from Jay S. Guyther

Note: This paper was originally presented at the European Radio Symposium on November 4, 2009 in Istanbul, Turkey.

Part 4: Required - Lower Costs and Larger Samples

I’m a big proponent of portable meters for radio audience measurement, having been at Arbitron in 1992 when the project was initially announced, later responsible for the first tests in the U.K. and subsequent international licensing and more recently having co-authored the U.S. business and implementation plan. The advantages of more frequent and timely reporting of data, the availability of more granular data for programming decisions and the increased accountability with advertisers and agencies are undeniably tremendous benefits for radio.

However, while Arbitron, TNS, BBM and GfK have successfully commercialized portable meters in a number of countries, the implementation has not been without its challenges and detractors.

The differences in audience estimates between electronic measurement and either dairy or day-after-recall are real and can not be trivialized. Audience estimates are the currency for radio advertising transactions and have significant impact on both buyers and sellers, often times having a personal impact on an individual’s income. However, new and different measurement systems should yield different results. Otherwise why make the change in the first place. Portable meters measure exposure as opposed to recall. More accurate measurement often captures exposure to stations that goes unreported in recall methodologies. Time spent listening to stations often decreases since people tend to overestimate their actual listening in diaries. Continued education and training is imperative and goes a long way in helping to recalibrate the currency.

As radio audiences continue to fragment, there will be an increased demand for larger sample sizes. In the U.S., Arbitron’s PPM panels are one-third the size of the previous quarterly diary sample. And while the PPM panels yield the same or greater statistical efficiency, many are concerned that narrow demographic analyses are no longer available since they require an absolute minimum number of panelists. The only way to significantly increase samples is to reduce the equipment or operational costs, while not sacrificing quality. Reduced costs could be achieved through improved compliance, but in actuality, compliance levels are probably already as good as they will get. Data from GfK shows that 88% of participants wear the Mediawatch for 10 hours a day and Arbitron data indicates on average, 75% of panelists carry the PPM a minimum of 8 hours a day.

A significant cost reduction could come from incorporating the detection technology in a consumer device, such as IMMI and IPSOS have done with their cell phone meters and Arbitron has long range plans for developing. However, as noted earlier, there are complex research and technology issues to be resolved such as microphone sensitivity and battery consumption. Improved compliance is also not assured given that current research indicates people do not always keep the cell phones with them, especially when at home.

As media companies find new and creative ways to create and distribute audio content, comprehensive measurement will become more difficult. Both encoding and matching will face challenges reporting ‘personalcasts’, or personalized content created from on-line libraries and sent to an individual’s mobile phone. Methods to encode Internet-only audio streams or collect signatures for matching when audio is distributed via ‘closed-loop’ platforms, like podcasts, must be developed. Media consumed via mobile phones and wireless/Bluetooth headsets could also be missed unless an effective way to intercept the audio is invented.

And while all the portable meters can measure both radio listening and TV viewing from the same panel, multimedia measurement will likely mean compromises. Questions to be asked and debated include:

• Is hearing the TV audio an acceptable definition of ‘viewing’?
• Will TV audience measurement continue to require a wired connection to the traditional set or PC to measure other on-screen activities?
• Can the very different interests and requirements of radio and TV broadcasters be satisfied with one measurement system?

It’s been 17 years and $200+ million in development costs since the concept of a portable meter was first introduced. Today, there are at least six legitimate systems competing for the attention and acceptance of media researchers. Portable meters have proven to be flexible in their implementation and reliable in operation.

The transition to electronic measurement for radio has been long and challenging. Broadcasters and their advertisers are beginning to realize the benefits of more frequent and timely reporting and greater accountability. But increases in radio advertising revenue will be necessary to justify the costs.

For continued success, development of new generations of meters must be continuous and come much faster. The costs, both the equipment costs and the system operating costs, must be greatly reduced so sample sizes can be substantially larger making electronic measurement  more attractive and affordable to more markets.   

The ability to efficiently and inexpensively embed signal identification technology into a consumers existing personal electronic device, while maintaining the necessary research quality, is the only solution to building large cost-effective panels. That doesn’t yet exist. Let’s just hope it doesn’t take another 17 years to get there!

{ Comments on this entry are closed }