It’s predicted that by 2022, online videos will account for 82% of all global consumer internet traffic. YouTube is currently the second largest search engine globally, and as of 2017, 60% of Australians view online videos every day.
With stats like that, it’s no surprise that 70% of marketers plan on using video marketing in 2019. But with 400 hours of new content being uploaded to YouTube every minute, standing out from the crowd is no mean feat.
Thankfully, SEO best practices ring true on YouTube, just as they do on Google. To get decent organic exposure on YouTube, it’s recommended that you:
- Ensure your titles, tags, descriptions are relevant to search queries
- Include closed captions
- Mark up your videos with structured data (so Google can understand exactly what they’re about).
The third point above was more intentionally phrased than you might have thought, because believe it or not, Google is now actually smart enough to “understand” video content.
And I don’t just mean in the rankings sense, I mean in the video sense. In this wonderful age of AI, Google can parse, categorise and understand many of the elements in your video, including what the shots are of, and how often certain items appear – and for how long.
This is all thanks to Google’s Cloud Video Intelligence platform, which has only been around for 2 years. At the moment, the platform is aimed at media businesses and corporations that have petabytes of video content to sift through. As the API “watches” through videos, the algorithm labels entities (such as dogs, flowers, and people) and makes the catalogue of its findings. The software even transcribes speech, tracks objects, and detects text.
You can try the API here and see how it works with a sample set of videos.
Currently, the API is only available privately to certain businesses, and typically at varying costs. But the fact that technology like this currently exists makes it worth more than a wee bit of consideration. Content creators and marketers may need to brace themselves for upcoming updates that could have a huge effect on video SEO.
Videos have been taking top positions for user queries on Google for some time already, however there’s no official word from Google as to whether or not they’re using this tech to rank videos. But why wouldn’t they? If they’re able to scan through YouTube’s petabytes of content to find obscene or copyrighted material, wouldn’t they also be indexing each video for its contents and relevancy?
If so, this would be a strong reason for marketers to up their rich content game. If not, we still know that the Cloud Video Intelligence platform is primed and ready for action, so it’d be wise to stay prepared for any big changes.
In my opinion, setting a solid foundation will always be hugely important. Ensuring relevant keywords are used in titles, tags and description will always lay a good foundation for being discoverable in search.
But now that Google can transcribe speech and label objects in your video, it needs to be able to connect the dots cleanly. So if you’re making an instructional video on how to make a paper plane, a video selfie of you talking through paper planes won’t cut it. Google would much rather endorse a video that shows a paper plane being built.
If you’re integrating video
content into your marketing plan, the key is to have your video, keywords and
audio all singing the same song. If Google finds they just don’t match, the
odds are that your rankings will suffer.