Analytics are important for understanding what is going across a network, however, it’s impossible to sift through a multitude of data points and detect issues across ad and content performance. This is especially true with publisher platforms running on increasingly large tech stacks.
For a publisher to scale their streaming video business they need instant insight into the root cause of issues, a guided course of action, and triggers to get fixes in place as soon as possible. Without automation tools, time is spent trying to discover the root cause of issues rather than providing a great streaming experience. When experience is impacted, likelihood of viewer churn increases.
Conviva’s solutions have always gone beyond simple data points, to give publishers actionable insights. Our priority is to continue to build automated tools that streamline the work needed to improve video performance. We’ve found that automation is particularly important on the ads side of streaming video. The connected TV audience is more engaged with video ads than audiences on any other platform, which is why CPMs (cost per thousand impressions) have continued to rise for these platforms. Detecting issues within ad breaks not only helps to mitigate viewer drop off but ensures publishers are optimally monetizing this valuable inventory.
Conviva’s machine learning algorithm continuously monitors leading indicator metrics such as ad errors, ad slates and creative bitrates to automatically find and alert on ad issues at an individual ad creative level. The algorithm triggers alerts whenever anomalies in the data are detected. Issues detected by Conviva’s AI Alerting system include bad creatives that aren’t delivering properly, high levels of ad slate that indicate poor fill rates, and trafficking errors that resulted in creatives serving on unsupported endpoints.
Alerts can be routed into other workflows such as Webhooks or PagerDuty within the user interface to accelerate the time to resolution. Looking ahead, alerts will be able to plug directly into some ad servers to make monitoring creatives during live campaigns a seamless effort.
The shift in analytics from strictly data points, to guided data, to data that triggers action is important to note. Getting bogged down in data points keeps businesses from acting fast. As the streaming industry competition continues to heat up, the ability to make quick data-driven decisions will only become more important.
Examples:
How the real-time Ad AI Alert can be automated:
Alerts provide second-by-second reporting:
The above example shows an auto ad creative that has consistently high buffering. The high buffering not only impacts viewer engagement on the video but also monetization for high CPM auto campaigns. Ad AI Alerts automatically identified where the issue is occurring whether the line item (ad ID), creative ID, ad server (ad system), or video player.