True OTT Ad Monitoring Requires Real Testing Devices

The Rise of Streaming Ads

The hottest trend in streaming video is one that’s been around since the advent of television: advertisements. After Netflix’s subscription-based model took the industry by storm in the 2010s, its recent decision to offer a cheaper, ad-supported plan came as a major surprise to most – but it was less of a surprise to those who have been paying attention to shifting attitudes in OTT video.

Ad-based streaming services like Freevee and Pluto TV offer similarly robust content libraries to Netflix without any admission fees. Instead, they depend entirely on revenue from ads that run throughout their content. FAST (free ad-based streaming television) channels do the same, but run their programming on a set schedule as a throwback to old school linear television.

A 2022 Comscore study found that American households were adopting advertising-based video on demand (AVOD) streaming services at a faster rate than subscription-based video on demand. A year later, Netflix created its own ad-supported subscription tier, as have fellow top streamers Max, Disney+, Peacock, and Hulu. Unfortunately, running ads in streaming video isn’t as simple as it was on cable television.

How Streaming Ads Work

Instead of being pre-packaged together, streaming ads are “dynamically inserted” into video content. They come from a different origin server than the core content being watched. This content is pulled from its original source into an online stream and then added to a video player. The video player inserts the ads in real time as viewers watch.

The most common form of dynamic ad insertion uses SCTE 35 ad cues. These cues are programmed into the content and trigger “ad pods” (or ad breaks) when the video player reaches them. The cues determine how long each ad pod runs and how many individual ads it contains. From there, the individual ads are selected (often by relevance to the viewer’s location) and pulled into the stream from separate online sources. It’s a small technological miracle that dynamic ad insertion (DAI) works at all – when it does work.

Common Issues with Streaming Ads

Streaming ad issues are so commonplace that they should be instantly familiar to anyone reading this article. Here are some of the top offenders: significant buffering before and/or during an ad; ads playing with a completely different picture quality or volume level than the surrounding content; ads repeating in every single ad break; and ads being triggered too early or too late or just being the completely wrong duration for the pod they’re filling.

These issues are elevated for free ad-supported streaming television (FAST) channels, that have a delivery path even more complex than AVOD services. FAST ad pods can feature slates instead of ads, which sometimes count down to the content returning and sometimes just present a confusing blank screen. The slates are also prone to crashing the stream. In Witbe’s own research, we recorded 30% of all FAST viewing sessions encountered an issue with their dynamically inserted ads.

The Importance of Resolving Streaming Ad Issues

Solving these issues is of vital importance to video service providers. Many AVOD and FAST providers rely primarily on revenue from advertisers purchasing ads. When advertisers have reason to believe the ads they purchased were affected by issues and did not reach their intended audience, they may withhold payment, request more ads be aired for free, or quit working with a provider altogether. Viewers who encounter ad issues might avoid the service in the future, losing ad impressions and leading to negative word of mouth.

So, what’s the solution? As any developer can attest to, there’s no single mistake causing all of these errors that can easily be resolved. Instead, there’s a single way to identify errors so that they can be fixed: ad monitoring.

Ad Monitoring with Real Devices

Errors cannot be fixed until they can be detected. For service providers, this makes reliable ad monitoring essential. Since the ad insertion is dynamic in nature, the errors themselves often are too: they will not appear the same in all circumstances or to all users. For this reason, ad monitoring should be performed on real consumer devices. Using the same apps and networks that viewers use means that teams can see the same issues that affect performance for their actual users.

For testing at scale, automation is an invaluable tool for providers. From there, ad performance should be evaluated by specific criteria. Teams monitoring dynamically inserted ads should monitor the duration of the ads, the amount of buffering involved, the video and audio quality of the ad compared to the content, and the appearance of any blank slates or stream crashes. The data from the monitoring sessions needs to be preserved for observing progress, and individual ad performance should be recorded for verification with advertisers.

The Way Forward for Streaming Ads

Running ads is a time-honoured revenue model for video services. Its recent resurgence is no surprise – it has sustained linear TV for almost a hundred years, so it has plenty to offer streaming video. Unfortunately, dynamically inserted ads often come with performance issues that can affect revenue, viewership, and brand reputation.

In order to improve ad quality, ad monitoring is a must. Monitoring on real devices leads to an accurate view of performance that can then be fine-tuned. For streaming video providers who rely on dynamically inserted ads and want to deliver a superior viewing experience, ad monitoring on the actual devices that are being used to watch streaming video is critical.