Mining Value From Seconds-Based TV Measurement

The expansion of premium video content platforms has publishers, advertisers and media agencies scrambling to navigate differences in reported audience metrics from multiple sources.

Programmers monitor content viewing across linear, data-driven linear, video on demand (VOD) and connected TV (CTV) with an eye toward maximizing profitability from advertising and subscriber revenue.

Marketers and their agencies, meanwhile, seek to reach their most desirable consumer targets by threading a needle across this quilt of diverse premium video conduits. And in the background, the deluge of viewing and consumer data during the past several years has stressed legacy ad operations systems for invoicing, bill paying, taxonomy administration and post campaign audience reporting.  

Later this year, another development will add to the combustible state of TV measurement and ad operations. In September, starting with the 2024-25 TV season, the industry is slated to adopt a more granular advertising currency that credits commercial viewer counts based on exact seconds vs. the traditional average commercial minute (ACM) audience within a program.



To that end, Nielsen will be providing clients with second-by-second reporting with the option for advertisers/agencies to maintain C3/C7 ACM metrics. During the past few years, second-level measurement has gained traction with publishers and media agencies executing TV buys using data sourced from alt currency firms including VideoAmp, and Comscore.   

The granularity of second-by-second viewer counts holds the promise of more precise audience measurement, providing transparency to report viewer counts to individual commercials vs. the average of all ads in a show.

But overlaying this finer measurement atop the rapid growth of advanced targeting data and addressable ad placement introduces more complexity for both video ad buyer and seller, in terms of pricing, reporting and processing. Let’s take a closer look. 

Parsing Seconds For Max Viewers

The move to second-by-second currency means that one ACM viewer count for each show will now be replaced by a multitude of individual commercial audience estimates.

For example, viewer counts for 42 individual ads would be tallied during a during a typical 60-minute national linear TV program. And viewing may vary significantly during this time, with a tendency for audiences to build momentum toward the latter part of a program that is viewed live, for example.

It makes intuitive sense that commercial pricing could toggle up or down based on audience movement and turnover throughout a show. But this dynamic is not a forgone conclusion, according to Ed Gaffney, Edge Consulting. 

Gaffney contends that cherry-picking commercials when viewing is elevated would result in CPM premiums for the most desirable spots. Conversely, he says that “for most brands any differences in viewing counts within the same program would wash out with random rotation of commercials within a show.”  

Defining impressions or views

A fundamental issue posed by second-based measurement is how to value commercials when an ad appears on screen for only four seconds out of a 15, for example. This can happen when viewers are changing channels and implies that impressions or views should be weighted to reflect their on-screen duration.

Hypothetically, advertisers could request a CPM discount off the full cost of an average commercial unit. YouTube, for example, charges lower CPMs for skippable video ads versus those that are “forced view.” There can also be a disconnect on how many seconds to credit unduplicated audience reach versus an impression or view.  Is it one second? Two seconds? 

Commercial pod position research has shown that ads appearing at the start of a commercial break receive more attention than ads on screen thereafter. Also known as “A” positions, these first-up placements are preferred by advertisers, who often pay CPM premiums to secure them. 

A 2022 TVision study (see below) demonstrated that viewers pay more attention to first-in-pod A position ads based on recorded face-to-screen metrics. Reporting of viewer counts at the second-level will reveal any audience variations across ads within a commercial pod, and when combined with attention-level considerations, could potentially impact pricing by pod position.  

Acceleration of audience-based placement?

As with digital ad placement, audience-based deals are designed to deliver impressions anywhere target consumers are available regardless of daypart, time period, program or platforms.

Mike McGuire, senior vice president-media at MSA (Management Science Associates), says “this approach is more akin to the digital model of ad transactions, although qualitative and strategic factors such as program viewer affinity and target reach goals will remain an important consideration in scheduling TV ads.”

Hypothetically, second-by-second measurement will afford both buyer and seller more flexibility to move closer to audience-based pricing models.  But publisher-side linear ad-serving systems are limited in their ability to execute schedules with any advanced degree of precision.

Howard Shimmel, president, of media advisory Janus Strategy and Insights, says: “Log optimization at a typical TV network is not highly automated posing challenges for blocking off specific minutes for data-driven linear (DDL) targets, for example."

Systems Upgrades Are Key

The path forward for second-by-second currency will likely be one of gradual immersion rather than a disruptive flip of a switch. Embedded industry practices and underdeveloped legacy operational systems for planning and executing TV ad placements underly the inertia for change.

In the near term, Nielsen’s ACM estimates are apt to maintain their position as the industry’s core currency for a number of reasons:

  • Systems needed for managing seconds Any potential benefits that seconds-based measurement holds for more effective insights and ad placements are trumped by the lack of sophisticated systems for extracting this value from the more granular data.  Second-based analysis of pod position, partial ad displays and delivery of advanced targets through DDL, for example, require massive upgrades in systems and analytic capabilities.  Then there may be challenges to scheduling linear ads based on more precise placement specs.  There’s also the potential issues on invoicing and billing to consider, given the fact that advanced target transactions are already stressing the current systems for execution and payment processing.  With the expansion of reporting audience viewing during the exact times on screen, publishers will also need to enhance the sophistication of their yield optimization systems, including at the log level.  Meanwhile media agencies are faced with evaluating and planning on the more granular data across multiple publishers and will look to cherry pick the most widely viewed air times within TV programs.  

  • Legacy ecosystem and operations Nielsen’s age-and-gender ACM ratings served as the industry’s currency for decades and are ensconced in financial and bill-paying systems which are already under stress to accommodate advanced target audience estimates.  Updating these systems is ongoing and impacts publishers, media agencies and advertisers’ procurement departments. 

  • Historical trend data Nielsen’s historical linear viewing data has been used for forecasting future linear TV audiences sold in advance during the annual upfront marketplace.  A problem emerges with constructing similar historical data sets for streaming viewing, given its relatively nascent audience measurement capability and the varying data sources.  

  • Standardization The ad industry will continue to work on streamlining audience measurement systems on an ongoing basis. CIMM (Coalition of Innovative Media Measurement) and its TV, advertiser and media agency members have developed blueprints for improving TV measurement while keeping pace with the constant stream of changes occurring in the research marketplace.  These initiatives are valuable but need to be paired with upgrading or even replacing legacy systems for TV audience analysis and ad placement operations.  

3 comments about "Mining Value From Seconds-Based TV Measurement".
Check to receive email when comments are posted.
  1. Ed Papazian from Media Dynamics Inc, February 27, 2024 at 1:20 p.m.

    Some interesting points, Gerard. And some questions. However ---and with all due respect---no matter what is decided advertisers aren't going to learn much about commercial viewing that they don't already know--which isn't a lot, based on our current "audience measurements". Second by second "viewing"data for commercials sounds really sexy until one realizes that all that is being reported is set usage by second---not viewing. Moreover, don't expect major differences between commercails or their placements  based on set usage findings. As was mentioned, sellers will mash all of it together by rotating all commercials in all positions---or by charging such high CPMs for special positioning that there's no ROI benefit. Even worse is the question of what to do when a person "views" a commercial for four seconds. I assume that this will be solved by the absurd idea of charging by the second, not by the commercial length----or unit. I can't wait to see what happens when they wake up the CMOs and "creatives" and try to sell the notion that each second has equal value, no matter how many seconds are "viewed". Poof!  So much for story telling in commercials.

  2. Darrin Stephens from McMann & Tate, February 27, 2024 at 3:47 p.m.

    Just a quick reminder that true "A" pod positions are nearly non-existent in cable, and becoming more rare on broadcast TV.
    Because the buyers let them get away with it.

  3. John Grono from GAP Research, February 28, 2024 at 12:35 a.m.

    Thanks Ed.   Saved my time raising similar issues.

    There is another consideration.   And I hope it has been considered and included in the new measurement.    There is an unwritten assumption that all seconds are equal.   Some examples to consider:

    • is a second at (say) 05:15:27 equivalent to a second at (say) 20:34:51 (just picking random times)

    • is a second in the Super-Bowl to a second in an old re-run of a (say) 1960's half-hour show that flopped

    • during the second is it just person or more ... or maybe none getting a drink from the fridge

    • and issue of latency - quick cable or fast internet would have a different one-second time-stamp per the clock

    • the frequency and/or recency of seeing the ad needs to be considered ... recall doesn't require per each second

Next story loading loading..