REAL CROSS MEDIA INTELLIGENCE FOR REAL CROSS MEDIA PLANNING The PPM contribution Roberta M. McConochie Beth Uyenco Arbitron s Portable People Meter (PPM) results provide real crossmedia duplication between radio and television. PPM differs from present currency-based estimation of random duplication between independent sources for local-market radio and television. This investigation, employing a variety of schedule variations for a variety of demographics, finds that random estimation generally overstates PPM s unified radio and television capture. Results show variations by target demo as well as by the scope of the plan itself. As PPM measurement is deployed beyond Philadelphia, its passive capture of cross-media flow promises a better basis for estimating the value of media-mix schedules.
2 Roberta M. McConochie, Beth Uyenco Background INTRODUCTION The past decades have brought revolutionary changes in media technology and contents as well as growing complexity, range, and specificity of consumers media preferences. In stark contrast, media-planning progress has been incremental and slower in pace. Though there have been advances in automation and a variety of tools have been added to planners repertoires, even in this richer planning world there remains a dearth of cross-media tools and appropriate respondent-level databases. The currently available crossmedia optimizers function on input derived from consumer-supplied recall data, or they fuse or integrate information from separate, single-media research. In other words, current systems cannot as yet fully account for individual consumer s cross-media flow, duplication, and interrelationships, nor can they help advertisers and agencies exploit consumers cross-media behaviors by identifying efficient, potentially effective, synergistic media plans that take advantage of the actual cross-media experience of consumer targets. PPM s real cross-media information provides an alternative guidance system. The authors have reported early investigations of PPM s radio and television cross media data (McConochie and Uyenco, 1 and 2). One of these identified small differences for broad dayparts and demos (McConochie and Uyenco, 1). These early investigations show the potential for PPM to contribute to cross media planning. Recap of Methods and Key Questions PPM measurement uses a panel design, tracking consumers up to one year and going forward, possibly over one year. PPM s participation is virtually passive. Panelists need only wear their meters. They do not need to write down anything or push any buttons, nor do they need to correctly identify the media entities they hear. The meter, which functions much like the human ear, registers all radio and television exposures, wherever they occur and whether or not the consumer has actively selected that channel. PPM picks up all audible media exposures, consistently, objectively, and passively. The portable meters recognize proprietary PPM codes emitted by media entities, broadcast TV, cable networks, and local-origination cable, as well as radio, which reach the Philadelphia DMA. Only media entities that are encoded can be measured by the PPM method. The codes are inserted into the audio portion of media contents. 1)
Real cross media intelligence for real cross media planning 3 Arbitron has maintained the PPM panel in Philadelphia in the United States at a size of 1,5 persons, aged six years and older, during 2 and the first quarter of 3. The PPM data reported here are from the Philadelphia DMA. The PPM system currently reports media estimates in the currency denominator of quarter-hours to facilitate the comparisons. However the PPM method estimates media use at the minute-level of precision. Results for this investigation are limited to a single U.S. market and to the intab sample for PPM and for Diary data. Moreover, these results, based on the nearly hundred encoded media measured in Philadelphia, may not generalize to other stations and channels in other markets. Additionally, as with most media research, these results may represent survey non-response bias as well as true relationship effects among the variables of this investigation. Nonetheless, the PPM panel results have shown impressive stability and consistency over time, in the United Kingdom as well as in the United States. The authors believe that these results offer far-reaching insights addressing the key research questions. The principle questions addressed in this investigation include: How does PPM s capture of duplication compare with that for random duplication estimation? How do these comparisons vary by demo both for broad buying demos and narrower targets? What are the patterns of variation for high-reach plans vs. lower reach? For more vs. less dispersed plans? RESULTS PPM Measurement Attributes Point to Potential Impact on Planning PPM s fundamentally different approach yields results that in some ways mimic today s currency. However, the PPM methods and results also differ, in some cases dramatically from currency. These similarities and differences frame the context for the investigation of duplication. PPM Captures the Real Media-Mix Potential over the Media Day In contrast to today s independent currency measures for radio and television, PPM simultaneously captures the cross-media flow among radio and
4 Roberta M. McConochie, Beth Uyenco television, including cable TV. The build graphs in figure 1 show how radio, broadcast TV, and cable build, converge, and diverge across the media day, for persons 6+ on both weekdays and weekends. Figure 1 PPM S MEDIA MIX POTENTIAL FOR RADIO AND TV OVER THE MEDIA DAY Pers 6+ AQH Ratings 4 3 1 5:AM 7:AM 9:AM 11:AM 1:PM 3:PM 5:PM 7:PM 9:PM 11:PM 1:AM 3:AM 6:AM 8:AM 1:AM 12:PM 2:PM 4:PM 6:PM 8:PM 1:PM 12:AM 2:AM 4:AM Broadcast TV Cable Radio Monday-Friday Saturday-Sunday Persons 6+ AQH Ratings, Philadelphia DMA, July 2 Obviously these build graphs have implications for planning, and they show the expected variation for individual demos. Figure 1 for total persons 6+, including children and teens, shows the convergence of midday ratings for radio and TV. Figure 2 for persons 25-54, the majority of whom are employed, shows a different daytime profile. (See figure 2.) The weekday media use curves for persons 25-54 shows substantially more accessibility via radio than TV, from 7am through 5pm. In contrast, the weekend media build charts show the close comparability of access via radio, broadcast, or cable. In addition to its unique capture of consumers real media mix, PPM also alters the capture of individual media components. For example, PPM captures the viability of reaching consumers via cable TV, more so than the present local meter-diary currency in Philadelphia (see figure 3).
Real cross media intelligence for real cross media planning 5 Figure 2 PPM S PERSONS 25-54 MEDIA-MIX POTENTIAL Pers 25-54 AQH Rating 4 3 1 Broadcast TV Cable Radio 5:AM 7:AM 9:AM 11:AM 1:PM 3:PM 5:PM 7:PM 9:PM 11:PM 1:AM 3:AM 5:AM 7:AM 9:AM 11:AM 1:PM 3:PM 5:PM 7:PM 9:PM 11:PM 1:AM 3:AM Monday-Friday Saturday-Sunday Persons 25-54, AQH Ratings, Philadelphia DMA, July 2 Figure 3 PPM CAPTURES VIABILITY FOR CABLE IN MEDIA MIX Cable Meter-Diary PPM Newer Bdcast Nets "Big Three" Nets. 5. 1. 15.. Pers 6+ AQH Ratings Persons 6+ AQH TV Ratings, Philadelphia DMA, July 2, M-Su, 6am-1am
6 Roberta M. McConochie, Beth Uyenco PPM s Ubiquitous Capture of Cable and Children s Viewing PPM s passive and portable media capture show about twice the persons 6+ PUT levels for total cable TV, in contrast to the meter/diary currency an eight AQH rating for PPM vs. the meter/diary four. A portion of PPM s higher cable capture comes from out-of-home viewing, nearly entirely omitted from the meter-diary currency. The PPM data have demonstrated that most persons in non-cable/non-satellite home are exposed to cable TV during their media days (e.g. Patchen, 2). The inference: non-cable panelists are exposed to cable when they are away from home. PPM also shows that younger consumers are more reachable by TV than estimated by the meter-diary currency (figure 4). Figure 4 PPM AQH TV RATINGS SHOW YOUNGER DEMOS MORE REACHABLE VS. LOCAL CURRENCY 4 NMR Meter-Diary PPM AQH PUT Ratings 3 1 P6-11 P12-17 P18-34 P35-54 P55+ May 2, Philadelphia DMA, Mon-Sun 6AM-1AM As with its capture of cable TV, PPM captures roughly twice the TV usage for children 6-11 and teens 12-17 as the active Meter-Diary method. As with cable exposure, a portion of children s and teen s media use is likely outside the home (Papazian, 2). Moreover the capture of young persons viewing by the set-diary method can be difficult and potentially bias prone as the entries may be made after-the-fact and by a person other than the child or teenager.
Real cross media intelligence for real cross media planning 7 In contrast to the PPM-currency differences in television, in radio the PPM method captures nearly identical total time listening (McConochie, Goerlich and Stinnett, 3) and similar demo radio-tuning levels as the diary (figure 5). 4 35 Figure 5 PPM AQH RADIO RATINGS SHOW SIMILAR DEMO REACH AS PERSONAL, PORTABLE RADIO DIARY AQH Ratings (PUR 3 25 15 1 Diary PPM 5 '12-17 18-34 25-54 35-54 55+ Spring 2, Philadelphia Metro, Mon-Sun 6AM-Mid PPM s Capture of Real Multi-Week Accumulation for Radio The PPM and Diary AQH ratings are similar for all the major age breaks. The similarity no doubt reflects the fact that both PPM and the Diary utilize personal and portable measurement. But from there, the similarities wane. PPM s panel measurement results in real accumulation of radio reach over time (McConochie, Jarvis and Dupree, 2). As shown in figure 6, PPM s passive panel measurement results in considerably higher station reach after only one week, with incremental add ons for subsequent weeks, in contrast to the diary s one-week, modeled data.
8 Roberta M. McConochie, Beth Uyenco Figure 6 PPM CAPTURES RADIO S REAL ACCUMULATION OF REACH OVER TIME Mean Diary Cume Mean PPM Cume Point Difference Index to Diary Day 7 7.51 16.48 8.97 219 Day 14 8.74 19.82 11.8 227 Day 21 9.45 21.93 12.48 232 Day 28 9.95 23.42 13.47 235 February 2, Persons 12+, Philadelphia Metro After one week of measurement, PPM s average station cume of 16.5 is more than twice that of the diary s 7.5, for an index of 219. After four weeks, the index is only slightly higher. Thus PPM s real cume build occurs early, in week one, rather than continuously as current reach models assume. Figure 7 graphs the PPM addition to diary cume, showing the cume contribution for the all-station average, the middle curve, as well as the high and low ends of the range of PPM cume addition for individual stations. Figure 7 PPM S CUME ADDITION CONCENTRATES IN WEEK ONE; STATION-TO-STATION VARIATION 6 High Side of Range 5 4 Reach 3 1 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 Average Added Points Low Side of Range Days February 2, persons 12+, Philadelphia Metro
Real cross media intelligence for real cross media planning 9 As shown in figure 7, there is considerable range of cume addition around that average. Figures 8 and 9 show two station-specific examples of PPM-to-Diary comparisons of PPM s real accumulation vs. the modeled accumulation based on one-week diary surveys. Figure 8 TALK FORMAT STATION S MULTIWEEK ACCUMULATION PPM VS. DIARY 1 9 Diary, 1st 7 Days Diary, 28 Days Modeled Maximi$er Schedule 8 PPM Weekly, Days 1-7 PPM Weekly, Days 8-28 PPM 28-Day Panel 7 Reach 6 5 4 3 +8.2 1 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 Days February 2, Persons 12+, Philadelphia Metro Figure 9 CLASSIC/CONTEMPORARY STATION MULTIWEEK ACCUMULATION 1 9 Diary, 1st 7 Days Diary, 28 Days Modeled Maximi$er Schedule PPM Weekly, Days 1-7 PPM Weekly, Days 8-28 PPM 28-Day Panel 8 7 +35.6 Reach 6 5 4 3 1 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 Days February 2, Persons 12+, Philadelphia Metro
1 Roberta M. McConochie, Beth Uyenco Both the PPM and Diary data show two sets of lines. The solid lines show actual 28-day data for PPM and seven-day data for the diary. The dashed lines show modeled data, based on each week for PPM, to assure that the 28-day sample results were stable. For the diary, the modeled data were run through Arbitron s software which models the cume build off of the current diarybased assumptions. Both station examples show the majority of PPM s cume build occurring early in the four week period, with smaller, incremental gains across subsequent weeks. And both examples illustrate the range of station-to-station variation. Obviously, PPM s real reach data have important implications for planning multi-week media schedules. PPM s measurement attributes, its capture of both radio and television on a single measurement platform, PPM s measurement of out-of home viewing, particularly for cable TV and for younger viewers, its capture of radio multiweek accumulation differ from currency measures, the input data to today s media schedules. These attributes imply that PPM would provide a source of refinement for duplication estimation. Variations in Schedules for this Investigation The OMD-Midwest planners developed systematic variations on original realmedia schedules as the design for this investigation. The variations on these schedules were run on May 2 PPM respondent-level data through Stone House Systems T-View software. There were three plan-level variations for the four different broad demo groups, as shown in figure 1. Figure 1 GRPS HELD RELATIVELY CONSTANT FOR DEMO VARIATIONS BY PLAN REACH Heavy Medium Light 15 Gross Ratings Points 1 9 6 3 Broad Demos, May 2, Philadelphia Teens W 25-54 M 18-34 Adults 55+
Real cross media intelligence for real cross media planning 11 The schedule-input GRPs were comparable across four key demo groups: Teens, Women 25-54, Men 18-34, and Adults 55+ based on real brand schedules reported in Nielsen Monitor-Plus. Each reach increment was approximately twice the level of the next lower variation. One challenge to using real schedules as a basis for this investigation was keeping the proportions of radio and television relatively constant. They are similar, as shown in figure 11, though not perfectly equal. Figure 11 PORTION OF RADIO GRPS ALSO FAIRLY CONSTANT OVER DEMO/REACH VARIATIONS 6 Heavy Medium Light Percent Radio in GRP Mix 4 Teens W 25-54 M 18-34 Adults 55+ Broad Demos, May 2, Philadelphia To the extent that there are duplication differences between PPM and Random Duplication, they may reflect the variations in input media mix by demo as well as the effects of PPM s real duplication data. Random Overstates Duplication Slightly for Broad Demos per PPM s Real Reach The first set of investigations holds the plan s reach-level constant at the mid range. We start by looking at the broad demo results.
12 Roberta M. McConochie, Beth Uyenco Figure 12 PPM S REAL RADIO-TV DUPLICATION OVERSTATED BY RANDOM MODEL PPM Real Reach Rand om Duplication 1 8 Percent Reach 6 4 Teens W 25-54 M 18-34 Adults 55+ Broad Demos, Medium-Reach Plans, May 2, Philadelphia All four broad demos show that Random Duplication produces some overstatement, ranging from a low of one percentage point for adults 55+ to a high of eight percentage points (89 for Net Random-based reach vs. 81 for PPM s real, unified Radio and TV capture). For teens, the overstatement of Random was six points; for women 25-54, it was four points. Though these differences are slight, they are consistent in direction. The extent of overstatement is not an estimate. Rather it is the actual difference between two different computations against the same database. Narrow Demos Show More Random-Duplication Overstatement For the Teen subgroups, the girls and the boys, the boys show higher levels of duplication vs. the girls, and so further increase the overstatement for total teens (figure 13).
Real cross media intelligence for real cross media planning 13 Figure 13 GREATER RANDOM OVERAGE FOR NARROW TEEN DEMOS VS. PPM PPM Real Reach Random Duplication 1 8 Percent Reach 6 4 All Teens Teen Boys Teen Girls Medium-Reach Plans for Broad Demos, May 2, Philadelphia Total teens show an overall six percentage point overstatement, based on an estimate of 88 reach for Random, vs. PPM s 82. For teenage boys, the overstatement is ten points. For girls, it is four points. Similarly, men 18-24 and men 21-29 show greater overstatement than the broader 18-34 demo. Figure 14 SIMILARLY GREATER OVERSTATEMENT FOR NARROW MEN S DEMOS PPM Real Reach Random Duplication 1 8 Percent Reach 6 4 M 18-34 M en 18-24 Men 21-29 Medium-Reach Plans for Broad Demos, May 2, Philadelphia
14 Roberta M. McConochie, Beth Uyenco The younger male target, men 18-24, shows 11 points overstatement: 81 for Random vs. 7 for PPM. Men 21-29 show nine points overstatement. These contrast with the eight-point overstatement of Random for the broader men 18-34 demo. As with the men, women also show more overstatement with Random Duplication both for the narrow women 18-24 break and for the 12-24 target than for the broader women 25-54 group. Figure 15 WOMEN S NARROWER TARGET SHOWS SIMILAR OVERSTATEMENT PATTERN PPM Real Reach Random Duplication 1 8 Percent Reach 6 4 W 25-54 W 1 2-24 W 18-24 Medium-Reach Plans for Broad Demos, May 2, Philadelphia In contrast to the women 25-54 four point overstatement (94 reach for Random vs. 9 for PPM) the narrower women 12-24 target shows a seven percentage point overstatement for Random duplication estimation (89 Random Duplication vs. 82 for PPM). The 18-24 target shows a five percentage point overstatement. Thus, the demo variations show that Random Duplication consistently overstates PPM s real duplication. Moreover, the level of overstatement varies according to demographic target. In general, the narrower the target, the greater the overstatement. (Figure 16).
Real cross media intelligence for real cross media planning 15 Figure 16 TEEN BOYS SHOW GREATEST % OVERSTATEMENT VS. PPM (MEN 18-24 AND 21-29) Teens Men Women Percent Random Overstate 15 1 5 B r o a d (T o ta l T e e n s, M 18-34, W 25-54) N a rr o w (T e e n B o y s, M /W 18-24) Na r ro w (T e e n G ir ls, M 21-29, W 12-24) Medium-Reach Plans by Demos, May 2, Philadelphia Narrow targets tend to show more Random-Duplication overstatement than the broad demos, with only one exception, teenage girls who show somewhat less overstatement than the total teen demo. The data points here show the percent overstatement for the broad demos (as opposed to the point overstatement presented above). As shown in figure 16, men 18-34 show the highest percent of Random overstatement among both the broad and narrower target demos. The men 18-24 show 1% overstatement vs. 7% for total teens and 4% for women 25-54. Similarly, among the narrow demo targets, men 18-24 show the highest overstatement, 15%. Men 21-29 are next highest with 12% overstatement and teenage boys with 11% overstatement are third. Results lead the conclusion that Random Duplication estimation does not describe the cross-media uses of men, particularly younger men, as well as it does for women. These results suggest that planners would need to adjust schedule goals differently for different target demos. Thus far, we have focused on a single level of schedule planning one of medium reach. We turn now to the impact of variations in plan level low, medium, and high. The medium plans aim for twice the GRPs of the light plans. The heavy plans aim for twice the GRPs as the medium, as illustrated in figure 12 above.
16 Roberta M. McConochie, Beth Uyenco Overstatement for Plan Level: Some Uptick for Higher Reach Schedules; Some Demo-Reach Interaction Variations in the level of plan reach show some uptick in overstatement for higher-reach schedules, though there are variations by demo target. The first set of results for reach variation focuses on the broad demo groups (figure 17). Figure 17 PERCENT BROAD DEMOS OVERSTATEMENT: SOME UPTICK FOR HIGHER REACH Teens Men 18-34 Women 25-54 Adults 55+ Percent Random Overstate 15 1 5 Light Plan M edium Plan Heavy plan Broad Demos, May 2, Philadelphia Two broad demos, men 18-34 and teens show increases in overstatement for Random Duplication with increases in the scope of the plan. For men, the percent overstatement is 4% for the light plan vs. 1% for both medium and heavy plans. Similarly for teens, the overstatement for light planning is only 4% vs. 7% for medium and heavy schedules. Adults 55+ show this tendency as well, but more weakly, with 1% overstatement for the light and medium schedules vs. 3% for the heaviest reach plan. Women show a slight uptick in percent overstatement only for the medium reach plan. In contrast to the broader groups relatively consistent patterns of overstatement, the narrow targets show considerable variation by plan level (figure 18). However, the plurality of narrow demos shows the highest overstatement for the medium-level plans. In particular, the male demos show the highest overstatement for medium-reach schedules. The singular exception to this pattern is women 18-24.
Real cross media intelligence for real cross media planning 17 Figure 18 MOST NARROW DEMOS MEDIUM/HIGH REACH SCHEDULES SHOW MAX OVERAGE Percent Random Overstate 15 1 5 Teen Boys Teen Girls Men 18-24 Men 21-29 W 18-24 W 12-24 Light Plan M edium Plan Heavy Plan Narrow Demos, May 2, Philadelphia Four narrow demos show the highest percent overstatement for the Mediumreach plan: men 18-24 with a 15% overstatement for the Medium plan s Random duplication; men 21-29 with 12% overstatement; teenage boys with 11% overage. Two of the women s demos show their peak uptick for the highreach plan: women 12-24, with 1% overstatement and teen girls with 6% overstatement. In contrast, women 18-24 show exceptionally high overstatement relative to Random for the lightest-reach schedule. This probably has to do with the particulars of the schedule and programming events. But these demo-reach variations suggest the complexities of random duplication for schedule variations and narrow targets, complexities that PPM measurement may help address. Overall, the impact of schedule reach level on Duplication overstatement depends upon both the target demo and the schedule variation. In other words the extent of Random duplication s overstatement hinges on the specifics of the plan reach and on the particular target. Therefore correcting the overstatement may require more than simple adjustments to the standard formula. In addition to taking stock of schedule reach and target, viable planning must also consider the impact of the density or dispersion of the schedule.
18 Roberta M. McConochie, Beth Uyenco Schedule Density/Dispersion Mix also Shows Complex Overstatement This investigation examined different schedule variations based on a template of medium-sized schedules for three of the four key demos: teens, men 18-34 and adults 55+. We varied the number of stations/outlets for both radio and TV plans from few to many. The variations included different mixes of cable TV as well as broadcast and radio. Of course daypart variations, peak vs. off peak, were included as well. In other words, systematic variation of plan density and dispersion is, in itself, a complex project. Not surprisingly, the results show a range of difference between random and real PPM net reach levels. Figure 19 shows preliminary results for the total teen demo. Figure 19 SCHEDULE DENSITY/DISPERSION MIX AFFECTS EXTENT OF RANDOM OVERSTATEMENT 1 9 Teens' Schedules by Density Reach with Random Dup 8 Reach 1+ 7 6 PPM's "Real" Net Reach 5 4 3 WB Prime Cable NBC Day 3 dprts/2 4 dprts /1 Med TV Pr ime NBC Med ium Da ytime Light TV +WRTI- +Radio +Radio stns stn +Radio +Radio Pr ime TV +Radio +WRTI- FM +Radio +Radio +Radio +WRTI- FM Schedule Type FM Despite the abbreviated labels giving the merest hint of the actual schedule variations, the results indicate a range in the extent of random overstatement vs. PPM s real duplication. There is wide-ranging variation depending on the specifics of the schedule density and dispersion. According to the preliminary results, these variations are influenced not only by density/dispersion but by other factors such as demographic group, the size of the schedule and the proportions of radio to TV weight. As with the plan variations above, there
Real cross media intelligence for real cross media planning 19 appear to be complex interactions among all these variables that affect the net reach and the overstatement of the plan. The correct estimation of a plan s reach thus appears to depend upon the interaction among this entire set of variables density/dispersion, demo target, schedule size, the proportion of radio to TV weight. More data analyses are underway at the time of this writing, specifically seeking to understand the disparities between real PPM net reach levels and those produced by random duplication. The significance of understanding is particularly important to U.S. planners who rely on syndicated models to determine the reach of mixed-media plans. CONCLUSIONS The authors previous paper (McConochie and Uyenco, October 2) demonstrated that current U.S. models for estimating local TV and radio reach produce enormously varied results. These estimation systems have built upon models using non-continuous demographic audience data, i.e., paper diaries which track either viewing or listening for one-week. Furthermore, these random duplication methods provide planners the only way to estimate the reach of a mixed-media plan. Clearly it is only through databases such as PPM that we can understand the appropriateness and limitations of random-duplication estimation for mixed media deliveries. Our work thus far suggests that current models not only over-estimate reach but do so to varying degrees, depending on the specifics of the plan. Further work needs to be done to determine how various schedule characteristics affect these differences. ACKNOWLEDGEMENT The authors would like to acknowledge the contributions of Kevin Killion, Stone House Systems, Inc., and Chris Heider, Arbitron, Inc. in preparing this paper. FOOTNOTE 1. More details on PPM methods and technology are available at Arbitron s website, www.arbitron.com
Roberta M. McConochie, Beth Uyenco REFERENCES McConochie, R.; B. Goerlich and S. Stinnett. (3). 21st Century Measurement of 21st Century Media. In press for ARF/ESOMAR. June. McConochie, R., T. Jarvis and L. Dupree. (2). Sex, Drugs, and Rock n Roll. Proceedings of the ARF/ESOMAR WAM, Cannes, June. McConochie, R. and B. Uyenco. (1). Contributions of PPM to Cross Media Planning. Proceedings of the ARF WOW, Chicago. McConochie, R. and B. Uyenco. (2). Trash & Treasures of Cross Media Duplication. Proceedings of the ARF/ESOMAR WAM, Cannes, June. McConochie, R. and B. Uyenco. (2). Trash & Treasures II: Improving Media Plans Deliveries. Proceedings of the ARF WOW, New York, October. Papazian, E. (2). TV Dimensions 1. New York, Media Dynamics, 2. Patchen, R. (2). Arbitron s Portable People Meters Update. ARF. NY October. THE AUTHORS Roberta M. McConochie is Director of Consumer and Industry Trends, Arbitron Inc., United States. Beth Uyenco is Director, Communication Insights, OMD-Midwest, United States.