Watch all video interviews from the Advertising Research Foundation Audience Measurement conference.
Sponsored by
Synopsis
Return path data is extremely rich with second-by-second behavior for extremely large samples. This paper explores the normative behavior of a metric called Tuneaway for a wide variety of media related variables. Tuneaway measures the holding power of a commercial. It is defined as the percent of available commercial time that viewers do not tune to. The norms for High Definition TV vs. Standard Definition TV, time of day of viewing, commercial position in pod, pod position in program, network type, programs and commercial length are shared. A deeper understanding of tuneaway across these variables can help to inform the media decision process.
Background
Kantar Media is actively developing RPD applications (Shababb, 2007; Katz and Shababb, 2008; Kaufman, 2008; Shababb and Drake, 2009). This data provides extremely rich second-by-second data from extremely large samples. These data have been used to investigate measures that can be used to better understand creative holding power. Tuneaway is a just such a measure of holding power. This measure was developed by Kantar Media in our measurement role for the Comcast/SMG addressable trials (Kaufman, 2008).
These studies join a growing search for ways to use return path data for commercial applications. These include ways to connect viewing to consumers (Bachman, 2009): via purchase (Kang, 2008),via geography such as zip code (Schmelling, 1998), for addressable advertising (Frey, 2008; Kumar and Steel, 2009) and for measuring
audiences for cable networks with samples that are too small for Nielsen to measure (Mullaney, 2006, Schechner and Vranica, 2009).
Studies that have investigated tuneaway have shown that there are lower levels of tuneaway for buyer targets (Katz and Shababb, 2008) that increased exposure frequency increases tuneaway (Shababb and Drake, 2009) and that tuneaway varies for the same commercial under different circumstances (Zigmond, Doari-Raj et al. 2009;
Zigmond, 2010).
The work by Zigmond (2010) explored removing the influence of media factors from commercial tuneaway estimates with the purpose of using the actual delivered tuneaway vs. the estimated tuneaway as a measure of a particular commercial’s efficacy. This paper turns that premise around and shares the normative values of the tuneaway for those media factors with the purpose of providing media planners with data to enhance their decision making. The variables that are explored include broadcast clarity (high definition vs. standard definition), position in pod, pod position in
program, daypart expressed as hour of the day, minute within the hour, program, and broadcast network
Research Question
The key question being addressed in this research is “How does tuneaway, or commercial holding power change across the key variables media planners use to select media?”
Findings
Tuneaway does vary across many variables. The variable with the largest variability is commercial duration. See Figure 1.
Figure 1: Tuneaway by Commercial Length
This finding is expected since there is four times as much time to tuneaway from a :60 than a :15. However, tuneaway is a percent of the audience that was in the audience 5 seconds before the commercial began. So, with each succeeding second there are fewer viewers that can leave the audience. If we plot a constant percent loss of remaining viewers, we have the red dotted line. This is a curve of diminishing returns.
What we can see is that the :15 has a lower percent tuneaway per second than the :60, as does each shorter commercial have as compared to longer commercials. In other words, if there were no audience gain across the seconds, there would be more tuneaway for a :60 than four :15’s.
Due to the high difference between tuneaway by commercial length, all further analyses were conducted on the :30 commercials so that the differences could not be attributed to commercial length.
In Figure 2, it is clear that where the commercial is in a pod really matters for tuneaway.
Figure 2: Position In Pod
The first commercials in a pod have the highest tuneaway. Common wisdom is that only the first and last commercials retain their audience. The data here do not identify which was the last commercial in the pod, so we can’t conjecture about the last commercial, however, clearly tuneaway for the first few commercials is much higher than those later in the pod.
Figure 3: Pod Position in Program
Figure 3 shows that the later the pod is in the program, the higher the tuneaway. Pods at the start of a program retain more viewers on average than later pods in a program. At the beginning of a program, perhaps viewers are more reluctant to leave.
Figure 4: Live+3 Audience(000)
As the Live +3 Audience(000) grows, tuneaway consistently falls with a bit more dispersion at higher audience levels.
Figure 5: Hour of the Day
There is a definite pattern across the day. There is far higher tuneaway during the day, with a dip in late morning. Starting at 6pm, there is a steady decline across the evening until it begins to return to high morning levels at midnight.
Interestingly, this is does not match rating size, since primetime is the period of steady decline.
Figure 6: Seasonality (Months)
Most of our analysis was conducted on the broadcast month of October (which includes a small bit of September) in order to collect a very large sample of commercial airings during a common period of time, which would have the same programs and environment. We added a week of data for three other months to see if there was a seasonal pattern.
The tuneaway for Live as compared to Live+3 was almost identical – making it impossible to see one set of data points over the other. The differences between months are not dramatic, but they are significant (see Appendix A for confidence intervals.) The highest levels of tuneaway are in August and the lowest in April.
Figure 7: Broadcast Definition (HD/SD)
In Figure 7 the tuneaway patterns for high definition and standard definition are compared. This is perhaps the most interesting pattern in this paper, in part because it is a new finding. Each of these data sets are mutually exclusive. SD Only in Cable Networks are networks that only broadcast in standard definition. The HD and SD broadcast and the HD and SD cable networks are the major networks that broadcast in both, reporting the tuneaway for each segment.
Ad supported cable networks that provide an HD feed show greater audience retention on their high definition channel than on their standard feed. The mean TuneAway for these HD channels is 3.73 compared to 6.8 for standard definition viewing. There is a similar, although less pronounced pattern for the broadcast networks: HD has 2.2 TuneAway and broadcast SD has 3.2. This might indicate these heavy viewing TV households are more engaged in television — or it might also reflect another secondary characteristic of this group’s behavior such as income or family size. Greater audience retention and its implied higher commercial rating is a powerful rationale for networks
considering adding a HD feed to their distribution.
Figure 8: Network Type
Network types also show distinct differences in tuneaway. We see that as Children get older, tuneaway increases from Kids to Family to Young Adults. We also see that Spanish Cable and Sports top the list of tuneaway, while Broadcast and Spanish Broadcast have low levels of tuneaway.
Figure 9: Top 15 Programs by Number of Spots
Tuneaway varies dramatically by program. These are the fifteen programs that had the most commercial spots – the programs that were used the most times. In this table you can see that we have rather large samples for these programs: ESPN News has almost 5,500 spots, and the smallest of the 15 programs, The Bernie Mac SHow has 600 spots.
You can also see that the tuneaway for these programs varies tremendously. Sometimes the tuneaway matches the type of network it is like Sports/News for ESPN News with very high tuneaway and sometimes it matches the kind of rating it has, such as NCIS or Law and Order with a high rating and low tuneaway. In general, there is a strong correlation between TuneAway and ratings size for these programs.
STUDY DESIGN
Data
This study was conducted using Kantar Media’s national DIRECTView service for twelve of the largest advertised categories’ key brand. The category/brands are: Automotive: Chevrolet; Communications: AT&T, Vonage; Department Stores: Macy’s; Insurance: Geico, Progressive; Misc Services: EHarmony; Prepared Food: Campbell’s Soup; Restaurants: Pizza Hut, Subway, Burger King; Retail: Home Depot. Kantar Media’s national DIRECTView service includes a 100,000 sample of return path satellite from homes that subscribe to DirecTV. This data was analyzed at household level, not at the individual set top box level.
The broadcast month of October 2009 was used for primary analysis. Results were replicated using representative weeks from 3Q 09, 1Q 10 and 2Q 10. All network commercial activity was tracked during this time across all channels and all dayparts.
There were 83 networks included in the analysis.
Commercial Occurrence data was provided by TNS-MI.
Analysis
The analysis began by analyzing the differences between the different commercial lengths. The vast majority of commercials were :30’s, followed by :15’s. See Figure 10.
Figure 10: Number of Spots by Commercial Length
The differences in tuneaway were dramatic by commercial length. Because of the large differences, the further analysis of this data was conducted using only the :30 second commercials. These were by far the largest sample of commercials.
Average levels of tuneaway were calculated for each variable shown with confidence intervals for these figures in Appendix A.
CONCLUSIONS
Tuneaway varies dramatically across many of the criteria used to select media and information used to develop a media plan.
Patterns of tuneaway can and should be used to help create better media plans.
The key conclusions for media planners are that: 1) higher rated programs have less TuneAway, 2) earlier positions in pod have more audience and less TuneAway, but not the first position and 3) there are dramatic differences by program genre.
The key conclusions for broadcasters are that shorter commercials have less TuneAway, and that position in pod is critical. In addition, that the HD feed may be a place for the committed viewer.
Author Bios
George Shababb
George Shababb is President of Kantar Media Audiences (formally TNS Media Research), and is responsible for the strategic planning and development of products and services related to digital audience measurement in the United States. Under Mr. Shababb’s leadership, Kantar Media Audiences has successfully pioneered the introduction of TV audience measurement services based on clickstream data sourced from digital set top boxes.
Mr. Shababb has been widely recognized for achievements in the field of digital audience measurement. For his work, he was named the 2009 silver recipient of the prestigious ARF Great Mind award for innovation. In addition, in 2007 and then again in 2009, he was named to the Mediaweek 50, which features the 50 most indispensable
executives shaping the future of media.
Mr. Shababb is a frequent panelist at industry conferences and has been featured in the Wall Street Journal (including a Media Q&A feature in the Business section), BusinessWeek Online, Advertising Age, Adweek, Mediaweek, TelevisionWeek, Multichannel News, among other key industry publications.
Joining TNS in 2000, Mr. Shababb served as Senior Vice President of TNS Media Intelligence (now known as Kantar Media Intelligence) where he spearheaded strategic planning and business development for advertising expenditure measurement.
Mr. Shababb has over 30 years of Sales and Service, Marketing and Operational experience working with leading information providers including VNU, AC Nielsen, The NPD Group, and Rx Remedy. Mr. Shababb is a member of the ARF Board of Governors. He holds a Masters Degree in Statistics from the University of Connecticut and is a named inventor on two patents for technology facilitating data collection.
Leslie Wood
Leslie Wood Research (LWR) is a media research consulting company offering two main areas of service: general media research consulting and programming and designing proprietary advertising systems that are simple to use and bring sophisticated thinking to communication planning and buying. LWR offers vast experience in media research, with programmers, statisticians, modelers, knowledgeable industry consultants, data processing/data entry staff and access to many industry services and data.
Leslie Wood has been a pioneer and innovator in reach & frequency, optimizers, fusion, single source data and return path data analyses. She is a co-chair of The ARF’s 360 Media and Marketing Council and the Analytics/ROI/Data Integration Committee.
LWR was deeply involved in Project Apollo, a joint venture of Arbitron and VNU with substantial support from Procter & Gamble. LWR built the user interface and software applications for PPM’s National Marketing Panel test in Philadelphia. This included being responsible for many of the business rules regarding weighting, data integration, defining statics/intab criteria, media measurement and reporting for that test and designed the software to process that data. She has also spearheaded a team of researchers to analyze the Project Apollo data for P&G. This team has been formed
into Media Trust, LLC.
Leslie Wood has a BS in mathematics from Hunter College and a PhD from The University at Albany in Information Science with a specialization in Expert Systems and Data Mining.
David Zornow
Dave Zornow has 30 years of media research experience working with broadcast and network cable, spot TV, print, radio and set top box data. He is President and Founder of TNG Research, a media research consultancy and app development company that provides information and analytical solutions for media buyers, sellers, vendors, and marketers. TNG Research provides applications development, solutions, and media research consulting for cable networks and ad agencies, working with research providers such as Arbitron, Mediamark Research, Kantar, OTX, the
Council for Research Excellence, and Nielsen.
From 2001-2006, he lead the CAB’s CONCAM group (Cable Advertising Bureau’s Committee on Network Cable Audience Measurement) as Chairman and Vice Chairman. Prior to founding TNG Research in 1991, Dave worked at Arbitron, MTV Networks, ABC, NBC, and in public radio.
In addition to research and app development work, Dave has been writing and publishing content in print and online for over 20 years. His articles have appeared in Cynopsis, Cable Avails, MediaPost, CableWorld, and Cable Sales Professional. Dave currently edits MediaNewsAndViews.com, a media industry collaborative blog with a
research focus.
Dave has a undergraduate degree in Computer Applications from American University and a Masters in Information Systems Management from George Washington Universtity. He is a former adjunct professor in the Computer Science Department of Montgomery College in Maryland.
