.com.
- Today, we're all gonna talk about triggering
algorithm-driven views and I wanna do a little disclaimer
because it'd just be right of me to do this disclaimer
but I'm giving you information on the data that I have
access to, which is quite a lot so I personally own
17 YouTube channels that we upload
23 videos a day on, then we manage
over 52 different other channels
that get a lot of views and so we get
a lot of different data.
We're always starting new channels and so on
and yesterday, I was meeting with the search
and discovery team at YouTube,
which the most brilliant people that work at YouTube.
A lot of the changes that happened right now
and that are currently happening, there's a lot
of experiments that are going on.
So last year there was 200 changes to YouTube
and I think we felt that a little bit
and I think Matt Gielen put it the best, he's like,
they didn't communicate a lot of that with us
and lot of people hurt, so that's why I'm here today
is to share, with the information that I received,
the knowledge for everyone here.
The second thing is there are, so there's 200 changes
last year but there was also 10x that in experiments
and YouTube does experiments before they implement
any of the changes to gather data and sometimes
it hits your channel for a while or videos
and it does certain things and it's really awkward at times
and you don't even know you're a part of it
so a lot of that's happening and it's 10x the 200 changes
and that's always, always happening.
The most notable changes to YouTube was in 2012.
A lot of them were these experiments that happened in 2012
and then there's a fundamental shift happening on YouTube
and that fundamental shift that literally transitioned
was going into what we know as the algorithm today.
Not much has changed about the algorithm
but the algorithm is AI, it learns, it understands,
and it expands, and so in 2013,
we pretty much had a lot going on.
We had even the glorious Google+ disaster (chuckles)
and there's a lot going on, what's interesting
is the core of the algorithm,
where YouTube actually puts the most weight,
if you did not catch the session that Matt Gielen
talked about today, he had an article
go up on Tubefilter that you need to check out.
I think it was Tubefilter or Tubular, one of those,
that went up today that you should read,
to have a little bit better understanding
of what's actually going on from the data set
that he has as well, but the metrics
that they actually are looking at, what they're
really focusing in on are video minutes watched,
which isn't a surprise, if you go
in your analytics dashboard and you see what are the areas
that they look at, there's six different areas
that they really have the focus on.
That's where you need to take particular attention.
Now in 2013, not a lot of people realize this
but they changed the view structure on it
to watch time minutes and so we knew that there's
a little bit more priority on that and we were able
to focus in on seeing how they'd actually changed that.
This one right here is one that most people don't talk about
but it's like the channel has authority
and it can have authority and it can post a video
that can get really out there and really suggested
and you can gain a lot of authority
based on your minutes watched for your channel,
whether it's today, this week, this month, this quarter,
this year or lifetime, and there's a lot of things
that will trigger events that will occur
to really get your videos be found
through the suggestion engine which we're talking about.
Also view duration and they wanted to make sure
that hey, there's a lot of bots out there that are people
are buying YouTube views that are just going out there
and triggering these views, we're now
gonna just kinda delete all those, which they did
in May 2013 and everyone was freaking out
that they were losing views but yet, they were getting rid
of the fake bots that were going on.
And they were really looking at other metrics
like video duration and session duration,
which is a session is where you actually bring someone
onto YouTube, that's when the session starts
and that's a good thing and we'll talk about some bad things
about that as well, moving on to another thing
is they really viewed content as the upload frequency
and if you notice since 2013, a lot of daily vloggers
actually received a lot of views and were getting promoted
quite a bit and you're like, what the world?
Why is that happening to me?
I have better content than them, why is that happening?
Well, one of the metrics that looks at is how often
are you actually updating YouTube to create
a new digital ID, now there's only two ways to do
digital IDs on YouTube, number one is to upload a video.
Number two is to create a playlist.
We'll get to that in a second, it's really important.
Then they also have a session Ends Penalty,
which deals with if you're sending traffic off the platform
then there's kind of a slap on the wrist a little bit.
It's not as intense but this is something
that they were able to track, that that time is there.
Now I wanna see a show of hands first off
is how many of you have YouTube channels, raise your hand.
Okay, wow, wow, okay, how many are you channel managers,
raise your hand, okay, right, perfect.
You're in the right room, channel managers.
Which one is the most traffic source
that you'll actually get on YouTube, does anybody?
- [Man] Suggested videos.
- Suggested videos?
Suggested videos.
It literally drives all of YouTube, the AI,
the AI basically under the hood
of YouTube is it's trying to predict exactly
what the right user, the viewer, wants to see next
and they will continue to tweak that
until they have it perfect and it will never be perfect
because people change over times but it's gonna predict
what would be the best type of content
for them to consume, it's in the best interest for you
and also YouTube to understand this
because that's where it's going and I know
there's some frustrations like, oh, my subscribers
are not being notified and oh, this or this or that
but I want to, like if I could do one thing today
is I would like to say to all of you don't worry
about your subscribers being notified.
What you want to do is have YouTube promote your channel
and things will change and you will see
some triggers that will happen that will really resurrect
channels from death.
So on the upload, a lot of people don't know
what happens on the upload and I wanna
go over some very simplistic things.
I'm not gonna get too technical on this
but when you upload, it creates a digital ID
and with that digital ID, it goes through the content system
to check to see if there's any copyright claims
or anything like that, right?
And as it's processing, 'cause that's what happens
when it's processed, then when you hit public,
basically what it does is it goes out
to some of your subscribers, now it doesn't go out
to all your subscribers but it goes out to some of them
and it goes out in several different ways
and they're able to access it in several different ways,
more ways than they were back in 2013 or even 2012.
Your content can be seen by more people.
And what happens, first off, is when you upload that video
and you start to see that you're getting a lot of traffic
and then it starts to bump up a little bit.
That just went out to your subscribers and when it starts
to really take off, it does this.
It serves it out to related videos on your video
so it's gonna go through your whole library
of your content, see what's actually in titles,
descriptions, meta information and say,
okay, do you know what?
They did a video on this, he's uploading a video on that.
Let's go ahead and make that a related video
and see what people will do.
Next, they are looking for this and I'm here
(chuckles) to tell you, I think Chad Wild Clay
was really talking about this last year sometime
but it's like they really just care about click rate.
They wanna know what the highest click rate is for that
and it's true, it's 100% true and we're gonna talk about it
but it's something that you can say,
oh, well, I'll just put up a great thumbnail
and a great title and we'll have great content,
well, not necessarily.
And then, two, as it looks at the click rate
and the percentages that are there and it looks
at how many people, your subscribers, that are engaging
with that content 'cause it's all about engagement
but not engagement through comments,
not through engagement through shares.
There's a lot of factors that are going on.
It's looking how is this content being received
by the people that are actually seeing it?
When it's positive, you're gonna get more visibility.
Now the visibility comes through extended reach,
through other suggested videos by other creators.
It's gonna go through more of your videos
and then also, it's gonna be more prominent
for your subscribers, even if they haven't engaged
in a while to see your content, okay?
So if I have a good click-through rate
and I have a good clickable title
and a good clickable thumbnail,
is that gonna give me success?
Well, absolutely not, absolutely not, the reason why
is because there's a factor that happens
and a metric that happens that there's certain amount
of time that occurs or certain amount of instances
that occur in the suggestions and seeing
the click-through rate and then when YouTube has enough data
and your attention's low, like you're not getting
the watch time that you need and people are dropping off
and they're able to get that data, what happens?
Well, it's what I call the Bounce and Drop Effect
and it goes through suggested videos.
Who's seen this before, that your video starts to take off
and it goes really, really high and you're like,
man, this is gonna be the most amazing video.
It's gonna bring me a ton of views, ton of subscribers,
and then it falls completely apart?
And this happens a lot, if you will go into your analytics,
I don't know why that black box is there.
I wasn't covering anything up, it's really not covering
anything up but you're able to see that
and it'll go a little bit of time,
and what the AI at YouTube's looking at, it says,
well, maybe that wasn't the right time to do a lot
of those suggestions so let's try it one more time
but let's do it at a different time
with a different time zone, with a different amount
of subscribers that this person has
or people that would like this content.
So we'd do another little bounce.
It's not gonna be as high as the first
but it will be prominent and this is all coming
through suggested video search and, I'm sorry,
suggested related videos and it'll go do a bounce
and if it's really bad content, I'm sorry,
but it's gonna do into the trash
and it will literally stay there, it will never resurrect.
It doesn't matter what happens.
They could have some other video that takes off.
It might do another little boost but it will always tank
because the other metrics that's looking at that
is really making the big difference.
I wanna talk about decay rates.
Not a lot of people talk about decay rate
but it's something that I find really, really,
really important and understanding decay rate is,
I had a lot of misperceptions about decay rate on YouTube,
what happens when you upload a video,
what happens when you publish,
how much is it being suggested?
And the more data they got, I felt like you know what?
I was actually misleading people and I wanna clarify this
because it's really, really important.
YouTube has its formula of how to actually suggest videos
to your subscribers and to out into different suggested
related videos and there's a point where it stops.
And I had the question, I had the assumption,
do you know what, YouTube, when I upload my next video,
because I would upload it at the same time every day,
that YouTube stopped promoting that one video.
I had the assumption that as soon
as I uploaded the next piece of content,
then YouTube would stop promoting this one
and start promoting this one, okay,
not a very good assumption so what we were able to do
is we have a few channels, five of them
that we actually own, to upload three videos a day
and after doing six months, eight months of different tests,
because I could do tests on these channels
as I owned them, it was really easy to see what could happen
and I found that with these channels,
we tried different times, early morning, mid-morning,
afternoon, we did every variable that you can and we
were looking at not just a U.S.-based centric audience
but worldwide and we tried it in different areas
at different times to see how was able to go
and what we found was this, there was nothing,
you could upload 10 videos a day and it would not take away
from your suggestions as long as the videos are performing.
If it's a bad-performing video, there's some other triggers
that actually happen on that and there is,
man, I just wish that black box would go away
'cause then you could see everything,
but there is a 24 hour decay rate.
It is like clockwork so when you upload a video
and you publish, you literally have 24 hours
of the best suggestions, suggested video placement
on YouTube ever and it will go and as soon as it hits
right at a full day, you'll drop every single time.
It doesn't matter if it's the most viral video of all time,
it will literally drop a portion of the suggestions.
And what it does is it takes in that information
and it goes a little bit of time but you can see
some of these are really, really good videos.
You had like 560,000 views in 48 hours
and you can see right at that 24-hour mark
that it just drops and it will recover.
And we'll talk about some places where it will recover
but you could have really good retention,
your retention could be going up and yet, at 24 hours,
it's gonna drop, everything's gonna drop on that
because there is less value on newer content after 24 hours
and so before, I would always say, hey, the first 48 hours
is the most important, well, no, it's the first 24 hours
that are the most important because it can set in motion
for good videos to really, really amplify
the suggestions that actually occur, okay?
And this has been done over, I could literally
do a presentation for five hours and just click
slide after slide after slide and you would see it.
There's another decay rate that happens.
You notice when you upload a new video and you're searching
in YouTube that there's a little new label that's there?
That goes away at seven days and at seven days,
we've seen the same thing, so you'll actually get momentum
on day one and then it comes down and it starts
to come back up and you start to go out further
but on day seven, that new label goes off,
less suggestions actually occur at that time
and you see a drop.
And some of these jobs are scary
'cause you have really views and this one video
had almost two million views in the seven-day period
but you could see that it went up and then it dropped
and then it dropped even further
and then it just kinda hovered down at the bottom.
Now, I don't want you guys to freak out about this
because YouTube, the platform, is a long-tail strategy.
This will pop again, this video will literally pop
again and I'll explain why and how.
But you can see after the seven-day mark,
I mean you're having a great, great success but then boom.
And all that will come back down to it, if you wanna look
on your own channels, just see, hey, what's happening
in real time and then after a few days,
literally look at what was the main traffic source,
and see the traffic source kind of drop there.
But you can see the labels here.
Oh, we got Team Edge here right at the top
but you can see the little label, they literally disappear.
Now, the next thing, this is probably the biggest change
that affects everyone in this room more so
than the ad apocalypse and it is the creator ads,
or I'm sorry, the creator videos, when you upload a video
in your content, you always have the Next Up
at YouTube would say, hey, do you what?
We'll find the best content for the viewer to watch
but you guys get some presence on your videos
in the suggestions that are there.
And this happened in the fall.
We noticed it actually this time last year
where they started to experiment with it
and we're like wait a minute, where are these things going?
And then I think PewDiePie or someone did
a really big call-out in October-ish but yeah,
it's no longer there so that's a big deal
because right now, even some really good videos,
you can see that there are no more suggestions
for that creator, and I bet there's a lot of you going,
my gut just dropped, oh, but you probably noticed.
You probably noticed the change in your channels
and you're probably wondering, where is my views going?
Well, the bottom fell out of your suggestions
that's going on, now, I do have a client (chuckles)
that is very popular and they have good audience retention.
They own all 20 spots of the 20 suggestions you
can get and it's possible to do it if you understand
how to navigate people through your content
and you're being a little more strategic,
says, oh, I'll just throw up this card.
But you're being more strategic in some of your processes
so you can blame that on mobile because there's
more mobile viewers and the way that YouTube
interacts with content is 100% mobile's fault,
so if you don't like it, just smash your phone and walk out.
But it is a good thing, guys, I'm here to tell you
it is really, really good and I'm hopefully,
the whole purpose of this is to educate you enough
that you can make the right choices
so that you can actually make really solid incomes
on YouTube and not have to worry
about a lot of things but a lot of the stuff
is best practices that creators forget about once they taste
success and they ride the joyride and then it falls out
and they're like, oh wait, I need a,
oh, this is just crazy, but if you do the right things
every time and you have the right strategies,
then it can go from there so this is in my order
of importance of the best things that you can actually do
on YouTube and I really stand by this
because of all the different triggers that occur on YouTube.
The first thing that I would do is spend
a lot more time on the thumbnail and title.
And if they're not gonna click it, they'll never come
into your ecosystem, they'll never come into the system
that triggers everything else and yes, you need to have
clickable titles, you need to have clickable thumbnails
and if I was to spend more time on it,
I would do two or three times more the time
in creating the thumbnails and making sure
it's really, really engaging, we saw one from Team Edge.
I think they have some of the best thumbnails on YouTube.
And I like 'em personally because they tell a story.
They're very engaging and it sucks their demographic
right into their ecosystem.
So this is one of my clients, Studio C but you can see
if you get the right thumbnail that's engaging,
that tells the story and says, well, I wanna see
what they're doing and if you have to put
a little text on it, it's fine,
don't overemphasize text, but spend more time on that.
Now with channels, see, I like to experiment
just as much is Google and YouTube likes to experiment
and I wanna see what would happen if we just fixed
the thumbnails and we just fixed the title.
Instead of creating more content, that we just focus
in on that, this is what happened to a channel.
Look at it, you can see the point where we actually changed
the thumbnail strategy and the title strategy
and you can see the growth after that.
Was it immediate, no.
Did it happen, yes, how long did it happen?
It was literally a month, month and half.
And there's certain things that happen
that got the right factors and we'll talk
about those factors so that YouTube would actually trigger.
The next is the most underutilized tool
or function on YouTube is playlists.
I have met with thousands of different creators and brands
and not one of them has actually yielded
the full potential of playlists and Matt was saying,
hey, now they're in favor with YouTube.
Well, they've always been in favor of YouTube.
They just didn't have as much weight.
They have more weight now and so actively creating
a playlist strategy would be wise.
You can see with the channel that had no playlist strategy,
they had two playlists, all videos and another one
and we created a playlist strategy.
Look at the growth at that, and it's not just videos
in playlists, that's one of those indicators.
That means YouTube's bot is automatically adding videos
from our playlist into other playlists
and it's getting more video views but look at the clicks
on that, I mean it's crazy, so next, tags.
Tags, I've heard so many people, well, you know,
this creator never puts any tags in their video
or they use the same four tags, well, I'm here to tell you
that that creator will hurt at one time.
It might not be today, it might not be tomorrow
but they'll hurt and the reason why is because YouTube
needs data and they only have the data that you provide them
and they have the data they currently have
and they have the data of the viewer
that's actually watching your content and so when
you first upload the video, if you don't put tags in your
videos, then it can't go to related videos
on your channel, it can't go to related videos
on other people's channel
and that alone, there's no character limit,
go up to 500 characters but just make sure that they're
relevant, if they're not relevant, you will hurt.
They're actively going after spam and stuff
so you wanna make sure that they're relevant.
I generally do about 350 to 450 characters for my tags.
This definitely helps with suggestions
in the first five days but it could be only two
or three days, it depends on how much data's coming in
that YouTube knows, okay, we don't need
any more meta information, we can now take it off
of the information that we currently have and serve
that out to where we need it to be.
Next is cards.
You can see our card strategy.
Peoples say, oh, these cards, they suck.
I like annotations better, whatever, but just doing
a simple card strategy, look at how many clicked on cards.
Over a quarter of a million people
actually clicked to go to see something else.
Now think about this for a second.
If YouTube's AI says you know what?
This video's great and there are people clicking
to go to this next video, why don't we suggest that
as the next Up Video, okay, think about that for a second.
And card strategies of when they actually hit
are really good, and I'll tell you an easy strategy
that most of you probably don't think yet.
Your first card could be at 0000 when it comes up.
Your second card should be about 10 seconds
before your audience retention drops,
so your audience retention, if it's three minutes 25 seconds
then you want to literally have it three minutes
is when your first card or your second card hits.
That just entices people and you wanna adjust it
and the biggest mistake is that you always put it there
and you might have higher audience retention or lower
and you wanna just keep on inching it
as you more audience retention on that.
Next, you can't see this but it says end-screen elements
so at the very end having end-screen elements
and we were one of the first people
that was actually in the beta.
There's like five other channels that was on the beta
so we actually have a lot of data on this
and what actually works and we've tested every variable.
I'm here to tell you that this is the combination
that I do 100% of the time, I would not vary it
from this set except for when I really wanna push
another video and I put more emphasis
but I would do it for a small period of time
and then change it back to this set, which is the latest
upload and I would put it on the left-hand side
'cause it's a higher conversion side
for some reason and what's best for viewer.
And then the third option would be your subscribe
so they can actually subscribe.
If you do more than three, you better
justify it 'cause the more options you give,
the less they'll actually click, okay?
And just adding a simple end-screen strategy,
you can see what that actually do for a channel
and this is pretty impressive that you can get
over 1.5 million clicks on something
that you wouldn't have had before and that's going
to your content, going into your systems that are there.
So let's talk about, man, I don't have a lot of time.
Let's talk about triggering suggested videos.
And with that, one of the main factors
and there's thousands of different factors
but there's a lot of weight that actually occurs
for certain factors and the ones that we noticed
that are the most prominent, the one we actually have
a lot of data on are the factors that I'm gonna show you,
which is increased watch-time velocity.
Matt calls it, Matt Gielen calls it view velocity.
I like to do watch time because I think there
is a difference between that and there's a lot of data
that supports it too, it's like the watch time of a video,
watch time of the channel, so on and so forth.
Related meta information upon upload, this is a trigger.
It's gonna trigger all your videos first
and it's gonna say, hey, these or similar topics
and it's similar engagement and there's a similar types
of people that view it and so let's put it out
to other content that they engaged with.
This AI is really, really smart.
Related meta on other people's channels so, for example,
I can guarantee you that when someone watches a video
on my channel that there's a high probability
that Tim Schmoyer would get a suggestion
in one of my related videos and I would for him
because we have a lot of similar content.
We've collaborated together, we've put annotations
and end-screen elements and cards pushing
each other's traffic and we're trying to promote each other
because that's how you grow on YouTube
and so it's gonna out to those suggestions
where it's at, two, upload frequency minimum three times
a week, like if you really wanna see massive growth,
you can't do less than that, now, I know a lot
of content creators have higher production
and I've worked with a lot of higher production crews
and channels and sometimes it's not feasible.
It's not even financially feasible or time feasible
that they can actually do this.
There are other little strategies that makes sense
that you can post more frequently, it's just you have
to get over, well, that's not what I want the channel to be.
It's like what does your viewers want to consume
is the big thing and that should evolve
as you evolve as a content creator.
Next is the viewing history and behavior.
I'm here to tell you that YouTube has a profile on everyone
in this room and in this building of what they like,
how long did they engage with content
and when I say engage, it's like they might hit rewind.
They might watch it two or three times, they might share
it with someone, they might embed it on their website.
YouTube has a lot of this data and they understand behavior
very well and that behavior, this is what I wanna explain
is if you understand the behavior of your viewer
and you help the viewer do certain things,
you're gonna see numbers like what we're talking about here
'cause it's just gonna through the roof.
Main factor and this is one of the triggers
that really disappoints is YouTube looks
at your last performance of your last video.
So you have your last video that just tanks (chuckles)
like it's just bad it just literally bombs,
that next video you have is not gonna be suggested as much.
And so don't expect that to get this high volume.
Now whenever we have a Hero video that we put a lot of time,
effort, energy, and money in, we actually look
at what are the first, the next four videos before it
and what's the video after it and we put that a part
of our Hero plan for that because you have to lead
up to it and get momentum and and really know your audience
well enough of what they'll actually consume
and engage with and then literally spend
more time on your thumbnails.
Triggering too, when you actually create a playlist
it creates an ID, when you actually add a video
in a playlist, it does some pretty amazing stuff.
So this is what I want everyone to do tonight
or today sometime, just go at six videos
into a brand-new playlist and do that just today
and tomorrow and see what happens.
You'll be amazed and it's just, it could be random videos.
It doesn't have to have any type thing.
It's just create that instance that it's
actually happening and see what triggers
from your current views, also on the related content
that you're actually putting out there.
New playlists, I stress that if you're uploading
content daily that you should create
new playlists two or three times a week.
These are new playlists and there's a lot
of content creators says, oh, Derral, you can't do that
because my subscribers will be inundated.
Your subscribers don't care.
This is all for the AI for YouTube and realistically,
there's hardly any people that click the playlist
and see what playlists you have under that tab.
It just doesn't happen but they're searchable
and YouTube, listen, YouTube associates
those videos together.
And where do you think that they associate it?
And if you can get traffic between them,
then you're getting more suggestions when you do
related content that's in that specific playlist.
Also, updating cards is a very big thing.
I think a lot of people just put it in there
and it's static and they forget about it.
A great tool that everyone should own is TubeBuddy.
The reason why love TubeBuddy so much is
because I can change 1,000 videos and change five cards
on every video in about 20 minutes
where it woulda took me weeks to do something like that.
And you can go to TubeBuddy.com/go.
And that's an affiliate link, by the way.
(audience chuckling)
Next, updating end-screen elements.
If you know, if I know that, hey, do you know what?
I want more traffic to come in that seven-day period
'cause this one video took off and it was great
and then it had that 24-hour decay.
I wanna be able to do boost it and I goose it all the time
where what you do is you change all your end-screen elements
at the end of every video and point to that one video
and then also, there's a featured video one.
You push that to it, and you can get all this influx
of views that are coming from people
that are watching your content to a specific video.
Do you keep it on there for a long period of time?
No, don't keep it on (chuckles) more than a day
or you'll destroy what you're doing but then I would revert
back to what you had before.
And so keeping track of this is really important
and especially if you have a method behind that.
Now I only have like
six minutes, there's a lot of information
that we went through and two,
if everyone's on Twitter, if you literally will Tweet me,
if you don't get your question asked here,
just put Ask Derral on it and I will respond to it
when I get a little bit of time
but does anyone have a question here first off?
Dane, we'll start with you and then we'll go to you next
but I don't know, we have a mic, can we--
- Where? - Right here in front.
- [Dane] Derral, do you think meta data in playlists
is important or not, the descriptions?
- Anywhere I can put, go ahead.
Do it one more time.
- [Dane] Meta data in playlists or descriptions
in the playlists important or not?
- Yeah, I do it 100% of the time and the same reason
is how is YouTube gonna know what these videos are about?
Here, do not put spammy titles in your playlist
because there's an automatic filter that kicks in
that if you put certain, like if you try to stuff it,
it will automatically give you a community strike.
Trust me, I've been there, I've gotten a strike before,
yes, even me, but yes, I like to fill everything out
because YouTube doesn't have data on the playlist
and the only way that you can give 'em data is within
the videos that are incorporated in that playlist but also
the meta information.
- (chuckles) Spammy titles, so like
for example if you're doing product reviews on cameras
or whatever use like Canon camera, amazing Canon camera,
vlog camera and that's all one title, that's a spammy title.
We had a question right here.
She has her hand up, fourth row back.
- No.
- So I would use all 10,
10 spots that you can do for the sections
is what they call it.
They don't necessarily all need to be playlists.
YouTube's going to announce or they announced some features
that are gonna be really cool that I can't talk about
that will help cross promotion of other channels
if you own those channels or you wanna partner
with some channels but I'd use all 10 spots
but I would be very hesitant of what you actually put
on the front page just because that's your front page
of your channel but keep it in the playlist section.
Let's do right behind, he had his hand up the longest,
that's with, yeah, you.
- [Man] You mentioned that the positive reception
of the video is one of those important things create
some financial success, I'm gonna assume that YouTube's AI
needs data (muffled speaking) positive in the form of likes,
or comments, or shares or something.
- Doesn't, that-- - Who would know?
- I was with the search and discovery team.
These are the people that are in charge of that division
and it has no relevance on suggestions, none,
but what it does is what the viewer does in the video
and how they engage with the video and what they do next
and where to go and how they interact,
that makes a huge difference and so you're thinking,
why isn't it just likes, comments, and shares?
Well, it's more than that because how they watch the video
based on the other percentages that you actually have
for that particular viewer, if it's higher than normal,
then they're actually gonna see more of that content
than they previously had, so if they watch a dog video
but they really watched the dog video
and they engaged with it a little bit differently,
for the next little bit, they're gonna see
a lot of dog videos, so right in the back.
Roberto, like right in the back.
- [Roberto] Microphone or--
- Yeah, no, no, no.
(audience chuckling) No.
- [Roberto] All right, thanks Derral, we talk a lot,
you and I, about user experience and then user behavior
is what's triggering things now, that's what the AI
is looking for, if you were gonna summarize
the user behavior that we should be focus on
in three to five points, what does that look like?
- Yeah, that's a powerful question and I wish I could do
a presentation all on that but I wanna make sure
that when the viewer watches it, I wanna predict
what they watch next, I wanna predict what comes up
in Next Up and the four suggestions following that
'cause if I can pull that 'em in there.
Everyone can pull out your phone and see
before the first swipe, that you'll see four videos
and when you swipe, you'll see all that content.
That makes a big difference if it's your content
before they move on to somebody else's content.
Great question.
Right here, let's see, we need another female.
Is there another female?
Oh, right there, right there.
- [Woman] Earlier, you mentioned card strategy,
how you (muffled speaking) 0000--
- Not for every channel but it might work for you, yeah.
- [Woman] Do you have a reason for that?
- Just because they're not gonna click on it.
But when it opens up for the next one,
it gives them an option below and above.
It's just real estate.
- [Woman] Do you have any particular strategies
that you use for cards (muffled speaking)?
- Like on my channels that I own or--
- [Woman] Where you place them other than average audience--
- Well, average audience retention is one
and then I basically look at different drop times
and I try to put 'em at each point so but keep in mind
there is the feature tab
that it brings up the feature video.
Don't compete with that 'cause they'll cancel
each other out, so you wanna make sure
that it's separate from that.
Right here.
- [Woman] You mentioned titles and the thumbnail.
You mentioned titles and thumbnails being super important.
Is there something that you've seen that works really well
with thumbnail and titles--
- Well, yeah, it's like understand your audience
so you have a tweenny audience or you have
a teenager audience, you gotta appeal to them
but ultimately, YouTube has some great resources on it
but it should tell a story, you should be able
to look at it, I think Vsauce and a few others
have some of the best thumbnails and titles out there,
very simple but yet it tells a story.
And the more that you can do that, it's really a good.
But check out Team Edge, like I really love their thumbnails
because everyone is a reaction story within a,
I don't even have to watch the video and I don't
but I look at their titles and their thumbnails
but yeah, great question, just right next.
- [Man] There's a rumor that if you schedule
or have your video (mumbles) too long it publish,
it won't perform as well, I should just put it public
immediately (faint speaking). - Yeah.
I used to believe that as of my whole life
until yesterday and--
(audience laughing)
It's true, because I noticed some changes to it
and I was able to get a little bit more data
from the right people and so I changed my thoughts on it.
I still need to test it so I'm still in the skeptic stage
but more leaning towards it doesn't matter anymore.
I think it's when you actually hit Publish and it's public,
not unlisted, not private but it's public that that occurs.
- [Man] That's when 24-hour clock--
- That's when 24-hour clock hits.
I did test that one out and that we noticed
that there's no difference between that
but yesterday it wasn't so, okay, way in the back.
We gotta go to the back and then we're
just wrapping up here.
- [Man] Really loved that last question.
What if you uploaded and you unpublished very quickly?
Is there any drawbacks-- - That's the worst thing
you could ever do in your whole life.
(audience chuckling) You wanna know why?
- Yeah, why? - Because you probably
shoot HD, right?
- Yeah. - Okay.
So what happens is until you see
every single resolution on, like you'd have
to click in the video and you have to click
to select a resolution, the different sizes of the video,
yeah, that would hurt you, so like I always do 4K,
60 frames per second, until I see that from all the way down
to the most bottom one, which is like 122
or something like that, then that's when I'm ready
to publish because what happens is not every viewer
has HD access and they might be looking
at 360 or whatever and they're not gonna
be able to see your video, they're gonna see,
oh, there's an error that's here.
And that's really bad through your click-through rate.
So they went on it, they went to go do it but guess what?
They couldn't see anything so they jumped off
and those are very bad signs so what I would do,
I mean, they've improved this but it might be an hour
or two on 4K 60 frames, everything else should do it
within 10, 15 minutes, you should be ready to go.
So then we had another question in the back with the beard,
just hand him, right over there, there we go.
That's a good-looking beard, by the way.
- [Man] Thank you, so my question is about playlists
and adding to playlists, and that is means using YouTube's
auto-add to playlist feature based on tags
from your own channel, will that help with suggestions?
- That's the only way that I would add to playlists
when uploading new content so it goes
to some general playlists and I like to do that with tags
and also in title that it just triggers that
and based on the content because every channel is similar
that it has similar themes, I want 'em to go
into certain playlists that build up those things.
So yes, that would make a big difference.
Think about this for a second, if you literally upload
a video and say, oh, crap, I forgot to add it to a playlist.
You will add it to a playlist in the video itself.
What do you have to do?
You have to hit Save, you don't ever want to hit Save
for at least seven days, so be very careful what you you do.
There are some data that it looks like they might've fixed
some of those things but as in the past,
it was very dangerous but you can add it to a playlist
if you go into the video manager, select just
the little radio icon and then up at the top,
say, okay, add it to these playlists,
that's a very good option for that
and then there's some other things you can do as well.
Okay, just two more questions and then I will be outside
until my voice is gone and we can answer that
so let's do a female over here
and then who really wants to have it on this side?
Okay, you over here, you, you, which one?
- Yeah, me. - All right.
Sounds good, and like I said, tweet me, I'm at,
you can't see it, it's like let's see if I can do this.
It's @derraleves and you can put #AskDerral on it, so sweet.
- Yeah. - Okay.
- [Woman] What is your titling site, do you, I wanna like
Google keywords and SEO and all that stuff
or do you care more about your personal audience--
- I care that when someone sees the title,
that they're like, oh, that's very interesting.
Oh, yeah, I wanna click on it and watch it
and then they look at the thumbnail and says,
oh, that's very relevant to what the title is.
That's what I care most about, to do stuffing
for positioning of search engine, mark my words,
I could care less if someone finds me in Search.
You wanna know why?
Because if I get 'em through suggestion and have them
click through that, I could get maybe a million times
more visibility than I could otherwise.
And even with my own content, like my channel
that I spend 20 minutes a week on
if I have time for my own how-to videos,
I really don't care for search, I just care for suggestion
just because it is my number one traffic source
so okay, last question over here in the hat,
reverse hat, super cool guy.
And then like I said, I'll be outside
and answer any questions or whatever.
- [Man] To your point about playlists,
if you have multiple channels and say we follow the strategy
(muffled speaking) algorithm would you recommend
playing videos from those different channels together--
- It really depends, it really depends if the channels
have similar content, if they don't have similar content,
I would completely avoid it.
There is a couple things is YouTube will see
how many people are watching your videos,
where in the world they're watching their videos,
like what country, how long they're actually doing it,
how they're engaging with it, and so you wanna be very,
I have one of those big mistakes that I made
that I decided, hey, this is gonna be great content
worldwide and I focused worldwide before I focused
into a core audience and now all my audience
is like in Indonesia and Saudi Arabia.
Not saying those are bad, the CPMs are just really low there
and so I just didn't miss that core audience that's there
so to overcome that, and this is a big factor,
YouTube's gonna look at data immediately
and so whatever your subscribers are
or wherever they're at is gonna help figure out where
that actually get suggested out in the land of YouTube.
Guys, thank you so much for this, being here.
(audience applauds)
I'll answer some questions, follow me on Twitter.
Ask the questions there, thank you.
All right, guys, don't forget to get that
free training at VidSummit.com/freereplays
and that's something that is 32 hours
of the best and brightest
minds in the industry talking about the business side
of YouTube and then also if you are interested
in coming to my next VidSummit, it is in October 10th
through the 13th and it's gonna be
the biggest and best event ever, it's 1,000-plus
people, that's just the business side.
It's not a fan experience, it's just 100% the business,
the brass tacks about how to really grow audiences,
monetize audience, work with brands, and what brands want
and also agencies is all sides of this business
and it's the business side of the industry.
And I'm really excited about it
so I wanna personally invite you.
Go ahead and check it out at VidSummit.com
and don't forget to pick up your free replays.
Guys, thank you so much for watching this long video
about the algorithm, hopefully, you found it useful.
If you have additional questions for me,
make sure you hit that Subscribe button
and the bell notification and also just tweet me a question
and when I get (chuckles) a free time, a moment or two,
I'll answer it, thanks, guys.
Không có nhận xét nào:
Đăng nhận xét