The Entire Longform Algorithm Explained in 300 Seconds...

[00:00] When you start a new long form channel,

[00:01] the first thing that happens is YouTube

[00:02] reviews your account to figure out if

[00:04] you're a real person and you're worth

[00:05] pushing out. If your channel looks empty

[00:07] or suspicious, YouTube has no reason to

[00:08] trust you yet and won't push your videos

[00:10] out anywhere beyond search. To avoid

[00:12] this, you need to make your channel look

[00:13] legit from the start. Warm it up, brand

[00:15] it personally, use an older, trusted

[00:16] email, and enable third-le feature

[00:18] eligibility. The less your channel looks

[00:20] like spam, the more willing YouTube is

[00:21] to push your content out beyond just

[00:23] search. YouTube is way more forgiving

[00:24] with long form than shorts. If you're

[00:26] not getting into the browse feed, your

[00:27] video is either bad or your channel just

[00:29] looks like spam. After you upload your

[00:31] first video, YouTube immediately

[00:32] transcribes everything you say, scans

[00:34] your thumbnail, and classifies your

[00:35] content. The algorithm is trying to

[00:36] figure out what your video is, who's it

[00:38] for, and whether it's safe to push.

[00:40] Then, over the next 24 to 48 hours, it

[00:42] continues to build a profile on your

[00:44] content with the main thing being your

[00:45] target audience. During this time,

[00:46] YouTube isn't really pushing your video

[00:48] out anywhere. It's just reviewing your

[00:49] data. That's why it takes a couple days

[00:51] before views start coming in on the

[00:52] first video of a brand new channel. This

[00:54] is where the metadata becomes the most

[00:55] important thing. YouTube's entire goal

[00:56] is to keep viewers on the platform for

[00:58] as long as possible. More time on

[00:59] YouTube means more ads watched, and more

[01:01] ads watched means more money to YouTube.

[01:03] So naturally, the algorithm's job is to

[01:05] find the videos that specific viewers

[01:06] would enjoy the most. To do that, it

[01:08] needs to know exactly who your video is

[01:10] for. If the algorithm can't figure out

[01:11] your target audience, it would rather

[01:12] push out someone else's content and

[01:14] actually understands rather than risking

[01:16] showing yours to the wrong person. Like

[01:17] I said before, you really need to focus

[01:19] on your metadata here. Your title needs

[01:21] to clearly tell YouTube what the video

[01:22] is about. And then tags, chapters, and

[01:24] timestamps all give the algorithm more

[01:26] context. Your thumbnail matters, too,

[01:27] because YouTube's AI is actually reading

[01:29] it. It scans for faces, objects, and

[01:31] text. The clearer everything is, the

[01:32] faster it matches your video to the

[01:34] right people. Typically, channels that

[01:35] write vague titles and leave

[01:36] descriptions empty or forcing YouTube to

[01:38] guess. And once again, YouTube doesn't

[01:40] want to take the risk showing it to the

[01:41] wrong person. Once the data is collected

[01:43] and the target audience is determined,

[01:45] YouTube starts testing your video

[01:46] through browse. They collect initial

[01:47] data and decide whether to keep pushing

[01:49] it or not. Similar to shorts, the

[01:50] algorithm pushes out your videos in kind

[01:52] of waves. YouTube is testing your video

[01:54] against small groups and watching how

[01:55] they respond. If they respond well,

[01:57] YouTube slowly ramps up impressions.

[01:59] Then, depending on how the next group

[02:00] responds, it either keeps ramping up,

[02:02] slowing down, or plateaus. And before

[02:03] you guys ask me about why my studio

[02:05] looks so much better than yours, I use

[02:06] New Studio, which is by far the best

[02:08] Chrome extension if you want to take

[02:09] YouTube seriously. So, how do we get

[02:10] YouTube to push your video as much as

[02:12] possible? Well, this comes down to the

[02:13] stats. There's three stats you need to

[02:15] be focusing on with the first being

[02:16] click-through rate. This is the percent

[02:18] of people who saw your video in their

[02:19] feed and actually clicked on it. Every

[02:21] person who sees your video in their feed

[02:22] is considered an impression. Picture it

[02:24] like this. Every single impression on

[02:26] YouTube is like a slot. There's a

[02:27] limited amount of slots in a given feed.

[02:29] So, YouTube is pretty selective with it.

[02:31] Why would they give a video that has a

[02:32] low click-through rate an impression

[02:33] when they could give an impression to a

[02:35] video that actually has a high

[02:36] click-through rate and it's proven that

[02:37] people are willing to click it? It's

[02:39] much more valuable for YouTube. Your

[02:40] click-through rate is determined by your

[02:42] thumbnail, your title, and your idea.

[02:44] This is the first thing a viewer even

[02:45] sees. So, you really need to focus on

[02:47] this. Your goal is to make your video

[02:48] the most appealing to click on out of

[02:50] every other impression on their feed.

[02:51] But the best way to stand out is being

[02:53] in a niche that has a ton of viewers

[02:54] willing to watch. To help with that, I

[02:56] have a playlist of a bunch of viral

[02:57] niches I find, and I'll be constantly

[02:58] updating it. These are niches I would

[02:59] use myself. A lot of you guys were

[03:01] asking me how I find my niches. So, if

[03:02] you're wondering, majority of these

[03:03] niches are found with algrow. But what

[03:05] percent should you be shooting for?

[03:06] Well, the average CTR on YouTube sits

[03:08] around 4 to 5%, and anything below

[03:10] three, YouTube definitely won't push it.

[03:11] The next stat the algorithm focuses on

[03:13] is the retention graph. This is by far

[03:15] more important than CTR. You want this

[03:16] graph to be as flat as possible, as high

[03:18] as possible, and the initial dip to be

[03:20] as small as you can manage. The graph

[03:21] will always go down over time because

[03:22] viewers just naturally leave no matter

[03:24] what you do. If you're losing too many

[03:25] viewers at the start, your hook is bad.

[03:27] If your graph is too steep, your

[03:28] storytelling and video ID is bad. But

[03:30] after the algorithm pushes your video in

[03:31] the browse feed for a bit, it starts

[03:33] being shown in the suggested feed. This

[03:34] is the sidebar when someone's watching

[03:36] another video. YouTube really only

[03:37] pushes your video in the suggested feed

[03:38] after it tests it in browse first

[03:40] because the suggested feed is much more

[03:42] dependent on your target audience. So,

[03:44] YouTube naturally waits until they have

[03:45] more data before showing your videos

[03:46] there. These impressions in the

[03:48] suggested feed are much more valuable to

[03:49] YouTube because their goal is to keep

[03:51] people on the platform for as long as

[03:52] possible. Now, these two stats are the

[03:54] most visible things you can improve. But

[03:55] beyond that, the algorithm tracks a ton

[03:57] of different things. Engagement, what

[03:58] the comments are about, likes, shares,

[04:00] subscriptions gained or lost, returning

[04:02] viewers, viewer satisfaction surveys,

[04:04] not interested clicks, search terms,

[04:06] whether your video started a session,

[04:08] whether it ended it, and what the viewer

[04:09] did after watching. And those three

[04:10] things I said at the end all fall into

[04:12] session time. If your video makes

[04:13] someone spend more time on YouTube, this

[04:15] is the strongest signal you can send to

[04:16] the algorithm. This is why linking your

[04:17] viewers to a playlist or another video

[04:19] at the end helps so much, even if it's

[04:21] not related to you. You can see this in

[04:22] action on my videos. The first long form

[04:24] video on my channel went viral. One of

[04:25] the reasons for that is because at the

[04:26] very end, I linked viewers to a playlist

[04:28] with a bunch of viral niches. This

[04:30] playlist got tens of thousands of views,

[04:31] and because my video, even though it was

[04:33] relatively short, I got viewers to spend

[04:34] more time on YouTube, and the algorithm

[04:36] rewards that. As you continue to post

[04:37] more videos and start gaining traction,

[04:39] you start to build an audience. YouTube

[04:40] shows your next videos to people who

[04:42] watch your previous ones first, so you

[04:44] always have a solid seed audience going

[04:45] into every upload. But if you're

[04:46] struggling finding ideas, once again,

[04:48] you can use Algro for this. You can

[04:49] actually integrate with Claude, and

[04:50] you'll be able to instantly analyze all

[04:52] your competitors through using it. So

[04:53] check it out, and if you want a free

[04:54] personalized channel review or need help

[04:56] with anything, join the Discord.