Why ‘TV Everywhere’ Needs Content Analytics

Many video providers are exploring Internet content delivery as an alternative to or complement to their traditional modes of delivery. For them, “TV Everywhere” provides an opportunity to capture audience share through other devices from other places and at other times than traditional TV viewing.

How can operators make sure they monetize the investments required to implement TV Everywhere? What will they use to demonstrate viewer engagement to their advertisers? Reliable and complete analytics are not just useful; they will be critical to the success of TV Everywhere.

Audience Metrics

The key measure operators need to understand is audience size: the number of peak concurrent and total unique viewers broken down by content. However, there are substantial challenges in getting these measures and in breaking them down to draw informative conclusions.

First, many of the newer video streaming technologies, especially those used for delivery to iPads and other mobile devices, break video streams into 10-second chunks and encode each chunk at multiple bitrates. With this, the player can adapt to changes in network speed or device load by dropping down to a lower bitrate chunk or maximize viewing quality by switching to higher bitrate chunks.

That works beautifully in some cases but, for traditional video delivery reporting, it creates a nightmare of data. A single half-hour episode of video could result in the delivery of 1,000 unique files. Top 10 or even top 100 lists by filename don’t provide any insight. Instead, analytics packages must put these files into meaningful collections and report by these groups.

Furthermore, typical video log files record usage by IP address. Multiple users may sit behind a single public IP address. A truer measure of unique viewership requires the tracking individual player devices in addition to the IP address. Here are more examples of the benefits of high-quality analytics:

Analytics can provide precise counts of ad views. Player logic can be managed by the publisher to enforce ad views, whether pre-roll or interstitial.

Despite the complications, analytics for Internet delivery have the potential to provide a much richer picture of audience engagement than typical broadcast measures.

Sophisticated analytics can provide measures of viewing time for any content source by minute or even second. They can provide measures of audience by geography, by country and state or Nielsen DMA metro. They can show how viewers are connecting, whether by mobile network or through broadband connection. Analytics also can provide some or all of this data in real time so that program managers know immediately where their audience is.

Measuring Quality

As a premium offering by broadband providers, TV Everywhere must offer a high-quality viewing experience while ensuring that quality when many parts of the delivery chain may not be directly under their control.

Key metrics of quality include delay-to-play start time, play start failures and counts of re-buffering – where the video pauses or stutters because the network or player can’t keep up. Tracking these events by content, by client “eyeball” network, by the Content Delivery Network (CDN) partner or by geography – all possible via analytics – can be used to manage and improve service quality.

Nielsen Ratings?

Reporting and analytics for broadcast television is dominated by a single set of metrics from a single provider. This is sufficient for a single business model – ad sales – supporting a consistent user experience – linear TV. However, TV Everywhere will evolve with a multitude of user experiences and business models. One size will not fit all, and getting a handle on unique viewers and experience quality will call for robust and flexible measurement tools.

jae H. Roh is senior director/Product Management overseeing online tools at Level 3. Contact him at [email protected].

Real-World Situations

>> Premium Sports A major professional league that has retained its Internet delivery rights relies on real-time reporting of viewers. Live game video is available on a subscription basis for the season on the PC, the iPad and the iPhone. Managers receive e-mail updates every 15 minutes of peak viewership and delivery volumes during live game simulcasts. The next day, they use a custom desktop app provided to report viewers and viewing time by game across multiple devices and multiple delivery technologies to view trends across the season. The league also uses metrics of delivery quality to manage CDN delivery, allocating the majority of their traffic to their network partner demonstrating the highest-quality viewer experience.

>> Live Simulcast A major satellite broadcaster simulcasts several dozen channels over the Internet to PCs, mobile devices and set-top boxes. It also uses Content Analytics reporting. In addition to audience counts by channel, it can break out viewership by video quality level to gauge its investment in encoding formats, and bitrate by device type. The broadcaster can monitor enforcement of geography-based rights restrictions on content distribution, for which there also is a sophisticated slate/blackout management system. Aggregate traffic statistics showing net flows of IP traffic help it manage relationships with other operators.