AndoMedia presented a preview of new metrics at RAIN Summit East last week. In keeping with the format of the Summit, Patrick Reynolds of Ando gave us a few snapshots of data. The online listening audience measured by AndoMedia grew from 204 Million in May of 2009 to 234 Million in August of 2009 – and according to Reynolds, that growth was attributable to increased listening, not to increased number of stations being measured. After the presentation he also told me that number does not include Pandora’s audience, which would surely have an impact.
The average Internet radio listener streams nine sessions per week, and 77% of listeners stream every week. I particularly like that last stat. An enormous percentage of Twitter users signed up, tweeted a few times, and have yet to return to the medium. Twitter retains only about 40% of its users from one month to the next. The fact that streaming audio is keeping ¾ of its audience active on a weekly basis is very promising for continued growth.
Ando is about to make significant changes to its measurement, and Reynolds talked a little about some of the new terms they will be using. Replacing Average Quarter Hour will be the term Average Open Sessions, which will count all sessions of at least a minute in length. This unit is closer to the actual data they are collecting and requires less manipulation. It’s also more in keeping with terms used in other digital media metrics.
According to Reynolds the changes have been blessed by MRC, the board they are working with for accreditation of Webcast Metrics. He promises a new ranker soon, following a four month hiatus while they worked through some of these transitions.
Recently there was a piece in Radio Business Report (here) about the need for Internet radio to create an industry governing council that would coordinate a voice for the industry on certain topics. One of the mentioned topics is audience measurement. There are at present two companies providing measurement for Internet radio stations, and the methods vary significantly.
Interactive measurement standards
One of the greatest assets that Internet radio has is the fact that it is an online, interactive, highly measureable media. All of its audience is online and can be motivated to interact with the advertiser. Audience can be precisely measured, allowing for greater accuracy in campaign planning. Impressions can be measured and accounted for exactly as delivered, eliminating the need for posting.
This advantage is the key to acceptance of Internet radio. Educating research departments, buying departments, and advertisers about this advantage, and providing them with the language and tools to evaluate Internet radio from an informed perspective, may be a challenge, but trying to simplify it by making Internet radio look and smell like traditional broadcast radio is short sighted. Radio’s slice of the pie is shrinking – and certainly some of the reason for that is its lack of measureability. Internet radio should be sold, bought, and priced based on impressions delivered just like all other interactive media. The most precise measurement tools available should be used for this purpose.
Two Companies, two methodologies
Currently there are two companies measuring Internet radio – Comscore and AndoMedia. Arbitron purchases Comscore data and releases it under their own name (or brand). Comscore’s data comes from their panel based research on website usage. The sample is small – Comscore’s monthly panels of web users are often as small as 80,000 persons, but even if they are 100,000, 21% of those persons are streaming on a monthly basis (according to Arbitron’s Digital Radio Study). That means on a monthly basis they’re watching what about 20,000 persons are streaming. With a sample that limited, they’re going to be able to measure broad network listenership, such as to Yahoo Launchcast, or the entire CBSRadio streaming network. They can’t measure the audience of a standalone station or smaller network. They just won’t be able to get the qualifying intab for measurement. They also won’t be able to detail listening within any network to specific channels.
AndoMedia is a relatively new independant company that provides third party server based measurement to subscribing stations. Many large companies are subscribing to the data including CBS, Citadel, Katz, and Clear Channel. Data is converted into radio metrics such as AQH, Cume, TSL. It’s been a sticking point for Arbitron’s advocates that demographic definition is added to the raw AQH and Cume by using national format profiles that assign percentages to age cells and genders, and therefore it’s also an estimate. That is true. However, isn’t an estimate that is derived from actual server stats more reliable and preferable to an estimate that is not? I’ve spent hours with Radio buyers, supervisors and Research analysts who tell me it is.
Each methodology has pros and cons. One provides panel based data with the recognizable Arbitron name, and the other provides more precise station by station measurement from a Web 2.0 startup. Agreeing on a measurement methodology that can unify the message would benefit all of Internet radio.