Friday afternoon, Dana Rehm, NPR's Senior Vice President for Marketing, Communications, and External Relations forwarded a link to a graph used on the washingtonpost.com site and asked me to take a look at it. This graph immediately fell into the category of "things that make you go.....hmmm."
While we in research love to tell a good story when we have one, the market share data displayed do not jibe with anything I've seen. The graph shows that NPR was listened to by more than 25% of the available audience in 2008 and portrayed a 10-year growth rate exceeding 100%. Wow! I like that story, but in 2008 NPR programming was heard by 10% of adults on a weekly basis and showed a 10-year growth rate in audience share of 45%. Not a bad story, but more modest.
So, super-sleuth Dana traced the graph back to its source: Jon Peltier. In September 2009, he posted a "chart busters" column critiquing the visual display of NPR data in a Fast Company article. As part of the critique, and he makes several good points, the author generated a market share of his own design in order to demonstrate a point about how to effectively highlight audience growth graphically. The NPR market share data, however, are not based on actual share or ratings as provided by the radio estimate provider of choice: Arbitron. Given the lack of connection to the industry accepted measure of radio audience, it is my sincere hope that this particular graph is not quoted in the future.
All that said, when I took a closer look at the original audience growth graph posted in the Fast Company article, I saw a less than entirely accurate comparison. Oh, why must I be thorough on a Friday afternoon? This graph indicates a 1998 NPR programming audience of 13.5 million and a 2008 audience of 25.4 million. But wait! I see an asterisk on the 2007 data. In addition to regular programming (30 minutes or longer), the 2008 figure inclues listeners to newscasts, and here's the problem - the 1998 figure doesn't! NPR began collecting newscast audience data in 2004, and we cannot reliably collect data prior to 2004. Contrasting a 2008 figure with newscasts to a 1998 estimate without newscasts isn't an apples-to-apples comparison and as such artificially inflates the growth rate. As the chart below shows, the 2008 audience should have been quoted as 22.4 million for the purposes of making a fair comparison to 1998. Lesson du jour - please triple check your audience "facts" with your friends at NPR.
Lori Kaplan is the Director of NPR's Audience Insight & Research group.