21st century matchmaking: getting your content to the right viewers on YouTube

Last time we had a look at some of the ways product placement is used in connection with marketing in today’s media environment - with a particular focus on how endorsements are increasingly being used in an unconventional and, often, seemingly counter-intuitive way.

Today, we turn our attention to how brands, as well as content creators, are ensuring their content reaches the right audiences. Let’s take a look at just how viewers are presented with relevant content and how you can benefit from that.

Advertising has come a long way since the days of the single screen cinema intervals or the so-called Golden Age of consumer goods. Before the 21st century, those in marketing faced a significant hurdle in ensuring their ads were not only effective, but also shown to the appropriate audience. Particularly when it came to cold calling and junk mail, consumers were nearly universally allergic to personalised ads of this kind.

Marketers had very limited options when it came to ensuring that promotional video content made it into the right hands - short of initiatives such as scheduling ads during a particular kind of film, TV programme, or sporting event (think beer at the Superbowl or Guinness during the rugby) in order to target a particular kind of viewer.

Today, things are different, and it’s all thanks to the Internet.

If you’ve logged into YouTube in the last few years, you may have noticed that certain videos are pushed towards you during certain kinds of activity. You will have a range of recommended videos on the homescreen, as well as next to videos. Even more recently, YouTube introduced autoplay, where a relevant video will play automatically once your current video ends. Some of these recommendations seem, at times, uncannily accurate - and that’s no accident.

In their own promotional material, YouTube boast that “YouTube streams over 4 billion videos a day, and our targeting tools are so precise that you can show your ad to folks around your corner or to anyone around the world.” Using a series of (rather opaque) algorithms, YouTube’s aim when it comes to recommended content is twofold: to provide users with highly personalised video content (with the aim of delivering a seamless, quick video service to audiences), and to provide brands with a means of distributing their content to the right people (with a goal of generating revenue).

There are a number of main challenges content creators face if they wish to take advantage of this, which need to be overcome for videos to become recommended. These are detailed in The YouTube video recommendation system, a white paper developed by Google staff. To summarise, these include:

  • Insufficient metadata, such as incomplete or irrelevant titles, tags, descriptive information;

  • Gauging viewer interest - there is no way of telling if a user actually engaged with a video or liked it based solely on the fact that they watched it;

  • Data sources - these include content data and user data;

  • Keeping content fresh - YouTube’s database of videos is so vast that only ‘fresh’ content will generally become popular.

So how does it work?

YouTube’s recommendation system does the bulk of the work when it comes to resolving these challenges. It combines an analysis of information about the video (its category, what its topic is, as well as ‘video quality’ according to its audience macro-reception) with information about individual users’ viewing habits, using a method known as associate rule mining.

By analysing the sort of content viewers watch in a single session, YouTube is able to determine what sort of thing they might want to watch next - while accounting for diversity of videos and interests. Put simply, videos that are too similar are removed from video queues, while those that fall too far from a user’s viewing interests are not included. YouTube therefore pushes content to users based on their general or overall interests, rather than solely on the last video they watched.

So how can I make it work for me?

Kissmetrics have produced a fantastic guide to optimising your channel, which you should read in full, but I will also summarise it here.

  • If YouTube recommendations are determined primarily based on groups of interests, rather than individual sectional interests, then organising your content into categories can be of major help. That’s why Kissmetrics recommend creating descriptive playlists of videos with a focus on keywords for titles. In their own words: “Playlists are beneficial because they help facilitate more views to your content by helping define the subject matter of a video for YouTube’s search algorithm when it comes to the playlist’s title and description.

  • As we’ve discussed, the algorithms rely heavily on metadata, and getting metadata correct is therefore absolutely crucial. Titles, tags, descriptions are all key.

  • Use annotations - these are there for a reason! Relevant calls to action (“subscribe here”, “see more fun stunts!”, or “thumbs up for garlic bread, comment if you’re against” are all classics), and links to relevant content will provide your content with a huge boost. That raises an important point - include calls to action in your videos as well! That’s why bloggers are constantly encouraging their viewers to get involved in a conversation in the comments.

  • Finally, brand your YouTube channel well. We’ll be getting onto the specifics of that next time.