The first major research study presented at Country Radio Seminar, which runs Feb. 28 through March 2 in Nashville, offered insight into the kind of behavioral research programmers will be tapping with the advent of electronic audience measurement. Speaking to a packed room on opening day of the convention, Jon Coleman, founder of research company Coleman, demonstrated how listener behavior is affected when new music is played on the radio.

Conducted in Houston from May to November 2006, the study is the musical
companion to research presented last year by Coleman that examined how
listeners react when commercials come on the radio. In both studies, Coleman
cross-referenced minute-by-minute listening data from Arbitron's Portable
People Meter with monitored airplay from Media Monitors.

Involving just one station -- CBS Radio country KILT -- the seven month study monitored audience behavior from 6 a.m. to midnight, in response to 47 new songs the station added during the period. It compared the station's lead-in audience in the minute before the song aired to audience levels at various points in time while the new song was playing.

Naturally, the roomful of country programmers and Music Row label execs was
eager to learn whether new music increases, maintains or undermines

In the very first minute of the average new song, the KILT audience remained
basically flat, posting an increase of 0.2%. Coleman concluded that listeners are "sticking with the station and perhaps going into an evaluation mode" during minute-one. In the second minute the audience grew by 1.8% and by 2.2% in the final minute of the song.

"This says, generally, we don't have to be overly obsessed with adding new music," Coleman said. "However there are songs that do well and there are
songs that do poorly," he added.

The bulk of the results Coleman presented compared audience size in the minute before the new song aired to the second minute the song was on.

New songs from established artists performed substantially better (improving listening by 3.7%) than new songs from new artists (-0.2%). Up-tempo songs did considerably better (+4.1%) than slower titles, which had no impact. Mid-tempo songs increased KILT's audience by 1.9%.

While new music impacted audiences differently in different parts of the day, Coleman said it was difficult to attribute the changes solely to new music. For example, KILT's audience spiked by 5.3% when the average new song aired in morning drive. But that daypart is when listeners are turning on their radios en masse. The inverse is true later in the day, and new music audience impact percentages parallel those natural lifestyle listening patterns: New songs on KILT improved listening by 2.3% in middays but caused minor declines in afternoons (-0.7%) and evenings (-.2%).

Songs of all tempos performed better as they became more familiar to listeners. Up-tempo songs played 300 or more times increased KILT's audience by 7.0%, familiar mid-tempo songs by 5.1% and familiar down-tempo tunes by 1.3%.

The average new song's positive impact on audience behavior peaked in the
neighborhood of 400-499 spins, which generally increased audience size by
4.1%. At 500 spins, new songs still had a positive audience impact, but
only by 1.3%.

There were extremes in the study. One song caused KILT's audience to decline
by 25% and one caused a 25% spike. Yet the bulk of new music impact ranged
from +8% to -5%.

Attempting to quantify the impact of new music decisions on KILT's 12+ ratings, Coleman said that if programmer Jeff Garrison made perfect music decisions, KILT would have a 3.8 instead of a hypothetical 3.5, while poor decisions would net a 3.3.

"Playing new music in 2007 does not appear to radically change the audience
behavior positively or negatively, assuming we're testing and playing the right music," Coleman concluded. "The ratings needle is moved more by dramatic programming and marketing changes than by just the music."