This is my first article (of hopefully many) for Search Engine Watch focusing on some technical video SEO topics. Today we'll look at the role of metadata in optimizing video for YouTube based upon my experience in developing a YouTube Channel in July 2007. We must have done a few things right because 21 million downloads later the work I did is still ranking in the SERPs.
We'll use the YouTube Creator Playbook in these posts as you can't get much better info than when it comes straight from the horse's mouth, so to speak. Metadata is one of the bigger topics in the playbook; overall video optimization seems to be less important than the programming/production and more important than the social/community topics in the playbook.
I'll also introduce some terminology we'll use going forward. I use "audience approval" and "relevance to query" for YouTube in the way that on-page and off-page factors are used in discussion of conventional SEO. T
Both conventional and video search have factors in the ranking algorithm that require a totally different skill set. Any strategy for optimization must account for this difference in skill set, or it can get too involved with elements that don't push rank.
It is very much like conventional SEO in that some elements of the algorithm are very hard for SEOs to affect in a cost effective/non spammy way (i.e., user behavior metrics on YouTube and link/authority in conventional search engines).
From the early days (for me) of 2007, two big factors at work in YouTube search quality have remained true:
- Audience approval. YouTube rankings seem weighted to reflect audience approval and therefore the user signals from plays and viewing time and to a lesser extent playlists are where the ability to really move rankings exists. The community factors of comments, subscribers, friends and channel activity seem to be weighted but again these are pretty hard signals to isolate for a good idea of their true value in ranking.
- Relevance to query. Metadata is exceptionally important because it pushes user metrics as well as relevance because of the role that related videos play in the viewing behavior of users. Metadata play a role in relevance to the query/search algorithm and tags in particular are tied to the related video results.
Relevance isn't as big of a contributing factor in determining rankings as in conventional textual search.
One of the advantages I had as an early adopter of SEO was I had worked with no tools and no keyword or traffic data so back in '07 I was comfortable working with other data and applying it to video. I took what I knew from the SEO and PPC campaigns and to a lesser extent the site search data and applied it to the metadata for the YouTube videos.
We knew from a marketing standpoint our best audience searched for the model and brand so we targeted these in the titles and metadata. I think you can take some of the data and marketing intelligence gained from conventional campaigns as a kind of jumping off point for using the keyword research and discovery tools now available for a video campaign.
YouTube titles are a little trickier than a conventional search engine with shorter length and "click worthy" CTA's stating the purpose/goal of the video needing special attention because they get more “action” from users who are more likely to watch the complete video.
YouTube titles are similar to conventional titles in that you want to position primary keywords at the beginning of the title with compelling use of phrases and branding at the end of the title. Titles aren't as big a factor in ranking on YouTube as they are in conventional search.
The description is very different on YouTube from a conventional search engine in that it can be much longer and contain links to other pages or video. YouTube descriptions are useful for not only ranking but also providing lots of opportunity for cross promotion of content using links and promotional elements in the video.
The YouTube playbook advises putting the most “compelling” information at the beginning, using keywords and including show and channel information.
Check the tags on popular videos in a query result and you will greatly enhance your chances of entry into the related videos results. The related videos increase plays which directly affects factors weighted heavier for ranking.
That is also why the YouTube "influence" in Google universal search, which is thought by many to be a little too biased is entirely within reason. The other video "sources" don't provide the user data and quite frankly user data not relevance is the holy grail of video search because quality is determined by entertainment value and less on the ability to fulfill a task.
Closing Thoughts On Optimizing Metadata
So, there we have it! The beginnings of a strategy to optimize video metadata.
It is always imprtant to understand the role of various elements in the ranking algorithm. Understanding what and the degree of influence different elements of the ranking algorithm possess is the basis for all optimization strategy. By understanding each element's role in the ranking algorithm, you can use your time more efficiently and produce better results.