Measuring Success: The Multiplatform Analytics Challenge
Most streaming industry discussions encompass three main workflow steps— acquisition, content preparation, and delivery—and for good reason. Whether it’s a discussion around cameras, video mixers, encoders, transcoders, or content delivery network (CDN) infrastructures, we as an industry seem most comfortable solving technical issues.
There’s one workflow step, though, that we tend to be less comfortable with: measuring the success of streaming.
Part of this may well be entrenched in the streaming industry’s DNA from the early days. Back in the late 1990s and early 2000s, almost every year brought a claim that streaming viewership would soon surpass the primetime numbers of traditional over-the-air (OTA) broadcast television or even cable’s 500-channel universe. But when measurements were applied against the hype of the early streaming industry, the numbers didn’t add up, leading to an implosion of cost per mille (CPM) rates and a subsequent mass defection of advertisers for almost a decade.
The days of reaching levels close to television audiences have finally arrived, with the Rio Olympics representing another pencil mark in the adolescent measurement of streaming’s growth.
Within the first 5 days of Olympic coverage, NBC had racked up more than 1 billion minutes of streaming, surpassing in only 5 days the 818 million minutes of streaming delivered for the entire 2 weeks of the London Olympics in 2012. Impressive numbers, for sure, with the opening ceremony alone racking up approximately 42 million streaming minutes. Numbers like that indicated that streaming had finally reached a milestone: matching viewership for an average cable TV prime-time show. Still, streaming only represented about 8 percent of the total Rio Olympic viewership when OTA and cable broadcasts were added in.
Another reason that analytics seems to be a third-rail subject in the streaming industry is the fact that we still have fairly rudimentary approaches to measurement, with the vast majority of measurements focused around file-based assets used for on-demand content delivered via web browsers or over-the-top (OTT) devices.
When we discuss analytics, most industry professionals approach measurements from the standpoint of a single video on a single platform. This 1:1 approach used to make sense, since desktop viewing long dominated both on-demand and live streaming.
Recent growth in mobile delivery, though, demands a radical rethinking of analytics. Not only do we need to rethink how we analyze video-centric delivery across multiple platforms, but also how we think about fundamental measurements around unique viewers and event-based analytics.
Only recently have we begun to analyze multiplatform streams in a holistic way, and even that approach is mired in a video-centric analysis. To move beyond the quagmire of analytics data and toward an actionable multiplatform measurement methodology requires looking beyond the streaming industry itself.
Understanding the Language of Analytics
The first step in understanding multiplatform analytics may surprise you: Learn a new language. In fact, we need to learn a language spoken less by technical engineers and more by traditional online marketers.
The need to learn a new language may seem daunting or even unnecessary, but remember back to when you first got into streaming. In the streaming industry, we have a language all our own. Some of our terms, such as transcoding, don’t translate into other industries.
In fact, our language is distinct enough that Streaming Media provides a glossary of terms to allow beginners and experts alike to get on the same page when it comes to what can be a highly technical language. Some of the terms have even shifted over the few short years that we’ve had the glossary, requiring updates from time to time.
However, while we might talk about devices, events, platforms, and users as familiar terms for streaming, in the broader analytics sense, some of the terms have a different meaning altogether.
For instance, an event in streaming often means streaming a live event to paying customers (think of it as almost a synonym for pay-per-view streaming). But in the language of analytics, an event is defined as “any action performed by or associated with a user,” according to Amplitude.
An event could be as basic as filling out a free-trial form to view the latest piece of premium content, or as complex as a user’s viewing patterns when watching a single piece of long-form, on-demand content in parts across multiple devices.
The distinctions between analytics and streaming terminologies are key. A mastery of both languages is an essential step in setting goals to effectively measure multiplatform streaming, since the language and terms needed to report effectiveness must blend the two terminologies without muddying the waters.
Setting Goals
Once you understand the lingo, it’s time to set goals to measure success. But remember that goals will ultimately fall into one of three measurement areas: events, users, or videos. Before we look at those, remember to sidestep the “build it and they will come” trap that video content often falls into when it is solely distributed online.
Distilled, a media consultancy focused on organic search, analytics, and video campaigns, has one of the best quotes I’ve seen on combatting that mentality, which our industry unfortunately often seems to live by.
“‘Going viral’ isn’t a business goal, neither is having a million video views,” writes Phil Nottingham. “Views and virality are simply a means to an end, namely the right kind of exposure to the right kind of people.”
Even exposure itself, though, isn’t the end goal when it comes to the business aspects of a Vimeo or YouTube strategy. The end goal should be conversion and retention. Nottingham expands his initial comments, noting that even the savviest online marketer should think beyond exposure and toward financial profitability.
“Business goals should be measurable, concrete and (even if several steps removed) tied to financial success,” writes Nottingham.
Measuring Videos
As mentioned earlier, the traditional streaming industry approach to multiplatform analytics has been to measure a video across platforms.
The reasoning is solid if all videos are created equal, or if a content owner focuses on only one type of online video asset.
For example, think of a news website. If you’re a news organization and every day you offer a series of 60-second clips covering the news of the day, measuring video views might be an effective way to analyze audience interest in traditional news topics.
But most content owners have a variety of content they want to distribute online, via OTT or specific online video platforms. And if there’s a disparity in video types, length, or even content, an apples-to-apples measurement of disparate video types may not only be ineffective but also misleading.
One way to move beyond simple video viewing measurements, whether on a single platform or multiple devices and platforms, is to look at what happens once a video ends. Companies such as Tubular Insights highlight this type of measurement as one of their unique selling points, offering insight into viewers’ organic actions at the end of a clip, an episode, or even a full-length movie.
Measuring the organic behaviors can lead to more finely tuned calls to action. In the past few years, we’ve all seen a rise in the number of YouTube channels that end a video with the video’s stars suggesting other videos to watch, often providing a hotspot link within the last 10 seconds of the video that a viewer can click to take advantage of the call to action.
The approach that Tubular takes, offering insights into both viewership and subsequent actions, is grounded in traditional web analytics, which often come in the form of event-based analytics.
Measuring Events
What to measure and how to measure are the two main questions that any marketing or engineering type typically asks when it comes to analytics. These two questions, though, become slightly less important when it comes to multiplatform analytics. The overarching question becomes less of what and how and more of when— when to measure on a given device or platform, when to correlate the data between platforms, and when to audit the results to confirm that an analytics platform is properly calibrating measurements across apps that reside on all these platforms and devices.
Related Articles
Experts from Akamai, Comcast, Conviva, and Yahoo! discuss key analytics, metrics, and assessment strategies that will help you deliver more effective and profitable live streams.
28 Sep 2016
For two years an error in how Facebook calculates average view time led to dramatically higher averages. The problem shows the need for third-party verification.
23 Sep 2016
Objective quality benchmarks provide exceptionally useful data for any producer wanting to substitute fact for untested opinions.
01 Apr 2016
At long last, broadcasters, content creators, and advertisers will be able to compare a program's views across all platforms.
23 Oct 2015
Also, Adobe announces TBS as a Primetime customer; will drive live linear and VOD content for TBS apps and sites.
24 Sep 2013