{"id":295824,"date":"2023-11-15T12:02:07","date_gmt":"2023-11-15T12:02:07","guid":{"rendered":"https:\/\/wyzowl.com\/?p=295824"},"modified":"2024-06-12T15:57:41","modified_gmt":"2024-06-12T14:57:41","slug":"youtube-algorithm","status":"publish","type":"post","link":"https:\/\/wyzowl.com\/youtube-algorithm\/","title":{"rendered":"How the YouTube Algorithm Works in 2024 (+10 Tips)"},"content":{"rendered":"
<\/div>\n

Last updated on 10th January 2024<\/em><\/em><\/p>\n\n\n\n

In this article we\u2019re going to take a look at how the YouTube algorithm works in 2024. <\/p>\n\n\n\n

If you want to be successful on YouTube, then understanding how the algorithm works is a must<\/em><\/strong>. <\/p>\n\n\n\n

Why?<\/em><\/p>\n\n\n\n

Well, YouTube\u2019s algorithm actually <\/strong>drives 70% of what people watch<\/strong><\/a> on the platform<\/strong>. So if you want people to find your videos, you need to play by the algorithm\u2019s rules. <\/p>\n\n\n\n

And we want to help you! We\u2019re going to take a look at everything you need to know about the YouTube algorithm – from the history to misconceptions to how it works in 2024 – so you can make more informed video marketing decisions. <\/p>\n\n\n\n

<\/span>A brief history of the YouTube algorithm <\/strong><\/span><\/h2>\n\n\n\n

2005 – 2007 (pre-algorithm)<\/strong><\/h3>\n\n\n\n

Like most great inventions, YouTube was created to solve a problem. And the problem was that it was difficult to share videos online.<\/p>\n\n\n\n

Enter: YouTube. <\/strong><\/p>\n\n\n\n

Here\u2019s what the platform looked like in 2005:<\/p>\n\n\n\n

\"YouTube<\/figure>\n\n\n\n

Source: <\/em>Wayback Machine<\/em><\/a><\/p>\n\n\n\n

This was YouTube, pre-algorithm. Although, there were ways to get your video noticed. As you can see, there are tags on the right-hand side (the bigger the tag, the more popular it was). And we can also see videos featured on the home page. <\/p>\n\n\n\n

But, if you look closely, you\u2019ll notice that these aren\u2019t in any particular order that would suggest an algorithm at play. These are the most popular videos on a random day in 2005:<\/p>\n\n\n\n

\"YouTube<\/figure>\n\n\n\n

The video with the best ratings is 4th. The video with the most views is 3rd. The video that comes in first place only has one tag. So, it\u2019s really difficult to grasp how these videos were being organised. <\/p>\n\n\n\n

And, according to some<\/a>, it\u2019s because these videos were manually <\/em>curated in YouTube\u2019s early days. Basically, it was one person\u2019s job to trawl through all of the videos uploaded on YouTube and put their favourites on the homepage. Imagine having that kind of power today! <\/p>\n\n\n\n

Of course, as more and more videos were being uploaded, YouTube needed to find a different way to organise their video content. And they decided on views. <\/p>\n\n\n\n

Videos with the most views were suggested to all users, regardless of the topic or content of the video. Here\u2019s an example from 2007: <\/p>\n\n\n\n

\"YouTube<\/figure>\n\n\n\n

The problem with this, of course, is that you end up catering to only one group of people. This \u2018organise by views\u2019 method didn\u2019t (or rather, couldn\u2019t) take into account different cultures, different languages, and basically anyone else who had interests that didn\u2019t match that of the average YouTube user pre-2008. <\/p>\n\n\n\n

2008 – 2011 (the recommendation system)<\/strong><\/h3>\n\n\n\n

YouTube made a big change in 2008 when they started building their recommendation system. <\/p>\n\n\n\n

According to a 2021 article by Cristos Goodrow<\/a> (VP of Engineering at YouTube at the time), not a lot of people watched the videos in YouTube\u2019s featured section prior to 2008 because (as mentioned) they weren\u2019t relevant enough to individual users. Instead, the majority of YouTube\u2019s viewership came from search results and shared links. <\/p>\n\n\n\n

To change this, the team started working on a recommendation system. Goodrow explains:<\/p>\n\n\n\n

\u201cWe start with the knowledge that everyone has unique viewing habits. Our system then compares your viewing habits with those that are similar to you and uses that information to suggest other content you may want to watch. So if you like tennis videos and our system notices that others who like the same tennis videos as you also enjoy jazz videos, you may be recommended jazz videos, even if you\u2019ve never watched a single one before.\u201d<\/em><\/strong><\/p>\n\n\n\n

As such, the home page in 2009 looked more like this: <\/p>\n\n\n\n

\"YouTube<\/figure>\n\n\n\n

The algorithm was starting to take shape. But there were still some issues. Clicks and views continued to be the primary ways to rank videos in terms of popularity, and this led to an increase of what we now call \u2018clickbait\u2019 – the use of a misleading video title and\/or video thumbnail to increase YouTube video views.  <\/p>\n\n\n\n

By 2011, YouTube started tweaking the algorithm to try and figure this out. They learned that just because users clicked on a video didn\u2019t mean they watched it. They also built classifiers at this time to actively prevent certain videos from being recommended, for example, videos that were racy or violent. <\/p>\n\n\n\n

2012 – 2015 (introducing watch time)<\/strong><\/h3>\n\n\n\n

YouTube wanted to make more personalised recommendations, so they started to pay attention to watch time (which is the amount of time a user spends watching a particular video). <\/p>\n\n\n\n

Surprisingly, after implementing this they saw an immediate <\/strong>20% drop in views<\/strong><\/a>. However, they stuck with it, believing it was more important to focus on delivering value to viewers by showing them content that they usually watch (over content that was just clicked on a lot). <\/p>\n\n\n\n

With watch time, YouTube was able to start recommending users relevant content that they might like across multiple categories. Here\u2019s a YouTube home page from 2014: <\/p>\n\n\n\n

\"YouTube<\/figure>\n\n\n\n

As you can see, most of the content on this user\u2019s home page comes from recommended channels (based on their watch history). <\/p>\n\n\n\n

Another thing the YouTube algorithm started taking into account during this time is likes, shares, and dislikes. This information gave YouTube\u2019s algorithm more solid information on what users wanted to see, and what they didn\u2019t. <\/p>\n\n\n\n

2015 – 2017 (increasing social responsibility)<\/strong><\/h3>\n\n\n\n

By 2015, YouTube\u2019s user base had surpassed the billion<\/em><\/strong> mark and the platform started to consider their social responsibility. <\/p>\n\n\n\n

There was a shift towards \u2018responsible\u2019 recommendations as more people started heading to YouTube for their news content. In 2015, the platform took steps to prevent sensationalistic tabloid content from appearing on homepages. <\/p>\n\n\n\n

The idea of what was a responsible recommendation and what was not expanded further during this time, and in 2017 YouTube began evaluating the machine learning that powered fairness across protected groups, such as the LGBTQ+ community.   <\/p>\n\n\n\n

2017 – present (deep learning)<\/strong><\/h3>\n\n\n\n

As you can probably tell, the YouTube algorithm hasn\u2019t necessarily changed over the years – from one thing to another – rather, it\u2019s grown more intelligent and more layers have been added. <\/p>\n\n\n\n

These layers continued to build from 2017 to involve more deep learning and tighter control on harmful content and misinformation. <\/p>\n\n\n\n

One of the biggest changes during this time was a 2019 update to begin demoting \u2018borderline content\u2019, this is content that comes close to violating YouTube\u2019s Community Guidelines without actually doing so.  <\/p>\n\n\n\n

YouTube uses classifiers to identify whether a video is \u2018authoritative\u2019 or \u2018borderline\u2019, and they actually went old school with this, bringing in human evaluators to assess the quality of information across different videos and channels.<\/p>\n\n\n\n

<\/span>3 biggest misconceptions about the YouTube algorithm <\/strong><\/span><\/h2>\n\n\n\n

There\u2019s a lot of information out there about the YouTube algorithm, but not all of it is true. And some of these misconceptions can make it difficult to understand how the YouTube algorithm works. <\/p>\n\n\n\n

Here are 3 of the biggest misconceptions: <\/p>\n\n\n\n

1. The algorithm is there to find an audience for your videos<\/strong><\/h3>\n\n\n\n

While this seems perfectly logical, it\u2019s not true. In fact, the exact opposite<\/em> is true. The algorithm suggests videos that are a good fit for viewers, based on audience behaviour, rather than finding a YouTube audience that\u2019s a good fit for your videos. <\/p>\n\n\n\n

This is an important thing to understand and keep in mind when creating video content for your YouTube channel, because if your videos don\u2019t match what people are looking for then YouTube simply won\u2019t push your videos out to viewers. <\/p>\n\n\n\n

2. There\u2019s only one algorithm<\/strong><\/h3>\n\n\n\n

This can be confusing because many outlets – including this one – will refer to the YouTube algorithm (singular), but in reality it\u2019s actually a system of algorithms that work in conjunction with one another. <\/p>\n\n\n\n

For example, the \u2018search\u2019 algorithm  is different from the \u2018trending\u2019 algorithm, because they\u2019re both looking for different things. There\u2019s also algorithms for video engagement, channel stickiness, and so much more – that we\u2019ll get into in just a second! <\/p>\n\n\n\n

So, it\u2019s important to continuously monitor your video performance to see what works best for your YouTube channel. Just because one video performs really well with the \u2018trending\u2019 algorithm doesn\u2019t mean it will match up with what\u2019s required to succeed with the \u2018search\u2019 algorithm (for example). <\/p>\n\n\n\n

3. YouTube prefers longer videos<\/strong><\/h3>\n\n\n\n

While it\u2019s true that longer videos tend to perform better on YouTube than on other social media platforms that aren\u2019t video-centric, that doesn\u2019t mean you should create extra-long videos for the sake of it. <\/p>\n\n\n\n

YouTube likes engagement, watch time, and channel stickiness, basically anything that keeps viewers watching and keeps people on YouTube! <\/p>\n\n\n\n

It can be easy to mistake this as YouTube preferring longer videos, but if you create long videos that aren\u2019t watched in their entirety then this is likely to have a negative impact on your YouTube success. <\/p>\n\n\n\n

<\/span>How does the YouTube algorithm work in 2024?<\/strong><\/span><\/h2>\n\n\n\n

Today, the YouTube algorithm constantly learns from over 80 billion signals<\/strong><\/a> to help connect people to videos they love. <\/p>\n\n\n\n

Unfortunately for video marketers and video creators, there is no handbook that YouTube gives out explaining exactly how the algorithm works. But there is content here and there that gives tidbits of useful information, like this video directly from YouTube: <\/p>\n\n\n\n

\n