The ability to measure the success or failure of content has truly advanced in the media landscape. With AI and machine learning, there are now multiple ways in which we can tie what’s happening on screen with the metrics that accurately reflect engagement and success.
Netflix has been the master of creating micro-genres of its content (almost 30,000) and defining micro-demographics (approximately 2,000) to fine-tune its engagement with viewers which, with so many ways to get to content, will become more and more paramount. Using new tools and approaches to engage audiences must become standard for studios, telecom companies, cable operators (distributors) and broadcasters alike.
These tools can now be applied across the board for all content creators. Netflix (and Hulu and Amazon) may have the early advantage, but more traditional content creators are seeing how AI and machine learning can offer them a chance to catch up.
The true need for this extends far past the question of how many people watched a piece of content. Terms like “Sticky,” “Shareable” and “Bingeable” get creators much closer to a sense of what is working (or isn’t).
There are three steps in applying advanced audience analytics to content:
Add deep semantic tagging (video annotation) to content
Add performance (ratings, downloads, engagement) measures
Tie to internal datasets for relevance and actionable outcomes
Why is deep semantic tagging the first step? Because it’s the fastest way to get insights.
First, you have to know what was happening at Minute 22 inside of your content. This does not mean just the objects that can be identified but what is actually happening, on a level that accurately reflects how viewers are engaging with what is on the screen.
Tagging content on a frame-by-frame basis that goes beyond the superficial demonstrates what audiences are truly interested in or have no connection with. Tying what can be called “audience listening” to what viewers see allows for reliable measurements of success when it comes to the major elements of content, such as storylines, themes and character resonance.
This, of course, is one part of a holistic approach to audience engagement that also includes linear ratings and digital downloads that offer a more robust look at both how many people watched and how they reacted (see our article on Content Power for more detail about this). Beyond the pure total viewing number, audiences can then be micro-segmented and targeted with more content which, that’s right, is what Netflix does now.
Several examples of outcomes that we have observed over time include:
- How many cliffhangers are too many in a season?
- What supporting actors are contributing to engagement?
- What story lines are resonating with audiences?
- What locations draw the most conversation?
- What available content will be the most interesting to audiences?
More time spent with content is the cornerstone of surviving in this new landscape. And understanding what happened at Minute 22 will drive smarter adoption of content, reduce risk, guide greenlighting and grow revenues.