Digging Into Kaltura’s New Video Analytics

Yael Dovrat
Updated September 10 2019
The philosophy behind Kaltura's video analytics
Yael Dovrat
Updated September 10 2019

As video usage increases across organizations so does video data, how do we make sense of all this new information? With the right tools, we can now find answers to questions we never thought of asking before. Kaltura always provided analytics, but is now defining analytics as a key focus area and introducing new tools and capabilities around video analytics. With our new focus we defined 3 key pillars to guide us through this journey: “Get more data than ever before”, “gain actionable insights” and “see the full picture”. I’d like to share more about how these pillars come into play.

Starting at the Foundation of Video Analytics

Our renewed journey started at the base, the technology. We moved to a Druid database that is designed for analytics with low latency streaming ingest and time-based partitioning, which enables performant time-based queries and fast search and filter. We also changed how events are sent, stored and enriched using various technologies.

We increased the granularity in which the data is stored and how frequently it updates.

We made this investment in technology to allow us to continue to grow and collect more data points with higher granularity, without compromising performance and speed.

Data as a Building Block

Everything starts with quality data. We can create the prettiest-looking graphs but without quality data you won’t gain much insight. Usually video analytics show you the basics, like the number of plays. However, we believe there is so much more and looked beyond the plays. We wanted data on how the video is being created, how viewers are interacting with the video and how behaviors change over time. We also added many dimensions to allow the data to be sliced and diced, for example by geo location, devices that were used and more. You can now investigate your data with questions like what operating systems do I need to support? Where are my users watching my content? We want to enable our customers to dig deep into the data. We will continue to expand and add data points as we grow our platform.

Informed Decisions Based on Actionable Insights

How do we translate all this data into action? We are creating tools that will help identify what are the most important questions that aren’t being asked yet. We want to help our customers optimize their content and make decisions based on meaningful insights. One example from our recently launched KMC analytics is the Engagement score, how do you rank a top video? Is it by number of plays? How many viewers watched it? How much of the video was watched? We created a unique Kaltura Engagement score for videos combining several key metrics to help you identify what are the top performing videos in your organization and compare between videos over time.

We provide analytics from the account level with aggregate data all the way to a single video with a heatmap per user to better understand which parts of the video are actually being watched and which are skipped.

Identifying trends is another important aspect which we allow to do with the compare mode in KMC, which allows you to compare the entire dashboard between two time periods of your choosing. This visualizes trends and changes over time and enables comparing with seasonality in mind. With these tools you can transform your dashboard and analyze changes to viewer engagement and adoption of capabilities.

Bigger Picture Video Analytics

Video is everywhere, and better understanding video data is key. However, we understand that this is just one data point in a complex environment. With Kaltura’s open standard approach, we continue this with support of various analytics APIs such as Caliper and xAPI which allow you to easily connect your video data with your existing analytics systems and software and view your video analytics in a larger context. For example, connect your video data to your LMS data and use this information to track learning engagement.

It’s been an exciting journey so far, but we are just getting started! As we continue to grow our platform and add more capabilities, we continue to focus on these video analytics pillars to guide us. We will continue to expose analytics in different places and applications and use the collected data to create better experiences. With the new video analytics tools both users and admins can better understand their content and drive action.