Size Crucial For OTT Audience Measurement

As television shifts to multiscreen, on-demand viewing and advanced 1-1 digital-style targeting, advertisers and publishers note the shortcomings of panel-based TV “ratings” to measure over-the-top television streaming (OTT).

This begs the question: Are alternative approaches, like Automatic Content Recognition ( ACR) and in-app measurement, better suited for OTT?

To reliably capture even basic viewing behavior, an OTT barometer needs to work in an environment in which measurement challenges, like content proliferation, audience fragmentation and time-shifting, exist to the extreme.

But OTT stakeholders require more than basic audience metrics.

OTT offers a strong ad environment, combining the premium high-engagement content of TV with the targeting depth and precision of digital. Advertisers and publishers can define specific TV audiences by merging OTT viewer data with purchase behavior, web-browsing history and third-party information.



That moves TV beyond demographic-based contextual buys to advanced people-based targeting.

To power the activation of granular targets stretched across content, devices and time, OTT audience measurement requires massive scale.

Consider the GPS system in today’s cars. Fueled by prodigious amounts of granular data, GPS directs drivers to specific locations with pinpoint accuracy. A GPS system fed by a map of the U.S. interstate system would be of little value in finding precise target destinations.

OTT is similarly dependent on large data sets to execute precision TV targeting at scale. Smaller data sets rely heavily on “lookalike” modeling to achieve scale, replacing the filter of program-based targeting with the fuzz of probabilities.

Additionally, a viable OTT audience measurement solution must provide cross-device measurement.  Publishers need cross-device metrics to understand the size and composition of audiences. Advertisers require cross-device metrics to manage campaign reach, frequency and creative rotation at the consumer-level versus the device-level.

So how do the current OTT measurement options stack up?


Media-measurement panels lack the scale required to capture OTT viewing behavior scattered across myriad platforms and devices. A panel of requisite size would be too expensive to operate.

From an OTT perspective, the strengths of panels – representativeness and single-source measurement – are useful for cross-media-planning. Panel data may also improve the accuracy of probability-based OTT targeting and provide a high-level link to linear TV behaviors.


ACR offers a step up in scale from research panels, with audiences ranging from a few hundred thousand to a few million, depending on the system and use case.

ACR platforms do have scale limitations. ACR technology is installed only on certain brands of smart TVs (about 10% of OTT households) — any ACR platform accesses data from a subset of those. In addition, consumers must opt-in for their viewing data to be captured.

ACR has some notable strengths, including measurement of all ad-supported content on ACR-equipped televisions: OTT and linear. That makes ACR well-suited for certain OTT measurement applications, such as understanding demand for ad-supported OTT programming, gaining insight into cross-channel OTT viewing, and measuring the interplay between OTT and linear TV.

Mobile and desktop viewing are not captured for ACR households, nor is content streamed to non-ACR TVs, which compromises people-based OTT targeting and measurement.


Apps are the OTT equivalent of linear TV “channels.” Because all OTT content is delivered via apps, in-app measurement is inherently census-level, which maximizes scale. It’s why in-app is the standard in mobile measurement.

In-app measurement has historically been labor-intensive to install and maintain. However, easy-to-install software has recently become available, and OTT publishers are now migrating to in-app.

In addition to its scale advantages, publishers retain the rights to in-app data, unlike ACR. In-app also strengthens cross-device metrics by using connected TVs as anchors to tie mobile OTT viewing back to residences and individuals.

Because the install base is still growing, in-app data does not provide a comprehensive view of OTT audiences across publishers, limiting its value from a planning perspective.

Predicting a Winner 

OTT promises to revolutionize TV advertising, due to more precise targeting, more relevant messaging, and more timely campaign management.  Advertisers best able to deliver the right message to the right person at the right time will benefit most as TV migrates to OTT.  

Media companies that enable advertisers to take full advantage of OTT’s digital architecture will benefit disproportionately. Those companies will deliver highly granular audiences with precision, along with timely, detailed analytics. The audience measurement approach that best supports these efforts will become the OTT standard.

3 comments about "Size Crucial For OTT Audience Measurement".
Check to receive email when comments are posted.
  1. Ed Papazian from Media Dynamics Inc, July 26, 2017 at 3:37 p.m.

    David, I'm sure that there are ways to identify what content is appearing on the screens of various devices in minute detail and probably do so with far larger samples than Nielsen uses for national TV ratings---providing the alternative samples are truly representative and you solve the problem of getting all of the devices a household owns measured. But how do you determine who was using the device every time it is employed and how attentive---if at all---this person---or these persons---were, not only to program content but also to the commercials?

  2. Daniel Tjondronegoro from Beatgrid Media B.V., September 4, 2017 at 9:37 a.m.

    Interesting article. The TV landscape has changed. No doubt. Most marketers/CMO’s have cross-platform TV and cross-media measurement in the top 5 of their challenges list. Yes, there are shortcomings with panel-based measurement, but is the OTT landscape already big enough….

    Scale in measurement is essential, but as long as traditional linear TV is still strong, I don’t think we should compare digital (deterministic) OTT measurement directly to the traditional panel-based TV measurement approach.

    At Beatgrid Media, we are a small but innovative startup, and we’ve built one of the best smartphone-based TV audience measurement solutions in the industry. Cross-platform video measurement, at the respondent-level, is finally scalable. This is why more innovative brands leverage Beatgrid’s innovation. But… the big question is; are media buyers ready to work with separate data sets for 1) currency for media buying and 2) Newer (respondent-level) cross-platform effectiveness metrics to optimize ROAS?

  3. Ed Papazian from Media Dynamics, September 5, 2017 at 9:10 a.m.

    Daniel, I agree with you. Moreover, we should take notice of the fact that, at present, OTT amounts to about 10-12% of the average person's total TV/video usage time and, of this the largest share does not carry ads. With upwards of 40-50 OTT channels competing for a rather small piece of the viewing pie, it may be premature to demand more than Nielsen can provide---when it finally gets around to it. Who's going to pay for huge panels to measure  what are very small ad-supported audiences?For more on how OTT viewing is split among the major contenders----Netflix, Amazon, Youtube and Hulu--- as well as its demographics we have a brand new special report on the subject available at a very low price.

Next story loading loading..