Arbitron Radio Advisory Council Minutes

July 16-18, 2008
Teton Village, Wyoming

Arbitron Chairman/CEO Steve Morris opened the meeting with four key points:

1. Mr. Morris stressed that he and the company are proud of the PPM™ service. While Arbitron believes the data was statistically valid a year ago, he stated that the PPM sample quality has improved significantly in the past several months due to the company’s Continuous Improvement program. He thanked the Council for its advice and said Arbitron will continue to look for ways to raise the bar on PPM sample metrics.
2. Mr. Morris said that he recognizes that Arbitron’s relationship with radio clients has become marked by a level of distrust, and he feels that it is important to return to a more collaborative dialogue. Mr. Morris reported that the decision to delay PPM commercialization last November cost the company a significant amount of money and that it had invested even more money in the many improvement initiatives since then. Mr. Morris asked for the Council’s guidance in setting priorities for planned future investments designed to raise standards even more over time. He also asked Council members to raise any concerns with him directly, instead of going through back channels or the trade press.
3. Mr. Morris asked the Council to recognize that the proposals Arbitron will be making at this meeting represent a carefully considered balance of initiatives.  Now that PPM sample metrics have moved in a positive direction, MRC accreditation has become the top priority for PPM.  Balancing initiatives aimed at that primary goal with those meant to address other industry interests will produce business tradeoffs, and the inevitable by-product of those trade offs is that not everyone will be completely satisfied. He stressed, however, that MRC accreditation is the company’s number one PPM priority. He noted that there has been a lot of trade press coverage about achieving 100 DDI’s in a range of narrow demographics.  Mr. Morris said the company does not agree that producing 100 DDIs in individual demographic groups is required to provide a valid sample and, therefore, Arbitron will not guarantee 100 DDI’s.  However, he did say the company will continue work toward getting as close to 100 as possible given the constraints of random sampling.
4. Regarding Diary-measured markets, Mr. Morris noted rising client frustration with sample quality and that Arbitron is putting a number of significant service improvement projects on a fast track.

Several Council members challenged Mr. Morris on his statement regarding the need to achieve 100 DDI’s in narrow demographics. Mr. Morris pointed out that quota sampling, for example, could produce 100 DDI’s, but it would likely not achieve MRC accreditation because it is not projectable. Mr. Morris indicated that there are ways to work toward improved DDI’s through methods that Arbitron will describe during the meeting.

PPM Sample Size Strategy

Arbitron President of Technology, Research and Development Owen Charlebois described the methods by which Arbitron plans to increase its PPM samples.

He prefaced his remarks with some background points:

• Clients are concerned about PPM sample sizes, especially in smaller PPM markets and among narrow demos.
• PPM panels include Children 6-11, which is perceived by many broadcasters to have limited value.
• There is customer frustration over the belief that Arbitron calibrated its 6+ PPM sample based on 12+ Diary samples, which Arbitron believes is a misperception.
• Smaller PPM sample sizes restrict some customers from accessing data if minimums are held at the levels used in the Diary service.

Mr. Charlebois proposed a “holistic solution” to the issue of PPM sample sizes. Among these proposals:

• A commitment to higher benchmarks;
• An increase in the 12+, and therefore 18-54, sample size target, by 10%;
• A modification of the rules governing minimum In-tab reporting;
• Evaluating the concept of permitting client-purchased over-sample; and
• A long-term commitment to continuous improvement.

First and foremost, however, Mr. Charlebois declared that MRC accreditation was Arbitron’s number one PPM priority.

Mr. Charlebois then went into more detail about specific aspects of the plan:

Higher Benchmarks

• Mr. Charlebois said progress toward achieving the original benchmarks has been faster and further than anticipated. As such, Arbitron is prepared to raise sample guarantees and benchmarks based on current performance and a pipeline of new improvement initiatives.
• The proposal calls for the 6+ benchmark DDI to be at 90 from the start of currency in the market (currently the 90 DDI benchmark begins with the fourth currency month). The benchmark would then be raised to 100 DDI effective on the first anniversary of currency, based on a 13-month rolling average (the rolling average is used to account for seasonality in the results).
• The 18-54 guarantee would be 80 DDI from the start of currency (currently the guarantee starts with the fourth currency month). This guarantee would transition to a 90 DDI effective on the first anniversary of currency, based on a thirteen-month rolling average.
• The 18-34 benchmark would be at a 70 DDI starting with the first currency report, through the sixth currency report (currently the 70 DDI benchmark begins with the fourth currency month). The benchmark would then rise to 75 from the seventh through twelfth month of currency and then rise to 80 DDI effective with the first anniversary of currency, based on a thirteen-month rolling average. This benchmark would also apply to Black and Hispanic 18-34 demos in those markets that have ethnic controls.
• The 6+ annual In-tab rate benchmark of 75% would become effective with the first currency book (currently the benchmark begins with the fourth currency book).
• The 18-34 annual In-tab rate benchmark of 60% would become effective with the first currency book (currently the benchmark begins with the fourth currency book).

Mr. Charlebois said customers should expect variation in performance across time and markets. He also said Arbitron needs to ensure that overall research quality is not sub-optimized by focusing too much on one metric at the expense of another. For example, increasing the 18-34 DDI target could negatively impact P6+ In-Tab rates.

Arbitron Will Increase Sample Size Target by 10% at No Cost to Customers

Mr. Charlebois said the 10% higher target is the approximate equivalent of recalibrating the panel for Children 6-11. The implementation of the sample increases would begin in 2009, with expected completion by the end of 2010. An increase in 12+ targets would affect all demos proportionally, meaning that 18-34 and 18-54 targets are expected to increase by approximately 10% as well.

He said Arbitron is looking at several options for delivering this 10% increase in the 12+ sample size target:

• Increase the cell-phone-only sampling rate (currently at 10% for Philadelphia and 7.5% for all other PPM commercialization markets, except in Houston-Galveston, which has no cap)
• Arbitron is conducting a test this fall that uses an address-frame sampling approach for cell-phone-only households. If the test is successful and the initiative implemented, it would have a modest sample size impact on 18-54 and a larger impact on 18-34 In-Tab samples. Mr. Charlebois noted that as the cell-phone-only rate was increased, the landline sampling rate would be reduced so that the overall size of PPM panels based on 6+ would remain the same with better distribution in 18-34 and 18-54.
• Use sample stratification
• This initiative calls for sampling randomly within a demographic group, and requires full enumeration of PPM households at the “sample prep stage,” the first step in the PPM sampling process. Sample stratification will permit Arbitron to adjust the sampling rate (either higher or lower) for specific demographic groups, e.g., lowering the sample of 55+ persons. Mr. Charlebois pointed out that although the MRC supports stratification in concept, the specific implementation details of this change would need to be reviewed in detail with the MRC. Mr. Charlebois also noted that stratification is a less efficient, and therefore more costly, method of sampling than the current PPM sampling method and that there would be a limitation to how granular Arbitron could implement this change.
• Addressing Children’s sample
• Reallocating the children’s sample would result in a 10% increase in 12+ sample target overall, and to the 18-54 sample size target specifically. Arbitron envisions two scenarios with its 6-11 measurement:
• Remove 6-11s from the sample. To do so would require an assessment of the impact of removing children on 18-34 and ethnic household representation, and that such a change would require a review of the impact on sample quality with the MRC.
• Retain 6-11s and raise the 12+ and 18-54 target anyway. The idea here would be to find revenue to offset costs from sources other than existing radio customers.

Implementation and Impact of Sample Size Increase Plan

Mr. Charlebois noted that Arbitron is achieving many of its DDI goals across markets as a result of panels that are significantly “overbuilt,” i.e., well above the P6+ designed sample target.  He told the Council to expect Arbitron to start managing the PPM P6+ sample sizes closer to overall sample targets.

Based on the sample size increase plan Mr. Charlebois outlined, the currently expected timing of specific initiatives is as follows:

• 4Q 2008: Cell-phone-only address frame test
• 2Q 2009: Test impact of removing children
• 2Q 2009: Increase cell-phone-only sampling rate
• 3Q 2009: Decide whether or not to remove Children 6-11
• 1Q 2010: Begin installing new markets on 12+ base and replace children, or keep children and begin increasing 12+
• 1Q 2010: Implement stratification (if needed)
• 4Q 2010: Full impact of all steps realized and 12+ sample targets increased 10%

Reduce In-Tab Minimum for PPM Respondent-Level Data Access

Bill Rose, senior vice president of Marketing, discussed changing the In-Tab minimum for PPM respondent-level data access.  Mr. Rose said Arbitron planned to reduce the minimum In-Tab for respondent-level software from its current minimum size of 30 to 20.  Mr. Rose said the current Diary service minimum is 30, but when taking into account the standard 3:1 Diary to PPM ratio, the current 30 minimum is the equivalent of approximately 90 Diaries.  Mr. Rose stated that the proposed 20 PPM minimum would provide greater reliability than an estimate generated from 30 Diary respondents.

One reason for the change is that Arbitron’s summary data sets already generate standard demo estimates with no minimum requirement and that a change in respondent-level thresholds would simply match what is already available from summary-level services. Another reason is that the 30 minimum limits the availability of reports especially for highly targeted stations/formats.

Council Reaction to the Proposals

• The Council asked Arbitron to accelerate the timeline for increasing the sample target.
• The Council was appreciative of the movement on benchmarks, but it remained concerned that Arbitron is not doing enough to raise 18-34 DDI’s. It asked Arbitron to consider a 100 DDI benchmark for 18-34s.
• The Council felt that changing the 30 In-Tab minimum to 20 would lower the standard at a time when the industry wants the standards to be raised.  The Council asked Arbitron to consider keeping the minimum at 30 for custom geographies in its respondent systems but enable Metro-level reporting for Black and Hispanic demographics.  The Council felt these steps would keep guardrails in place for the respondent systems, while enabling consistency between software systems and equal access to demographic data.

Arbitron Responses to Various Council Requests About PPM:

1. Changing the Data Access In-Tab Minimum From 30 to 20

• Arbitron agrees to allow customers to generate PPM estimates based on Metro level, standard dayparts in respondent software consistent with the current practice for Arbitron ebookSM and summary dataset (no minimum for the Metro) and to add access to Black and Hispanic Metro-level information in race-weighted markets.  Arbitron will keep the current 30 minimum sample size for respondent-level data for geographies other than the Metro.

2. Accelerating 12+ PPM Sample Increase

• Mr. Morris said that Arbitron has not adequately clarified a misperception that Arbitron took valued sample out of 12+ and put it into 6+. He said that the Effective Sample Base of 12+ in PPM has always equaled the 12+ of the Diary. Mr. Morris said Arbitron does, however, agree that more sample will always be worth trying to achieve and will evaluate acceleration of the timetable for the 10% increase.  Arbitron will get back to the Council regarding that timetable by late August.

3. Raising 18-34 benchmark

• Mr. Morris said Arbitron cannot accept the notion of setting a 100 DDI benchmark for 18-34s. He said that Arbitron believes that at an 80 DDI, the data is solid and usable. Using random sampling as the basic platform of the PPM methodology, there will inevitably be some distribution of the DDI’s of specific demographic groups—some above 100 and some below.  He said Arbitron is currently achieving DDI’s among 18-34s that are greater than 80 as a result of oversampling—but continued oversampling is “nonsustainable.” Mr. Morris stated that Arbitron intends to work its sample down to 100 DDI for 6+, which may result in 18-34 DDI’s returning to the mid-80s. He cautioned the Council that the company is bringing to market four complicated and expensive initiatives, including some MRC-focused initiatives, and cautioned against trying to do something more on DDI’s that may have a negative effect on MRC-related metrics such as SPI.
• Mr. Morris also noted that Arbitron’s definition of “benchmark” does not represent its aspiration. Rather, a benchmark is meant to represent the level at which Arbitron will put money into remedial action. However, in a world of random sampling, Arbitron will not commit to get every demo to a 100 DDI.

Council Reaction to PPM:

There appeared to be confusion among some Council members regarding the In-Tab minimum for PPM respondent-level data access issue. Mr. Rose clarified that the proposal was designed to allow equal data access for black and Hispanic demographics.  With this proposal, these groups would get the same Metro-level granularity from the company’s respondent-level services as they do from its summary data sets. Arbitron indicated it will provide to the Council a report of what the sample sizes would look like at various levels of DDI.  A Councilmember requested that the company provide a similar report for Hispanic, based on English/Spanish-language dominance.

Regarding the accelerated PPM sample increase timeline, the Council remained concerned that Arbitron may not increase the sample sizes until the PPM commercialization is complete—the Council believes the industry needs it sooner.  Mr. Morris said that Arbitron will look for ways to accelerate the timetable and report its assessment to the Council in August.

Council Chairman Chuck DuCoty reiterated the Council’s desire to see 18-54 and 18-34 benchmarks at 100. Mr. Morris said that MRC accreditation is the company’s top priority and the company will need to balance initiatives designed to achieve that objective with the desire to narrow the difference between 100 DDI and current levels.  For this reason, Arbitron will not agree to a 100 benchmark for either 18-54s or 18-34s.  Mr. Morris reiterated, however, that Arbitron will continue to strive to narrow the difference and that it will provide the Council with timelines for the initiatives that the company expects to use to help improve sample performance.

Diary Day

The Council dedicated a full day to Diary service issues. Arbitron consultant and former RAB President/CEO Gary Fries told the Council that many of Arbitron’s smaller market customers weren’t happy with the service; they didn’t see a return on their investment and believe ratings have increasingly less relevance to their business needs. Many of these comments came from the Small Market Operators Caucus, which has met several times under Mr. Fries’ leadership.

Mr. Fries said two particular items of interest to Diary customers are the excessive “slicing and dicing” of data and whether they were getting sufficient ROI from Arbitron. Jay Mitchell, the publisher of the Small Market Radio Newsletter, helped Fries to estimate the amount of small market station revenue attributable to Arbitron ratings. After some probing, they estimate that approximately 30% of their business is impacted by Arbitron ratings.

Diary Service Changes

Arbitron Vice President of Research Policy and Communications Dr. Ed Cohen said improving young adult proportionality requires a holistic approach that begins with cell-phone-only respondents. Cell-phone-only households are a principal reason that young person representation has fallen over the past four years. An estimated 15.8% of households had only cell phones in their homes in late 2007.  
Cell-phone-only households are MORE likely to be…

• Unrelated adults
• Aged 18-29
• Students
• Lower income
• Hispanic or black consumers

Cell-phone-only households LESS likely to be…

• Homeowners
• 45+
• Middle and upper income
• Live in the Northeast

Dr. Cohen said Arbitron is working on reaching cell-phone-only households with an address-based sample frame. The preliminary research looks promising, according to early data from a Spring 2008 test.  Arbitron is targeting the start of cell-phone-only sampling in the Diary service in 2009.

Dr. Cohen cautioned that, while it is a promising tool, cell-phone-only:

• Is not a panacea
• Population estimates are limited at best
• Weighting will be on a 12+ basis

He also reported that Arbitron’s efforts to amend the Telephone Consumer Protection Act, which currently prohibits calling cell phones with a predictive dialer, are ongoing but the status has not changed since the last Council meeting.

Dr. Cohen also updated the Council on these other sample quality initiatives:

• Treatment Redistribution: This initiative redirects premiums away from 55+ households into younger households. It is based on household enumeration. A test is scheduled in Fall 2008, with a more complete roll-out planned in 2009 pending successful results.
• Second Chance Diary: This initiative was introduced to continuous measurement markets in Spring 2008. Early data indicate that it has a positive effect on response rate and proportionality.
• Second Chance to Consent: This initiative provides another opportunity to be in the survey for those who do not agree to participate.  It is being tested in Summer 2008 and implementation date is pending results of this test.
• Young Male Promised Incentive: This initiative provides an additional $10 promised incentive for 117 noncontinuous markets starting in Spring 2008.  Early results look promising.

Qualitative Update and Discussion

Tom O’Sullivan, Arbitron vice president of Diary Market Development, and Gregg Lindner, executive vice president of Scarborough Research and Operations, discussed initiatives to provide more qualitative information to clients in Diary markets. Mr. O’Sullivan pointed out that as much as 70% of the radio revenue in markets ranked 101+ is locally driven. Further, as “ratings based” revenue declines in importance, Arbitron needs to develop tools to help broadcasters grow their local business.  Scarborough provides local-market-specific qualitative information, allowing broadcasters to prospect more effectively, provide media comparisons and better understand the needs of their advertisers.

Improvements to Arbitron’s qualitative services:

• Updated Qualitative Diary: Arbitron has updated the Diary-based qualitative questions effective with the Fall 2008 survey. This information will be available via Maximi$er® and Tapscan™ Web.
• Arbitron ebook Radio Marketing Guide Prototype: Arbitron is prototyping new Radio Marketing Guide reports in ebook. This new report will utilize the existing qualitative questions within a market providing broadcasters with simple qualitative profiles for key business segments. The prototype for selected markets is planned for Fall 2008.
• Scarborough Expansion: Scarborough is evaluating doubling the number of its mid-market tier markets.  The mid-tier service is a more affordable version of Scarborough tailored to midsize markets.  Today it is in 15 markets with four to be added in Fall 2008.  Scarborough has the bandwidth to add 20-40 additional markets in 2009.
• Gathering Qualitative and Quantitative From the Same People: Scarborough and Arbitron are collaborating on a test that combines Diary ratings with local market consumer behavior data using a single respondent source.

Reduce “Slice and Dice”

Mr. O’Sullivan told the Council that the Small Market Owner/Operator Caucus expressed concern over the ability to analyze data using very small Diary sample sizes.  They requested that Arbitron develop ideas to reduce “slice and dice” of data to create more stable estimates for key buying demos.

Arbitron asked the Council’s feedback on the following ideas:

• Limited reports to “must have” buy/sell demos
• Stop reporting of single surveys in two-book rolling sample markets
• Limit county and zip code breakouts
• Expand two-book rolling sample system into the standard two-book markets

Pierre Bouvard pointed out that Arbitron’s small market service was started with the idea of providing customers with data for the key buying demos only. It was only after repeated customer requests, along with higher sample sizes, did Arbitron software permit the analysis of smaller demo cells.

Council members’ questions/feedback:

• Improved sample size would minimize the need for the slice and dice discussion.
• Another said, “To paraphrase James Carville, It’s the sample, stupid.”  If you fix the sample proportionality in 18-54 demos, the “bounce” you will see in the estimates will be lessened, eliminating the need for the “slice and dice” discussions.
• History shows that users attempt to reproduce age and sex cells if not published by Arbitron, which leads to confusion in the marketplace.
• Even if reported estimates were limited to broad buy/sell demos, stations will continue to depend on target age/sex cells for their core audience. Therefore, preventing access to these core target cells could make it more difficult for stations to fully understand their performance for both programming and sales purposes.
• A proposal to substitute rolling sample reports in two-book markets had already been rejected by the marketplace. Many of these markets view themselves as more like large than small markets and have concerns about losing the granularity of data that individual samples and reports provide.
• Arbitron should speak with agencies in the smaller markets to get their viewpoints.

Arbitron Responses to Various Council Requests About Diary Markets:

• Mr. Morris acknowledged the sentiment of the Council that sample sizes in key demographics are the most immediate concern for Arbitron’s clients. He pointed out that Arbitron has several key initiatives in the testing or implementation stages right now that are meant to address that concern.
• Regarding benchmarks for the Diary service, he noted that Arbitron has never had them before, other than a 90 In-Tab guarantee for Diary markets. He said the Diary service is currently running in the 80’s for 18-54 proportionality. However, he said the 18-34 demo is clearly the area of concern. Given the sample improvement programs Arbitron is putting in the market in the next six months, he is not inclined to address benchmarks right now. He stated that once the company sees the results of the initiatives in the field, it will then discuss with the Council what appropriate benchmarks might be. He also agreed to consider narrowing the Diary guarantee to 18-54, but again Arbitron will be better able to do this after it sees what the cumulative effect is of the initiatives now headed toward implementation.
• Mr. Morris reiterated that the immediate issue is sample improvement, and while adding more qualitative will be crucial to an optimized Diary service, it must be a second priority.
• Based on a Council request, Arbitron is investigating pilot markets in which to test the Scarborough mid-tier service.

Council Reaction to Diary Initiatives:

Chairman DuCoty said the Council came away from Arbitron’s Diary presentations “feeling unfulfilled.” He said the Council recognized Arbitron’s commitment, but the industry needs a greater urgency relating to Diary sample quality.  The Council remains “very, very concerned” about issues over sample. He stated that if Arbitron can establish benchmarks and 18-54 sample guarantees for PPM, then it should do so for the Diary.  Mr. Morris stated that Arbitron is open to the concept but the company needs to focus first on improving Diary sample quality before tackling the Diary benchmarks. 

Mr. Rose promised to get back to the Council in September to report on the result of the Spring 2008 initiatives. At that time, the company will also update the Council on cell- phone-only implementation plans.

Council Business:

Lynn Bruder, General Manager of WRDW Philadelphia, joined the Council representing Youth formats.

The Council declined a request from the Radio Network research committee to add a Council seat representing radio networks. Chairman DuCoty said the Council’s mission is to deal with locally licensed radio stations and their issues only.

Arbitron eBook^SM, Maximi$er® and PPM™ are marks of Arbitron Inc. Media Rating Council® is a registered mark of the Media Rating Council. TAPSCAN™ is a mark of TAPSCAN Inc., used under license.