The marketing director’s question came at the end of a routine media presentation. “That’s all fine,” she said, “but how many people are going to see the ad? And how often are they going to see it?”
In essence, the advertiser wanted to know the campaign’s reach and frequency. In the early days planners got the answer from a set of graphs that worked for all television schedules. Today they must choose among several different systems that have their own strengths and weaknesses.
Three broad classes of tools are used to estimate TV reach. Deciding which is best depends on how much precision is needed and how much is known about the plan being evaluated.
* Persons Cume Study (PCS)-the least precise but best for general planning: Usually, planners don’t know much more than the target audience and the number of gross rating points (GRPs) in each daypart. They might also know which networks will be used, but not the individual shows. Nielsen’s PCS was designed for this sort of general planning.
Studies were conducted over four weeks in November and December (skipping Thanksgiving) in 1996, 1999 and 2001. The PCS provides the raw data for IMS Telecume, Telmar TVPlan, Stone House T-View and FCB’s Reach-2000. All PCS-based systems allow planners to create custom target audiences by age/sex and various household characteristics. They estimate schedule reach and frequency nationally and in spot markets, but planners must create groups of markets to ensure adequate sample size. The PCS is the only source of spot TV reach.
All of the systems allow planners to create customized dayparts that reflect an advertiser’s particular buying strategy. Dayparts can be defined in terms of individual broadcast or cable networks and groups of programs classified by the median age of the viewers.
Alternatively, planners can use prefabricated “generic” dayparts that lump together all broadcast networks and all cable networks. They are convenient for advertisers who use a broad range of networks, but they overstate the reach of more typical buys that are limited to those most demographically targeted.
Systems using the PCS are fast, easy to use and accurate enough for most general planning, but they have serious limitations. Although fringe and news programs can be selected individually, prime time and daytime are based on groups of shows. The studies cover November/December only, so there is no measurement of the second season or sports such as basketball and baseball. They are not available for at least a year after the reporting period, preventing analysis of current programs. Finally, all estimates are based on models using curves or personal probability calculations that are typically accurate to within 3 to 4 reach points (modeled estimate vs. a count of Nielsen respondents).
* Person-by-Person Tape- more precise, but requires more knowledge: When planners need to estimate the reach of current programs, complex custom dayparts or campaign periods other than November and December, they use systems based on Nielsen’s Person-by-Person Tape, which comes out 52 weeks a year, two weeks after broadcast. These systems include KMR X*Pert, IMS Optimax, Telmar Transmit, Stone House T-View and proprietary systems at some ad agencies. Although they provide reach/frequency estimates, they are best known for their optimization capability that identifies which mix of dayparts offers the most reach for a given budget. On the other hand, these systems are limited to one minute per quarter-hour (the mid-minute). Estimates are based on models, not on actual tabulations of exposed viewers. There is no way to break out spot markets. And finally, although the systems can handle individual programs, they work best with dayparts composed of groups of shows.
* NPower-the most precise analysis of spot-by-spot detail: The most precise national reach estimates are obtained from the reach/frequency module in Nielsen’s NPower system. Planners enter schedules spot by spot, detailed down to the date and minute of each telecast. Although keying in large schedules is a tedious process, NPower is by far the most accurate because it literally adds up the weighted number of Nielsen People Meter respondents who were watching during the specified minutes. It is the only way to answer questions such as, “What is the difference in reach between concentrating four units in the first half-hour of a show vs. scattering them over the entire telecast?” It also provides “only-only-both” analyses of two-program reach, tabulated frequency distribution and other statistics that are unavailable from the less precise but easier to use systems.
Roger Baron is senior VP, media research director, FCB/Chicago.