0:02 I've been doing YouTube for nearly a
0:03 decade now, and over the course of my
0:06 career, I've heard all sorts of advice
0:07 tossed around. Some of this advice
0:09 dramatically reshaped the way I thought
0:12 about the YouTube algorithm. But some of
0:14 it, looking back, was just complete
0:16 nonsense. And what better way to
0:19 structure a video with a list of bad,
0:22 mid, and good things than a tier list.
0:25 Let's get into it.
0:27 We're going to start things off with F
0:30 tier. And this tier is for the truly
0:32 terrible advice. Advice so bad that
0:35 following it will hurt your channel. And
0:36 there's not going to be too many things
0:39 in this tier. Most advice is at least
0:43 applicable in some niche cases. Uh or at
0:45 least is at the very worst just a waste
0:47 of time. Uh but the stuff in this tier
0:50 is worse than that. It will mess with
0:53 your analytics. It will derank your
0:55 videos in the algorithm. Don't follow
0:58 the advice here. But to start us off, we
1:00 have you should delete bad, old, or
1:03 underperforming videos. So, this is
1:06 disastrously bad advice. But to fully
1:08 explain why, let me give you a quick
1:10 rundown on how the YouTube algorithm
1:12 works. This will, of course, be
1:15 massively simplified. But to be honest,
1:17 the core concept is really not all that
1:19 complicated. The YouTube algorithm works
1:22 by comparing viewers watch histories and
1:24 looking for patterns. So the most basic
1:26 pattern would be something like this.
1:28 Viewer one gets on YouTube and gets
1:30 shown a random selection of videos on
1:34 their homepage. They watch videos B, D,
1:37 and G. Viewer two gets on YouTube and
1:41 watches videos B, D, and G. Viewer three
1:45 gets on and watches videos B and D and
1:48 then logs off for the night. YouTube is
1:51 going to put video G onto viewer 3's
1:53 recommendations front and center. the
1:55 next time they log on because it noticed
1:57 the pattern that people who like videos
2:00 B and D also tend to like video G. There
2:02 are of course hundreds if not thousands
2:05 of additional variables, but all of
2:07 those additional complexities layer on
2:09 top of that core concept, patterns of
2:12 viewership. YouTube calls this co-watch
2:15 behavior, and it is the main driving
2:16 force behind the recommendation
2:19 algorithm. If you can understand that
2:20 concept, you can probably follow my
2:22 rationale for basically every tip in
2:25 this video. So, the reason that deleting
2:27 old videos is bad for growth is that it
2:29 also removes that data from your viewers
2:31 watch histories, which means there's
2:32 less of a pattern for YouTube's
2:34 algorithm to pick up on to use to push
2:36 your videos to new people. Now, of
2:38 course, there can be genuinely good
2:40 reasons for deleting an old video. if
2:41 you, I don't know, accidentally doxed
2:43 yourself and included some sensitive
2:45 information, yeah, probably don't just
2:46 leave that up for strangers to see for
2:49 your own safety. But in general, it's a
2:51 bad idea. And if you really just don't
2:52 like your old videos and want them gone
2:55 for reasons of taste, uh, that's okay,
2:58 too. I suppose you can also just unlist
2:59 them, which won't remove them from
3:01 people's watch histories, but that also
3:03 means that if they want to watch it,
3:04 they can just go back through their
3:06 watch history and find it. So, it's not
3:08 a good solution for really sensitive
3:10 information, but if you just don't want
3:11 it to show up on your channel, it does
3:13 work. I'm personally of the opinion that
3:15 it's a good thing to let audiences go
3:17 back and see how much you've grown as a
3:19 creator. But I understand that not
3:21 everyone can stomach the cringe of their
3:22 teenage self, but if you're just
3:24 deleting it because it didn't get the
3:26 view numbers you wanted, you're making
3:28 the problem worse, not better. I
3:30 actually know a few YouTubers who delete
3:32 low performing videos to make sponsors
3:33 think that they hit higher view counts
3:35 more consistently than they actually do
3:37 so they can use that data to negotiate a
3:40 higher rate. But this is playing with
3:42 fire in my opinion. And second in F
3:45 tier, we have the advice all engagement
3:47 is good engagement. I think everyone who
3:49 thinks about this for more than a second
3:52 knows it's not true. But still, every so
3:54 often I'll see a creator going back and
3:56 forth bickering with one of their fans,
3:58 and they'll invoke this as a sort of
4:00 attempt to reframe the whole interaction
4:03 as like a gotcha. Like the viewer will
4:06 be criticizing or insulting some aspect
4:08 of the creator's video, and the creator
4:10 will counter, "Aha, I got you. You see,
4:12 you thought you were damaging my
4:14 reputation by insulting me, but actually
4:16 you've just given me more engagement,
4:17 boosting my influence even further.
4:21 Haha, jokes on you. But this is usually,
4:24 frankly, a pathetic attempt to stop
4:27 people from criticizing them. But still,
4:28 there are some people who earnestly
4:30 believe this. I've even heard of a few
4:33 creators who will include one or two
4:36 spelling mistakes on purpose in order to
4:38 generate more engagement in the form of
4:40 comments that are correcting them
4:42 thinking that that is helping their
4:44 recommendations reach more people. It's
4:46 it's just not I don't know how else to
4:49 say it. Uh but the people who literally
4:50 design the YouTube algorithm have said
4:53 as much. It doesn't work. Don't do it.
4:55 Now, of course, if a creator posts
4:57 something that's so bad that the badness
5:00 of it becomes a spectacle in and of
5:04 itself, that can override the negative
5:07 effects of negative engagement. But even
5:09 in those cases, usually the big winners
5:11 in that scenario are going to be the
5:13 commentary channels that farm that
5:16 misfortune for views. That's honestly
5:18 the only advice I've heard which I do
5:20 think actively hurts your channel's
5:21 ability to rank highly in
5:23 recommendations. Let's move on to D
5:26 tier. This tier is for advice which I
5:28 don't think actively hurts your channel,
5:31 but can indirectly hurt it by nature of
5:33 just being a complete waste of time.
5:35 This is going to be a pretty packed tier
5:37 because honestly, most of the advice
5:40 that I hear as a creator falls into this
5:43 category. First in Dtier, we have you
5:45 should post at specific times or days of
5:49 the week. For like 99% of YouTubers,
5:51 this just doesn't matter at all. Now,
5:53 sure, if most of your audience is
5:56 American and you post at 3:00 a.m.
5:59 Eastern time or you post during the
6:02 Super Bowl or you post during Christmas
6:04 morning, yeah, the amount of people who
6:05 are immediately dropping everything to
6:08 go watch your video is going to be
6:10 lower. But if a viewer likes your
6:12 videos, they're going to see it the next
6:14 time they open YouTube. And in fact,
6:16 your video already being up and ready to
6:17 watch when they sit down at their desk
6:19 for the first time to start their day
6:21 might actually play in your favor. It
6:22 might not be good for their
6:25 productivity, but still. The point is,
6:27 in the long run, a video is going to get
6:28 about the same amount of views that it
6:30 would have no matter what time it was
6:32 uploaded. So, don't stress about it
6:36 unless you make time-sensitive videos.
6:38 If you're doing a news show or, god
6:41 forbid, YouTube drama videos, you do
6:43 have to keep your topics relevant to the
6:45 current news cycle. And if you make
6:48 videos about internet trends and fads or
6:51 new releases in gaming and film, you
6:53 don't want to miss the boat on those.
6:55 Although, even in those cases, a good
6:57 video can still work after the fact as a retrospective.
6:59 retrospective.
7:00 But yeah, for most creators, this just
7:02 isn't something you need to concern
7:04 yourself with. I used to upload all of
7:05 my videos at 10:00 a.m. on Saturday
7:08 morning because that's when I personally
7:10 would have been most likely to just be
7:11 chilling looking for something
7:14 interesting to click on. But eventually
7:15 I stopped caring about it and just
7:17 started uploading whenever a video was
7:19 ready and I didn't see any difference in
7:22 performance at all. Next in Dtier, the
7:24 algorithm will punish you for taking a
7:27 break. So in general, no. There's no
7:28 amount of time you can take off that
7:30 would trigger some sort of massive
7:31 downranking of your videos in
7:33 recommendations. If your audience really
7:35 likes your videos, they're going to keep
7:37 sharing them with their friends and
7:39 probably even go back and watch some old
7:41 ones. As an example, at the beginning of
7:43 the year 2020, I had about the same
7:45 amount of subscribers as my friend Sam
7:47 from the YouTube channel Sam Onella
7:50 Academy. Sam proceeded to take a nearly
7:53 three-year long break, all the while
7:55 massively diffing me on both subs and
7:57 views. And this isn't me saying there's
7:58 anything wrong with the algorithm or
8:00 anything like that. His audience just
8:02 really likes his videos and likes to go
8:04 back and rewatch them. And when he
8:06 finally did come back after those 3
8:08 years, all of his new videos did
8:10 basically just as well as you would have
8:12 expected had he never taken a break at
8:15 all. Bill Wartz made what I think could
8:18 be reasonably described as the greatest
8:20 educational YouTube video of all time.
8:22 And even though he hasn't posted in
8:24 years, his channel still pulls plenty of
8:27 views. If he ever posted another history
8:30 of blank, I think the internet would go
8:33 bananas. Now, these channels are of
8:35 course outliers. I don't think everyone
8:36 can expect that level of audience
8:39 loyalty. But what I'm getting at is that
8:42 the performance of good old videos is
8:45 not negatively impacted by the lack of
8:48 new videos. Even channels whose videos
8:49 are highly time-sensitive still take
8:51 breaks from time to time, and they still
8:53 have an audience when they come back.
8:55 Now, obviously, you don't want to be too
8:57 inactive. If you just up and ghost the
8:59 internet for years at a time, it's not
9:01 just your audience who's going to
9:03 notice. YouTube channels are businesses,
9:05 and like any business, you're going to
9:07 have competition. If you've left your
9:09 audience with nothing to watch for years
9:12 on end, I would expect your competitors
9:14 to notice and for them to try and
9:17 emulate your style in order to capture
9:19 some of your audience. So, take breaks
9:21 when you need to. Don't burn yourself
9:24 out, but never think that success is
9:27 guaranteed either. Next in Dtier, just
9:30 do this one weird trick. All right, I
9:32 don't want to stoke drama, so I'm not
9:35 going to name names, but I do have a
9:37 bone to pick with. Okay, I'm
9:39 re-recording this bit because after
9:40 watching it back, I think I started to
9:42 come off way too heated for no reason.
9:45 Basically, what I want to communicate is
9:46 that there's a lot of videos out there
9:48 that can really just get you barking up
9:50 the wrong tree if you're not careful by
9:52 convincing you that there's a lot of
9:54 subtle tweaking that you can do to your
9:56 video's metadata. you know, things like
10:01 the video tags and the file names and
10:03 like things you can do with the YouTube
10:05 Studio editor that would allegedly
10:07 massively boost your odds of going
10:10 viral. And I'm sorry, but I do think
10:11 that that's all just a complete waste of
10:14 time. I'm not saying that adds zero
10:16 value at all, but at the end of the day,
10:19 if your videos are bad, there's nothing
10:21 that you're going to be able to do like
10:23 metadata manipulation wise to get people
10:26 to like them. And if your videos are
10:28 excellent, frankly, it would be pretty
10:30 hard to screw it up. Like, it would be
10:32 legitimately difficult to get people to
10:34 stop sharing them. And if your videos
10:36 are mid, you're much better off spending
10:38 your time learning how to improve your
10:41 craft than you are learning all these
10:44 subtle metadata manipulation tricks. You
10:46 should always prioritize the things that
10:48 your audience can see, not what
10:50 YouTube's algorithm can see. Next in
10:53 Dtier, one bad upload can kill your
10:55 channel. Thankfully, probably not.
10:57 Again, flip it around. How many times
11:00 have you, as a viewer, completely
11:02 written off all of a creator's future
11:05 work based on one bad video? Probably
11:07 not too many times. Although, if you've
11:09 been around long enough, maybe you can
11:11 think of one or two. But no, what
11:14 actually kills channels is a string of
11:16 uninteresting videos that leads to their
11:19 audience just not being as excited about
11:21 their stuff anymore. But yeah, don't
11:24 stress about singular flops. It's way
11:26 too much pressure to put on yourself.
11:28 And literally every YouTuber uploads a
11:30 stinker every once in a while. And last
11:33 in Dtier, monetization will hurt your
11:35 YouTube channel. It's crazy. I know a
11:37 handful of creators who just agonized
11:39 over the decision to either monetize
11:42 their videos or do sponsors. They
11:44 thought their audience would crucify
11:46 them if they did it, but Patreon wasn't
11:48 quite making ends meet for them. And so
11:50 eventually they felt like they had no
11:52 other choice. And then when they flipped
11:54 the switch, they got literally zero
11:56 negative comments about it. And the same
11:58 is true for sponsors. In fact, oftent
12:00 times I see the exact opposite happen
12:01 where they'll suddenly get a lot of
12:04 comments along the lines of like, "Yo,
12:06 our boy finally made it. Hype." It's an
12:08 understood milestone in a creator's
12:10 career, and I think most audiences
12:12 understand that. I've even heard of
12:14 creators faking sponsorships because
12:16 they feel like it gives them more
12:19 legitimacy to have them than to not. I
12:21 wouldn't do that. Faking sponsors is
12:24 also Dtier advice, but so is completely
12:26 writing them off as selling out. That's
12:28 it for Dtier. All the rest of the stuff
12:30 on this list is at least somewhat
12:33 applicable and definitely won't hurt
12:36 your channel. In Ctier, we have advice
12:39 that is potentially useful, but isn't
12:42 universally applicable. And first in C
12:43 tier, this is going to be a
12:46 controversial one. Collabs are a good
12:47 way to grow your channel. This one's
12:49 interesting because YouTube Creator
12:51 Studio actually tells you that this is a
12:53 good idea. I did a ton of collabs in the
12:56 early years of my YouTube career, and to
12:59 be honest, I don't think a single one
13:01 moved the needle at all. Now, I'm not
13:03 saying they can't work. They can make
13:05 for fun and entertaining videos, but for
13:07 growing an audience, one, they're not
13:10 that great, and two, you want to build
13:12 your own audience, not leech from
13:14 someone else's. That second point isn't
13:16 some moral judgment or anything. I'm not
13:17 saying you're not going to have any
13:19 overlap with other channels audiences.
13:20 Of course, you will, but I've seen it
13:23 several times where a creator will do a
13:25 collaboration and it'll do well and
13:27 they'll get a ton of new subscribers and
13:29 views and they'll get all excited and
13:32 celebrate. But when they upload their
13:35 next video, it doesn't hit and they
13:37 might even lose subscribers. And the
13:39 reason is because those subs weren't
13:42 made in earnest. They were subbed as a
13:44 show of support for a different creator
13:47 and for you as their friend, but not
13:48 because they actually connected with
13:52 your content. And that can be really
13:54 demoralizing. You may have noticed that
13:56 collabs in general are just not all that
13:58 common these days on YouTube. And I
14:00 think this is the reason. they're just
14:02 not all that effective in the long run.
14:04 Now, there are other creators who I
14:06 think do it well, but I think the reason
14:09 they make it work is that they do cameos
14:12 or games where it makes sense to involve
14:15 other creators. My friend Magic the Noah
14:17 does this amazingly well with his game
14:20 shows, but even then, after being in
14:22 multiple videos, I don't think I've ever
14:24 noticed the spike in subscribers or
14:26 views. So, do the collabs that make
14:30 sense, but definitely don't force it.
14:32 And especially don't chase down other
14:34 YouTubers asking to collab because it's
14:37 not even worth it. Next, in C tier,
14:40 censor your speech using euphemistic
14:42 slang. Look, I'll be honest, I really
14:45 wanted to put this one in F tier. I find
14:47 it immensely disrespectful when people
14:50 use such corny language to discuss
14:52 serious topics, but I will admit that
14:55 this is a matter of personal opinion and
14:57 that at this point clearly that type of
15:00 thing has become pretty firmly ingrained
15:03 in the online vernacular. I watched a
15:04 pretty interesting series of videos by a
15:07 channel called Ethmology Nerd, who makes
15:09 a pretty good case that these words do
15:12 serve a social function beyond just
15:14 dodging a platform's sensor bots. I
15:16 don't know. I'm not an ethics
15:18 philosopher, so I will leave it up to
15:20 y'all to decide what is and is not
15:22 appropriate for serious online
15:26 discussions. But what I do know is that
15:30 these supposed sensor bots do not exist.
15:32 There are no guidelines saying that you
15:35 cannot use terms like kill, sex, and
15:38 yeah, even rape and pedophile in your
15:40 videos. Now, sure, if you're using those
15:43 words in a graphic context, that's one
15:46 thing. But if the context is there, if
15:48 your video is about those subjects,
15:51 removing the word doesn't suddenly make
15:53 your video familyfriendly. A true crime
15:57 podcast about vicious evil murderers is
16:00 still a true crime podcast about vicious
16:02 evil murderers, even if you say unal
16:05 alive instead of kill. So either talk
16:09 about it or don't. Next in C tier,
16:12 premieres help your channel grow.
16:15 Premieres are fine. They can be a cool
16:18 way to build hype and even result in a
16:21 bit of bonus revenue from super chats.
16:24 But do they help you grow? Probably not
16:27 in a significant way. In fact, some
16:30 people, myself included, find them to be
16:32 a bit obnoxious. It's a bit of a
16:34 buzzkill when you're bored and looking
16:36 for something to watch and you see a new
16:39 notification from your favorite creator
16:42 and it's for a premiere that's in 24 hours.
16:44 hours.
16:46 Cool. I'd rather you just have told me
16:48 tomorrow. And if you're late to a
16:50 premiere, you got to either wait for it
16:54 to end or awkwardly try to rewind it.
16:56 Uh, meanwhile, chat is talking about a
16:59 totally different part of the video and
17:02 spoiling stuff. And yeah, premieres are
17:07 fine. Messy, but fine. Next, in Ctier,
17:09 you should engage with your audience by
17:11 doing update videos, Q&A videos, and
17:14 milestone videos. So, on my main
17:16 channel, Tearsu, I decided very early on
17:19 that I was going to never do these. So,
17:21 when I was first starting out, I created
17:24 a list of rules for myself to keep me on
17:27 track for growth and keep me out of
17:29 trouble. And the number one rule on that
17:32 list was to only ever upload exactly the
17:34 sort of video that I thought my audience
17:37 had subscribed for. My channel was about
17:39 looking at the natural world through the
17:42 lens of gaming and nothing else. My
17:45 assumption was my audience did not care
17:47 who I am. And I do think this rule
17:49 served me well. I mean, not to toot my
17:51 own horn, but for a while I was the
17:52 fastest growing educational channel of
17:55 all time and reached a million
17:58 subscribers after only about 18 months.
18:00 But of course, I have no control data to
18:02 compare against, and I know plenty of
18:04 other channels that have done just fine
18:06 uploading that style of video alongside
18:08 their main content. I think it makes
18:09 more sense to flip it again to the
18:12 viewer perspective and ask if your
18:15 favorite channel or actually let's say
18:16 not your favorite channel. If a channel
18:19 that you like but perhaps isn't even in
18:23 your top five uploaded a Q&A video,
18:25 would you watch that? But I guess this
18:28 also goes back to one of the earlier
18:30 points on this list, which is that one
18:32 bad upload isn't going to sync your
18:34 channel either. The performance of my
18:35 first video on this channel certainly
18:37 proved to me that there can be a ton of
18:39 value in doing these more personal
18:41 videos as I've been able to sinceense
18:43 connect with more of my audience and
18:45 fellow creators than ever before.
18:47 There's clearly merit to both approaches
18:49 and I look forward to learning more
18:51 about how to best make this genre of
18:54 content. So subscribe if that's
18:56 something you're interested in. And last
18:58 in C tier, basically everything you've
19:01 ever heard about fair use. I wish I had
19:02 the time and the expertise to try and
19:04 properly debunk all the misinformation
19:08 about fair use, but for now I will just
19:10 say go watch the Tom Scott copyright
19:13 video. It's crazy though. In my
19:14 behind-the-scenes video, I got a lot of
19:17 comments saying, "Why on earth would you
19:19 pay for stock footage when fair use
19:21 exists?" Look, I do wish some of them
19:24 charged a bit less, but ultimately I do
19:27 want an incentive to exist for nature
19:29 videographers to go shoot interesting
19:32 videos for me to buy. If it were legal
19:34 that I could just rip their videos and
19:36 run it through a watermark remover, that
19:39 would be cool for a short time, but
19:41 eventually these nature videographers
19:43 would just find better uses for their
19:46 time and film experience. I want to have
19:47 a good relationship with nature
19:50 photographers. I want them to be excited
19:52 to show me the latest cool shots that
19:54 they got, not just to hide behind some
19:56 flimsy legal defense while I benefit
19:58 from their work for free. As with most
20:01 things, the answer to the question of
20:04 why does this cost so much is because
20:07 it's worth it. Maybe not to you, but to
20:10 someone. All right, first in B tier.
20:14 Now, interested is interesting. So, this
20:15 is a common adage I've heard tossed
20:18 around in YouTuber circles a lot, I
20:20 believe it was first popularized by
20:22 Casey Neistat, but I'm not 100% sure.
20:25 So, while it is true to an extent, uh, I
20:27 learned the hard way that if you make
20:30 things too niche, too high concept,
20:33 people will bounce off of it. Still,
20:35 passion is by far one of the most
20:37 important things to showcase in your
20:40 videos. YouTube is pretty great at
20:41 showing people videos relevant to their
20:44 interests. And if you among all the
20:46 YouTubers for a given topic seem like
20:48 the most passionate of the bunch, you'll
20:50 be in a pretty good position. And the
20:52 reverse is also true. If people can tell
20:54 that you're not all that excited about
20:56 the thing that you're talking about,
20:57 they're not going to be excited about it
21:01 either. Next in B tier, shorts will help
21:03 your channel grow. So strictly from the
21:05 perspective of increasing the number of
21:08 views and subscribers on your YouTube
21:11 channel. Yes, this absolutely will work.
21:13 In fact, the numbers can be truly
21:15 staggering. But what can also be
21:18 staggering is how difficult it can be to
21:20 convert those shorts viewers into long-
21:23 form viewers. This is my friend Juliet.
21:25 She runs a channel called Girl Who Eats
21:27 Art. It's a shortssonly channel about
21:30 showcasing edible mosaic artwork. Her
21:32 videos have a combined 1 and a half
21:35 billion views and her channel has 1 and
21:37 a half million subscribers. For
21:39 reference, that is over three times as
21:41 many views as I have across all of my
21:43 channels. And she's only been doing it
21:45 for a year and a half. Juliet just
21:48 uploaded her first long- form video. An
21:50 8-minute long video of her touring an
21:53 art museum and interviewing people.
21:55 Before I tell you, how many views would
21:59 you guess that video got? 10,000,
22:03 100,000, a million? 431?
22:09 less than a tenth of a percent of her
22:12 subscriber count. So, what's going on?
22:14 Why did her video get so few views?
22:16 Well, I actually think the answer is
22:19 pretty simple. While most long- form
22:21 viewers do watch a few shorts here and
22:25 there, most short form viewers never
22:26 even open the YouTube homepage where
22:29 long form videos are shown. I've seen
22:30 some viral posts going around saying
22:32 that YouTube Shorts should have just
22:35 been a separate app. And if you ask me,
22:38 it basically is. So if the last thing
22:40 you watched on the YouTube app was a
22:43 short, the next time you open YouTube on
22:45 your phone, it'll open directly to the
22:47 shorts feed, so you never have to see
22:50 anything else. Now, if you ask me, it is
22:53 a bit ridiculous that shorts are also
22:55 shown on the YouTube homepage,
22:56 especially with these tall thumbnails
22:58 that take up double the real estate and
23:00 crowd out long form stuff. But in
23:03 fairness on the Smart TV app, which is
23:05 frankly where long- form creators most
23:07 prefer their videos to be watched
23:09 anyway, shorts usually don't show up
23:12 there. Anyway, the point is you can pull
23:14 insane numbers on your shorts, hundreds
23:17 of millions of views, but still struggle
23:20 to pull even a tiny fraction of a
23:21 percent of those viewers over to your
23:24 long- form videos. You really do have to
23:26 treat them like a separate audience.
23:28 Still, if the content of your shorts
23:30 specifically references long- form
23:33 content that people watching that short
23:36 might also like, you can see excellent
23:38 growth across your entire channel via
23:40 shorts. This is something that I myself
23:43 am far from mastering, but my friends
23:45 over at Extra History have been crushing
23:47 it with this strategy. So, yeah, shorts
23:49 are powerful for growth, but they're
23:51 easy to overestimate and tough to
23:53 convert into the audience you actually
23:56 want. On a similar point, my next piece
23:58 of advice in Btier is to ask your
24:00 audience to subscribe during your
24:02 videos. So, there's really no reason not
24:05 to do this. I do it. All the best
24:08 creators do it. And we do it because it
24:12 works. It makes the number go up and we
24:14 really like it when the number goes up.
24:17 The reason I include it now here is
24:19 because I think it's time to drop the
24:22 biggest truth nuke of the video, which
24:25 is that subscribers don't matter very
24:27 much. I wish they did. A YouTuber's
24:28 subscriber count is often their biggest
24:31 point of pride, and it certainly is for
24:33 me. Hitting a million subscribers was
24:34 one of the biggest milestones of my
24:36 life. But if you're going to ask your
24:38 audience to do something, you know
24:40 what's better than asking them to click
24:42 the I'd like to see more of this in the
24:43 future button? getting them to watch
24:45 more of your videos right away. In the
24:47 same amount of time that it takes to ask
24:49 your audience to click the red button
24:52 and tap the bell, you could just suggest
24:54 another video to them that they might be
24:56 interested in. Spending more time on
24:58 your channel right away is a lot
24:59 stronger of a positive signal to the
25:01 algorithm than just tapping a button.
25:03 Subscribers are great, don't get me
25:06 wrong, but it's easy to over prioritize
25:08 growing that number rather than growing
25:10 your channel's actual reach. Next in B
25:12 tier, focus on making sure your
25:14 click-through rate is as high as
25:16 possible. This is another number that
25:18 creators often over prioritize, and it
25:20 can be even more misleading. I can't
25:22 tell you how many times I've seen a
25:23 fellow creator complaining on Twitter
25:26 that they must be shadowbanned because
25:29 with a CTR as high as theirs, surely
25:31 YouTube should be pushing their video to
25:33 the masses. The problem is, while yes,
25:36 CTR does scale directly with how
25:38 interesting a video's title and
25:41 thumbnail are, it scales inversely with
25:44 how many impressions a video is getting.
25:46 CTR is always going to start high and
25:49 decline as your video spreads. So, if
25:51 you're looking at your analytics and
25:53 thinking, "I don't get it. My CTR is
25:55 crazy high. YouTube says this one's
25:57 clearly a banger, so why is it not being
25:59 shown to the masses?" Unfortunately,
26:02 that's kind of similar to saying like,
26:04 "I've gone on so many first dates.
26:06 Clearly, I'm hot stuff." So, paying
26:09 attention to CTR used to be more
26:12 important for creators because you could
26:15 take a dead video and then change up the
26:18 thumbnail and then look at the real-time
26:22 CTR data that YouTube would give you and
26:23 you could look for a change. If the
26:25 change swung positive, that was a good
26:27 thing. That was a really good signal.
26:29 Now though, you can just use YouTube's
26:32 AB testing feature to do the same thing.
26:34 Um, so there's not really a ton of
26:36 reason to pay super close attention to
26:38 CTR. Now, looking at your real-time CTR
26:43 can be useful for title changes. Um, cuz
26:45 right now, most people can only do it
26:48 with their thumbnails, but there is a
26:52 title AB testing feature in beta, so
26:55 you'll get it eventually. Last in B
26:57 tier, you should get your audience to
26:59 sign up for a newsletter. This is a
27:01 strategy I hadn't even heard of until
27:03 recently, but apparently a lot of really
27:05 successful creators do it. So, the basic
27:07 idea is pretty simple. Kind of like what
27:09 I was saying before, subscribers don't
27:12 necessarily mean that much anymore. Just
27:13 because someone is subscribed doesn't
27:15 mean they're for sure going to get a
27:17 notification for your next video. So
27:19 instead of relying on YouTube to push
27:21 your video out, you could just ask your
27:23 most loyal fans to join a mailing list
27:25 so that you have total control over that
27:28 connection to your audience. And this
27:30 does work. Data shows that the industry
27:32 average for opening these emails is
27:36 about 20%. Which is like 10 times higher
27:38 than the average rate that YouTube
27:40 viewers click on push notifications. So
27:42 it is a pretty reliable way to reach
27:44 your audience. The really clever
27:46 YouTubers will send out not just their
27:48 own videos, but a whole curated list of
27:50 the best videos that they've seen over
27:52 the past few days or weeks. That way,
27:54 their video becomes part of a pretty
27:56 concrete pattern of co-atching. And
27:57 anyone who organically encounters one of
27:59 the other videos in their newsletter
28:01 will have a much higher chance of seeing
28:03 their video next. So, pretty clever
28:06 stuff. The only drawback is, believe it
28:08 or not, sending a bunch of emails can
28:10 get pretty expensive pretty quickly.
28:12 Like potentially several thousand dollar
28:15 a year. So definitely useful, but not
28:17 something to worry about unless you're
28:19 already an established creator. All
28:21 right, we're leaving the mid tier now.
28:24 So look, I'm not trying to be a YouTube
28:26 guru or anything like that. I'm not
28:28 going to promise that these tips are
28:30 like guaranteed to grow your channel and
28:32 propel your career to new heights and
28:34 make you rich and famous overnight. But
28:36 I do think everything from here on out
28:38 is genuinely great advice that I would
28:39 give to anyone in the industry, both new
28:42 and old creators alike. First in A tier,
28:45 you should use the AB thumbnail testing
28:48 feature. So, this just works pure and
28:50 simple. If you're not AB testing your
28:52 thumbnails, you're almost certainly
28:54 leaving views on the table. Now, it is a
28:56 lot of extra work to create three
28:58 thumbnails instead of just one, but I've
29:01 seen some truly explosive growth on old
29:03 videos after successfully finding a new
29:05 winning thumbnail. I'm talking like
29:06 doubling or tripling the amount of views
29:09 on the video. There's also a title AB
29:11 testing feature in beta right now, which
29:13 opens up even more possibilities. I
29:15 don't have a ton more to say about this
29:17 one. It just genuinely is a really
29:20 powerful tool. Next in a tier, you
29:22 should translate your videos into
29:24 multiple languages. So, pretty much
29:26 every creator that does this that I've
29:28 talked to has said that it basically
29:29 doubled their revenue for very little
29:32 extra effort. Translated subtitles are
29:34 good, but if possible, what you really
29:36 want to do is find someone who can
29:38 provide a translated audio track, too.
29:39 The multi- language audio track feature
29:42 is really powerful and auto plays the
29:43 appropriate language based on viewers
29:46 location if the translation exists. I
29:48 really wanted to make this work for Tzu,
29:50 but because my videos are so
29:53 jargonheavy, unless I find someone who
29:56 specifically knows like the ins and outs
29:59 of gamer culture in that language, the
30:01 translations just don't capture the same
30:04 gaming connotations that they do in
30:07 English. So, yeah, it's tough and
30:09 finding someone who can do that is
30:11 difficult. People like that are hard to
30:13 find and expensive, but if you'd like to
30:16 see tiers you brought to your local
30:18 language, u pledging to my Patreon would
30:20 help me get that off the ground. All
30:24 right, and finally, S tier advice. So,
30:26 for advice in the highest tier, I
30:29 decided to go with more nebulous
30:33 vibes-based sorts of things because at
30:35 the end of the day, I really do believe
30:38 that having the right mindset and having
30:41 the right creative ambition will really
30:43 always trump any sort of back-end
30:46 optimizations or analytics tricks. First
30:49 in S tier, we have the 80% rule. This is
30:51 advice that Hank Green posted to YouTube
30:53 around the time that I was just starting
30:56 my career here. And I never forgot it.
30:59 The idea is simple. 80%. Everything
31:01 creative I do, I do my best to get it
31:04 80% of the way to as good as I can make
31:07 it and go no further. I just don't try
31:08 to get it to 100%.
31:10 >> Now, of course, I've heard tons of
31:13 similar proverbs to this, you know, like
31:15 don't let perfect be the enemy of good,
31:17 done is better than perfect, that sort
31:20 of thing. But I don't know, something
31:23 about the way that Hank puts it, plus
31:25 his track record with starting and
31:28 finishing a huge variety of massively
31:31 successful projects, made this one stick
31:33 differently. I'm sure there's plenty of
31:34 fields that this sort of advice can
31:38 apply to, but I think us YouTubers
31:40 really need to hear it more than most
31:43 because we really like to stress a lot
31:46 about the finer details of things and
31:47 tend to have a pretty well- definfined
31:51 picture of what we want a video to end
31:52 up looking like. And I think for
31:54 YouTubers like myself who upload
31:58 somewhat infrequently and spend a lot of
32:01 time per video, the biggest risk is not
32:03 that we make a bad video. The biggest
32:06 risk is that we don't make a video at
32:09 all. We either spend way too long
32:11 polishing something that's already
32:14 perfectly fine or we get halfway through
32:16 a project and realize it's not quite
32:19 turning out the way we liked it and so
32:22 we just start over. It's tough because
32:23 one of the pieces of advice that I see
32:26 passed around a lot in creative circles
32:28 is just make it exist first. You can
32:33 make it good later. But on YouTube, once
32:35 you publish a video, it's out there. You
32:37 can't change it. You just have to let it
32:40 ride. Mistakes and all. And even with
32:43 that being said, the 80% rule is great
32:46 advice. And the last advice on this tier
32:50 list, the algorithm is the audience.
32:52 Basically, the idea that every time you
32:54 would normally use the word algorithm,
32:58 swap in audience. So instead of the
33:00 algorithm likes this thumbnail more than
33:03 that one, you think the audience likes
33:05 this thumbnail more than that one. And
33:07 instead of thinking the algorithm isn't
33:09 pushing this video, you think the
33:12 audience isn't pushing this video.
33:14 This is advice that Todd Dup prey, head
33:17 of search and discovery at YouTube, aka
33:19 the guy that leads the team that manages
33:21 and designs the YouTube algorithm, says
33:24 every chance that he gets. I've kind of
33:26 been trying to condition you guys to
33:27 think this way over the course of this
33:30 video by frequently asking you to
33:31 reexamine all the pieces of advice
33:34 through the lens of an audience member
33:35 rather than the perspective of a
33:37 creator. The things that matter to the
33:39 algorithm are the things that matter to
33:41 the audience. And if something doesn't
33:43 matter to the audience, it's probably
33:44 not worth fussing about for the
33:46 algorithm's sake either. Personally, the
33:48 things that determine whether or not I
33:50 want to click on a video are, does it
33:52 look interesting? Is it by a creator who
33:54 I really like? Does it look like it's
33:56 about something that I've already
33:58 watched another video about? Do I have
34:00 time to watch something that's that
34:02 length? If the thumbnail looks like a
34:03 still frame from a video, do I want to
34:05 see the rest of that video? And that
34:07 sort of thing. There have been a lot of
34:09 videos on YouTube lately from YouTubers
34:11 complaining about low views and
34:13 speculating about possible algorithm
34:16 changes that are affecting them. And in
34:19 fairness, the video that I posted um a
34:21 few weeks ago was also about challenges
34:24 that I was facing as a creator. But I
34:26 hope the takeaway from that video was
34:27 that I don't believe that there's any
34:30 nefarious YouTube suppression going on,
34:34 but that the audience is just changing.
34:36 And yes, I did hear also about the
34:39 potential ad block shenanigans. And
34:40 those sorts of things do happen from
34:43 time to time, but as a creator, the
34:45 advice that I would give is that you
34:47 still have to just prioritize focusing
34:50 on the things that you can control. What
34:52 viewers want is a moving target, and
34:54 sometimes strategies and formats that
34:56 used to crush it suddenly don't work as
34:59 well anymore. I apologize if the S tier
35:02 section came off as a bit disappointing.
35:05 Uh, I get that just make good videos
35:08 doesn't really feel actionable. Uh, and
35:11 again, I'm not trying to be a YouTube
35:13 guru or anything like that. I'm just
35:15 sharing my opinion as a creator. I hope
35:17 this video helps you avoid red herring
35:19 advice and helps you focus on what
35:22 really matters. And that's the tier
35:24 list. I'm really curious how this video
35:27 does because I have no idea if people
35:28 actually want to just hear me ramble
35:31 about things or if the success of the
35:34 first video was an anomaly. I know I
35:35 kind of baited people into thinking I
35:37 was maybe about to retire with a title
35:39 and thumbnail like that. And I know
35:41 that's not a move I can pull twice, so
35:43 I'd be surprised if a video like this
35:45 does anywhere near as well. I promise
35:46 the next video on this channel will not
35:48 be about the YouTube meta. Let me know
35:50 in the comments what you think. And if
35:51 you'd like to help me fund an indie
35:53 gamedev studio, consider pledging to my
35:55 Patreon cuz we've got some cool things cooking.
