How Does the YouTube Algorithm Work in 2023?
Master the Algorithm to Grow Your Audience
It’s a simple question, really. But it can make all the difference for video content creators: How does the YouTube algorithm work? Knowing the answer can make the difference between going viral and staying invisible.
Currently, YouTube’s recommendation algorithm drives 70% of all views on the platform. But despite its influence, the algorithm’s inner workings remain mysterious.
When you’re publishing videos, you may be struggling to find the viewership you’re looking for. There’s a good chance that that’s because the algorithm doesn’t rank your content very highly. Learning how it ticks pays off.
However, learning more about the algorithm is also useful if you’re already enjoying some success. It can be a great way to take your channel to the next level and really profit from the creator economy.
Here, you’ll learn about some of the YouTube algorithm’s lesser-known secrets. We’ll discuss how you can use it to get your videos to the right viewers – and compete on the platform.
What Is the YouTube Algorithm?
At its most basic, the algorithm is a constant loop of collecting data and recommending videos to users. It gathers data from viewers and channels to learn who will watch, like, and share which kinds of videos.
And it collects tons of these data points – about 80 billion signals every day.
What for?
Ultimately, YouTube’s goal is to keep users on the platform for as long as possible. It does this by making it easy for you to find an endless supply of videos it thinks you’ll like.
The algorithm’s work appears in several places across the site and the app.
First, there is YouTube’s homepage, showing you rows and rows of video suggestions:
Second, the algorithm powers YouTube’s search function. If two different users search for the exact same phrase, the results will vary in content and ranking based on their individual history.
Third, there’s the “Suggested Videos”. These shows up in a column on the right edge of the page. The first of these video suggestions is also “up next” – that is, queued for autoplay at the end of every clip. The videos’ end screens themselves will show even more suggestion.
Finally, there’s YouTube Shorts, which gives you a practically endless feed of clips.
There are slight differences to how the algorithm picks videos for each of these places. That’s because they all serve slightly different purposes.
Technically, there’s no such thing as “the” YouTube algorithm. The platform actually uses a number of systems that each have their own specific job, and constantly optimizes all of them.
This goes especially for the Shorts algorithm. After all, it powers a fairly new feature. Given the competition by TikTok and Reels, it’s likely to evolve rapidly. For instance, it used to only take Shorts into account – which is why a lot of YouTubers then launched separate Shorts channels. That’s no longer necessary, though the editing on short-form video does have its own demands.
The underlying principles of the recommender algorithms are the same, even if they differ in some details. Let’s now go on to explore those principles.
The Principles: How Does the YouTube Algorithm Work?
YouTube funnels videos to users based on their feedback.
This feedback can either be explicit (clicks, likes, dislikes, subscriptions …) or implicit (video watch time, shares). Yes, YouTube does hide dislikes now, but it still uses them internally.
With your data, the algorithm automatically builds user profiles. In the background, YouTube’s engineers use fancy machine learning and artificial intelligence models, but we’re not after the technical details here. Over YouTube’s history, these details have changed a lot, as well.
Something interesting happens in the next step: It groups similar users together, to fill out the picture of what individual users may or may not like.
Subsequently, YouTube uses this profile to decide what suggestions to make. This is a two-step process: The algorithm first generates candidates, then ranks them.
Here’s a rough overview of this:
Step 1: Candidate Generation
Whether suggestion or search, the starting point is the same: YouTube’s vast video library, or corpus. The Shorts algorithm obviously works on the smaller, but rapidly growing set of short clips.
The first step, candidate generation, has the job of selecting just a few hundred out of billions of videos. The picks reflect the viewer’s implicit preferences and demographics.
But these implicit preferences are more than just watch history. It’s also how much they watched of each video.
This demonstrates an essential part of YouTube’s algorithm. A video watched to the end carries much more weight than those abandoned after seconds.
That’s one reason clickbait can backfire – the algorithm deals out a penalty for videos that get a lot of clicks, but little watch time.
Another critical measure of implicit preference is which videos you shared in the past.
The candidates generated in this step are all likely to be relevant for the viewer. Still, there’ll be too many to suggest all of them. Shorts on mobile, especially, has basically no space to work with.
This leads to the second step: ranking.
Step 2: Ranking Video Candidates
The variables that go into ranking are pretty similar to those of candidate generation, but a few new factors now come into play.
Consequently, your implicit preferences matter less. The algorithm will now prefer videos with more engagement: that is, more views, likes, shares, and comments.
In a nutshell: After generating a bunch of choices YouTube is pretty sure you’ll like, it ranks them by well they do in general.
Every video gets a score, and the algorithm ranks them all according to that score.
But then comes one extra step: mixing it up.
To ensure that the suggestions are as varied as each user’s interests, some randomness is introduced.
After all, every user has their favorite topics, most of us like a bit of variety. Humans get bored. And that’s why YouTube doesn’t just suggest more of what it knows we like, but spices it up a bit.
This introduced diversity works in two ways. First, it encourages watching a variety of subjects. Second, it also introduces videos from less popular channels.
So, even if a channel’s videos aren’t top-ranking, they still have a chance to make it into the recommendations. Because they might be just right for you, and that’ll keep you on YouTube. This goes double for Shorts.
After a video passes through this second step, the algorithm will finally suggest it to you – whether on the homepage, an end-screen, or as the next-up Short.
How Can You Harness this Algorithm to Generate Views?
It’s important to have this basic understanding of how the YouTube algorithm works in theory. Yet the more valuable lesson for creators is how to use it to succeed on the platform.
Here are six critical steps to getting your videos in front of an ideal audience.
1. Increase Viewer Watch-Time
One of the most important things we’ve talked about above is that it’s important how much of each video your viewers watch. This means focusing on interesting, engaging content that your viewers will enjoy.
In a nutshell: Don’t make videos longer than they need to be. Do everything you can to get your viewers to stick with you until the end. Keep them entertained throughout. Drop hints about exciting tidbits of information that you reveal later on. You might include a fun outro in every video.
This requires some reflection on the quality of your videos and the responses that they’re getting. YouTube Studio’s analytics can tell you pretty much exactly where attention starts to drop off. Do more of what works and less of what doesn’t!
The more viewers watch your entire video, the more likely YouTube will be to suggest it to others.
This even goes for Shorts: Engagement on long-form content also boosts your chance of getting your short-form content in front of people. Your subscriber count and similar metrics matter much less for it.
Pro tip: When you post your videos is also important. If you time it right, the initial boost of your fans watching the whole thing can push your content up in the recommendations. This feedback circle is even stronger with Shorts.
(See also our full guide to getting more watch time!)
2. Get Viewers to Watch Multiple Videos
Viewers who’ve already watched several videos from your channel are much more likely to see more of your videos in their suggestions.
The first step to earning multiple views is to put out a lot of engaging videos within your niche (we’ll talk more about niching down in a moment). While it’s true that this is a lot of work, you’ll want to publish regularly anyway.
There are also a few key tactics to encourage this type of engagement.
One easy strategy is just to mention past videos on your channel. Simply say something like, “If you enjoyed this video, then you’ll love the one from last week about Topic X.”
Include as many links as you can to your other content. Here are just some suggestions:
- Pin a comment with a link to your most popular video.
- Create a playlist of audience favorites and of related videos.
- Include an end screen with links and thumbnails of similar videos.
Thinking carefully about your content strategy makes this much easier in the long run. If you produce videos that are related to begin with, it’s much simpler to create these connections. They’ll also look natural to your viewers. Video series are excellent for this.
Take vlogmas, for example: Many vloggers do daily videos and weekly round-ups in the run-up to Christmas. This gives their audience a lot of closely related content to watch!
All this encourages viewers to check out more of your videos. And that, in turn, makes it more likely that they – and similar users – see your content suggested.
3. Focus on Shares, Not Just Subscribes
Getting subscribers is vital in building an audience, of course. But getting viewers to share your content is just as important.
Like watch time, shares are an essential part of the implicit feedback recommender algorithms consider. As we’ve seen above, implicit feedback features in both candidate generation and ranking.
Thus, it has a huge influence on which videos your audience sees on their homepage.
Beyond that, shares are an organic way to expand your channel’s viewership. On the one hand, you get the strengthened position of implicit recommendations. On the other, direct links to your videos will spread all across the web!
So, encourage your audience to share your videos on Twitter, Facebook, Instagram, or anywhere else they like to hang out. You can even produce short teasers and vertical remixes that are perfect for Shorts, Reels or TikTok. And: This signal boosting can attract even more shares on those platforms.
4. Build Credibility Within Your Niche
New YouTubers often post videos about anything and everything that piques their interest. It’s an integral part of finding a niche that works for you and connects you with your ideal audience.
But – unless you’re really just doing this for yourself – you’ll need to settle down a bit eventually. It’s really difficult to build and keep an audience if your videos are all over the place.
Once you’ve determined your niche, it’s important to build credibility within that niche. Improving your reputation will drive viewers to more of your videos, giving you better engagement and better implicit feedback.
Becoming a credible authority takes a lot of time and videos to achieve. Good planning and consistent publishing give your channel the best chance at success.
5. Expand Video Topics to Hit Multiple Niches
One lesser-known way to increase organic reach is hitting several niches at once. True, it’s crucial to find a focus. But if you find overlapping subjects, they might pull in otherwise disinterested users.
Let’s take a video that focuses solely on beauty products as an example. Only people who have viewed videos about beauty products are likely to see it suggested.
Now consider a video about a celebrity’s favorite beauty products. There’s a good chance it’ll also get suggested to fans of that celebrity – even if they don’t usually watch videos about beauty products.
You don’t need to force this tactic on every single video. Still, try to include content that will interest various people for a variety of reasons.
Here’s an example: YouTube cook Andrew Rea regularly recreates dishes from popular movies and books – and makes it fun. Fans of Everything Everywhere All At Once, for example, had a field day with his Everything Bagel.
6. Create Positive, Emotionally Compelling Content
Finally, YouTube’s algorithm increasingly rewards positive, emotionally engaging content.
It’s true that YouTube has been accused of promoting hateful or dangerous content in the past. But recent changes to the algorithm have shown a shift in ranking. There is now more of a preference for videos likely to encourage good feelings in the viewer.
Ethics – and a broader trend called reactive creativity – isn’t the only reason for this.
YouTube actually found that total watch time goes up when controversial content is downranked. Just as important, for both YouTube and you, is that advertisers avoid borderline content.
This gives the platform a solid motive not to suggest emotionally negative content. While YouTube doesn’t show dislikes anymore, the algorithm certainly takes them into account.
This advice applies to more than your video content itself, though. It also applies to titles and thumbnails. Avoid titles with clickbait, shock value, or controversy to drum up interest. Don’t use ALL CAPS, and don’t make false claims to mislead potential viewers.
Your videos are most likely to get suggested if they are authentic and connect with your target audience.
Keeping Up to Date: How the YouTube Algorithm Works Changes Constantly
There is one more thing for content creators to consider: The YouTube algorithm changes practically every day. In a typical year, its engineers make hundreds of modifications to the algorithm.
However, as a creator, your job is to publish great videos, not keep up with the ever-changing algorithm. That’s where we come in. Not only does the team at Tasty Edits stay up-to-date on every change. We also help make sure your content is ready to take advantage of them!
Whether you’re just getting started or looking to take your online presence to the next level, Tasty Edits can help.
Want more tips for producing video content that viewers will love (and earn you money)? Reach out to us.
