Measuring Reach on Radio

by Peter Kowalski

-- Editor's Note: The following was provided in response to a client's query, and it makes a good blog post, too.

Radio broadcast content remains among the most difficult to record, archive and retrieve, and of the traditional news media, is the least frequently (and accurately) monitored. Radio content is tracked using speech to text systems, which rely on voice recognition software that is often error prone. The sounds involved in the key phrase "Red Cross" are fairly unambiguous, and I would anticipate that your organization would have less trouble than something initial-based like the ASPCA, or something based on a proper name like Susan G. Komen for the Cure. As opposed to print or Internet monitoring, though, we would put very few restrictions on a radio broadcast search, instead relying on human readers to filter through segments for information that may or may not be pertinent to a particular investigation.

There are four major providers for radio broadcast content:

-- Cision tracks major markets and national news shows, but in my experience with using Cision in a limited capacity for a higher education client of ours, I would not recommend using their services for radio monitoring. Their accuracy is extremely low.

-- VMS tracks national networks (e.g. NPR, PRI, ABC Radio, Dow Jones Radio) and stations with news, talk, business and financial formats in the top 20 metros. They admit to the challenges inherent in radio monitoring, but an independent study found them to be 40% more accurate than Cision in retrieval.

-- National Aircheck is a niche company, and has monitoring specialists who record and monitor news, talk, business & financial radio across 2,500 stations (all major FM + almost all AM in the top 50 metros). As they specialize in radio, they have a large client roster that includes many of the largest public relations firms.

-- TVEyes is one of the newer entrants into this space, and tracks a handful of the national networks (Bloomberg, NPR, AP Radio, CBS News Radio).

Arbitron data is most commonly used to provide estimates for reach. In an integrated study, I tend to prefer Cume Persons, which is the average number of different people who listened to a station for a minimum of five minutes in a quarter hour. These persons are counted only once, no matter for how long they listen (so say, they listen for 45 minutes). This differs slightly from Average Quarter-Hour Persons, which is more useful for estimating reach in a rotating ad spots in a specific daypart or timeframe. Since we know when the news item aired, we would use Cume Persons, as it is comparable to other metrics like Unique Monthly Visitors (Web) or Impressions (Print). If you are new to Arbitron data, a handy reference guide can be found here. This data is often segmented within demographic groups, but the extent to which each provider can provide this level of detail would have to be more closely determined.

As with print and online data, radio reach data would be enriched with information about position within a particular show, dominance within a segment and other measures of prominence and dominance that are specific to radio and relate to how individuals remember aurally-received news.

Data providers are slowly getting to the point where segments can be listened to online, which is always preferable in analysis as tone of voice, pacing and similar factors can alter perceptions of information. It took TV and cable monitoring services some time to get to the streaming video point, and as streaming audio becomes more widely available, I consider it an asset to analysis.

Peter Kowalski is Director of Research Strategy at KDPaine & Partners. His research interests include international inter-media influence, agenda setting and network analysis. He holds a Bachelor of Arts in Public Relations from the USC Annenberg School of Journalism.