Back to Blog

How to Measure Video Engagement Metrics With a Proactive Monitoring Strategy

laptop showing video engagement metrics

When viewers have an issue with performance, quality, or availability, the results can be catastrophic to the business. Not only do they watch less content, but they also abandon more content and even cancel their subscriptions. For OTT services that rely on advertising, that’s detrimental to the bottom line. The same holds true for SVOD, as low video engagement metrics generally increase subscriber churn rate. 

The prevailing solution to improve viewer engagement is back-tracking: an issue is discovered and operations spend the next hours or days sifting through log data to find out what the problem is. However, this isn’t a very effective way to mitigate the negative impacts of reduced video engagement. Sure, you will resolve an issue, but it doesn’t fix any of the unhappy customers who have already experienced it and eventually ended up abandoning the content. The key is to monitor the video experience proactively by identifying problems before they reach the customer and implementing processes to prevent those errors in the future.

Which video engagement metrics reveal the state of the video experience?

Most OTT operators measure the video experience through metrics gathered from the player:

  • Time spent watching each video
  • Number of videos watched over a given time period 
  • Amount of time spent within the OTT platform
  • Average bitrate per title
  • Time to first byte

These metrics reveal a lot about what the user is experiencing within the player. For example, how long it takes for the video to start (and what the user does while waiting for it); how many bitrate adjustments happen and the time between each request; and, of course, how many buffering events occurred in each video session. These metrics, though, disclose more than just individual session issues. When aggregated and extrapolated across regions and demographics, they expose behaviours about content such as time spent on specific titles, time spent watching related titles, and more. 

Yet these metrics don’t provide the entire video engagement picture. Yes, the endpoint player is a critical part of the video experience, but there are other metrics, gathered upstream from throughout the workflow, which provide an in-depth understanding of video engagement. For example, collecting metrics from encoders, CDN caches, and network infrastructure alongside those metrics from the player will help NOC operations teams uncover why the video experience is less-than-expected for individual playback sessions or even entire geographies.

👉 You may like: How to Fix the Root Cause Analysis Problem in Streaming Video

Why more upstream monitoring is critical to improving video engagement metrics

Although streaming operations can improve video engagement metrics reactively by examining issues after they have happened in the player, they can also address video engagement issues proactively by looking at what’s happening further upstream. Rather than working backward from the player, operations can work forwards from within the workflow to ensure the metrics are within acceptable ranges. 

This upstream monitoring, though, can’t simply be random quality or performance checks. To truly take control of video engagement, you’ll need to implement a preventive system: a way to identify the issues as they happen and address them immediately, even before viewers experience the problem. That kind of system requires monitoring tools that provide a real-time, consistent picture of what’s happening throughout all the workflow components. Sure, player analytics are important because they reveal what the viewer actually experiences. However, player analytics don’t provide actionable insights to resolve the issues happening further up the workflow.

Proactive monitoring is key to preventing potential issues from affecting the viewer

When the insights derived from such a system are combined with visual indicators based on acceptable thresholds (imagine a dashboard that shows red, yellow, and green levels for key metrics up and down the workflow), operations engineers will be able to identify whether a metric is good or bad and how it impacts the player-level video engagement metrics. In this scenario, time that was traditionally spent triaging issues identified in the player video engagement metrics can now be spent resolving the underlying problems upstream (which would ultimately cause the problems in the player). 

Enabling a preventive monitoring solution to assure the highest levels of video engagement

Video engagement drives every OTT business, and the metrics that reveal the state of the viewing experience–which, in turn, tell you how engaged your viewers are–come from throughout the workflow. However, responding to the issues revealed by those metrics can’t be reactive. To stave off subscriber or viewer churn, you need to implement a preventive monitoring solution: a system to identify, triage, and resolve issues before they reach the end-user.

Synthetic probes are at the heart of the monitoring tools inside such a preventive system. These probes, simulating user requests, force those workflow components to generate the data, which combines to form the metrics and informs you about what video engagement might be. Think about it this way: if everything is working as designed, you will be able to predict video engagement based on what’s been revealed in your preventive system. 

Those probes, though, must be deployed across your entire workflow. You’ll not only need them within your own network but also within public clouds and even within partner networks (like a CDN). Managing your preventive system will then involve maintaining and scaling the probes to simulate what happens when your concurrency rate grows. If you are only measuring on historical access data, you’ll never be able to understand what happens when your scale expands because of a hot content title. A quality or performance metric, for example, may look one way under a certain load and entirely different under another. This is why preventive monitoring for video engagement is so important: there may be variables, such as scale, which you aren’t considering as a critical KPI, but that could have a dramatic impact on video engagement. A preventive monitoring system “prevents” those variables from negatively impacting these metrics by monitoring their eventuality. If you are operating without a preventive monitoring system, you’ll never see the impact of that KPI until it has already hampered the player-side video engagement data.

👉 You may like: The Complete Guide to Live Stream Monitoring

Get out in front of video engagement or you’ll get left behind

Video engagement is impacted by a lot of variables. From performance to quality to content type, viewers have many reasons for watching and spending time on your platform. The problem is that understanding video engagement is not cut-and-dry. It is a combination of server-side and client-side video engagement metrics, which are related through other vectors like scale and device. Remember that viewers may watch content more on their phones than on a big screen, or vice-versa. This decision may be a factor of performance and quality or just user behaviour.

That’s why it’s so important to gather as much data as possible throughout your entire streaming video technology stack. Correlated together, video engagement metrics from the player and other workflow components will provide valuable information about why viewers are watching, what they are watching, and how to keep them engaged. If you wait to act until after video engagement has dropped, you’re bound to negatively impact business metrics like the number of subscribers or ad impressions per viewer. 

To ensure video engagement stays within acceptable thresholds, you need to be proactive rather than reactive and implement a preventive monitoring approach. With synthetic users testing upstream workflow components and monitoring tools with clear visual indicators, you’ll be able to identify and address issues before those player-based video engagement metrics take a hit. As the old saying goes, “An ounce of prevention is worth a pound of cure.” Maybe in the world of streaming video, that translates to, “an ounce of prevention is worth a few minutes of MTTD.”

Want to know more about how to proactively monitor streaming data with Touchstream? Contact us here.