Showing posts with label internet radio statistics. Show all posts
Showing posts with label internet radio statistics. Show all posts

Monday, August 23, 2010

In Trenches of Audience Metrics



As published in Talkers Magazine - June 2010

In Trenches of Audience Metrics
Internet radio lights a way for talk radio
By Linda Mackenzie

LOS ANGELES ––Numbers! Numbers! Numbers! Shrouded in mystery of sample sizes and extrapolated numbers has there ever been a real count of audience size? It’s the never ending bane of radio – even in my arena as General Manager of an Internet-only talk network.

The real internet radio pioneers started broadcasting in the late 1990s. As a former 18 year datacom engineer with an 8 year stint as a talk radio host I started my internet radio network in 2002. At that time the majority of people knew very little about internet radio, however today, according to the Bridge Ratings Feb 1, 2010 study, Internet radio’s audience size is about 60 million a week.

Being in radio you can imagine the effort it took in 2002 to get quality talk hosts and listeners, much less advertisers. By 2004 when I started streaming 24/7 our audience numbers grew. Numbers now became part of my equation to get advertisers who were weaned on audience metrics. Being that old datacom engineer I set out on my quest to find real numbers.

I can remember being at the RAIN internet radio meeting held at the NAB 2004 Convention. The web-based radio audience size at that time was about 40 million a month. After the Arbitron presentation, I raised my hand and asked them a couple of questions which were perplexing to me at the time. “How can Arbitron determine numbers using an infinitesimally small sample size in comparison to overall audience size, how could these numbers reflect the listening audience of my particular talk station and why did I need Arbitron when as an Internet-only station I have actual numbers?” At that point the moderator came up to the podium and said, “We’re going to take a short break.” My mouth was agape as they walked off the stage leaving my questions unanswered amidst a plethora of murmurings from the attending audience. The speaker from Arbitron, who was extremely nice, came over to talk to me during the break and ended up by asking me questions.

Unlike any other broadcast medium, internet radio can get actual audience counts today.

This data is available on streaming servers. A streaming server can get the actual listener’s IP address, how long they listened from the time they log on to the stream to the time they log off, and even the zip code. All that information can be captured into what is known as the streaming server log. Administrators can specify how much or how little information they wish to capture into the server’s log file. All the way down to every last event or transaction that occurs.

However, processing the log to glean the desired information is quite time consuming and therefore expensive.

In 2004 our numbers were smaller and I received reports from my streaming aggregator displaying all this data. We had a 50 State breakdown, what cities in those States by listener count and how long they were listening. At that time we went to 108 countries with 82% of our numbers coming from 1,240 cities in all 50 states. We also had a count of our on-demand archived shows which are real streaming server numbers based on the individual retrieval of the listeners IP address. We were getting 8,000to 10,000 listeners a month on our archive.

Today with so many listeners there is so much data in any given day that it is impossible to process it. It’s too time consuming to warrant the cost. So at this point in time, even though the data is there, no one processes it. But it’s just a matter of time before the solution on processing these individual transactions is found and then we will get real numbers.

This leads us to a bigger question, if real numbers were available would we really want them and would they be ‘true’ numbers?

Internet radio has the capacity to reach more listeners than any other method of single radio delivery.

My network is distributed (carried or rebroadcast) on over 50 channels of distribution including the internet, WiFi, smartphones, private radio networks, mobile, podcasts, in dashboard of the BMW Mini cars and we even are on an Iphone App for an alarm clock.

As to ‘true’ numbers, now here comes the rub. On the streaming server, each one of these 50 distribution channels appear as one individual IP address. The distribution channel picks up my stream, broadcasts it out 24/7 to their listeners, they get the audience metric - which I never see. Take the case of my network being 1 of 300 content providers on the Microsoft Windows Media Player Radio Tuner. That makes us available on every Windows Media Player in the world, but I don’t know how many people are listening. I would assume we are getting a lot since they haven’t thrown us off, but the bottom line is that it brings us back to extrapolated numbers, at least for distribution channels.

Now let’s throw into the muddy waters of radio measurement, the advertisers. The advertising agencies seem to want ‘Page Views’ as their preferred measurement of decision for placing their advertising dollars. Page views are the website statistics of how many people request the website, people who may or may not listen to the radio stream. Most advertisers are unaware that the actual numbers of listeners come from the streaming server.

However it is still my opinion that any Internet audience metric is infinitely more accurate and better, especially for talk radio, than a sampled base audience metric system that is geared towards music.

Internet radio websites provide the ability to get real listener demographics

In 2007 I realized that although I knew the listener count and could somewhat glean from our show topics my station’s approximate demographics I didn’t have a male/female composition split or any real measure of the age group breakouts. So I initiated the Listener Clubhouse. People sign up for the Clubhouse, answer a few demographic questions and I get a view of my actual demographics which I can apply to my overall listener numbers. Surveys and polls can also provide a method for audience demographic accumulation.

Talk radio hosts interaction with listeners is a real source of audience measurement.

When all is said and done the most important key to audience measurement are the talk show hosts themselves. As General Manager I know my hosts are the core to my whole operation whether this is reportable or not. Our hosts get the on-air phone calls, emails, mail, appearance requests and interaction with their unique listeners on the network’s radio stream and web site, the host’s website and both our network’s and their own social media sites. That’s a lot of potential for numbers. That’s a lot of feedback. The activity that we get and the phone calls, albeit the ones in the middle of the night, show me we are doing a good job at our first priority which is - the listener experience.

In these times where face-to-face and social interaction is being replaced by the computer, I believe the voice of talk radio is going to grow to fill that gap of this lost interaction. People need and relate to people – and that’s talk radio.
Automated music, which can be gotten in so many various forms on the internet other than radio, (and don’t get me started about .mp3 versions that are playing on air that cut off the bottom bass sound of a song) it is amazing to me how talk radio seems to be getting the short end of the stick when it comes to current measurement practices. Maybe it’s time to re-think the whole measurement company paradigm. The data is there we just have to find a way to process it.

Linda Mackenzie is the Founder and General Manager of HealthyLife.Net Radio Network (www.healthylife.net). She can be emailed at linda@healthylife.net

Thursday, May 27, 2010

The 8 Truths of Internet Radio Audience Measurement

Page views DO NOT measure listeners
A page view is a unit for measuring Web readership. It is the count of how many times a web page is requested. Streaming Servers measure listeners who listen to the audio stream. Not all listeners who visit a web page listen to the Internet radio stream.



Streaming servers DO NOT measure all listeners if station is simulcast and/or syndicated on other distribution channels
A streaming server measures listeners but if a station is connected to outside distribution channels like WiFi, Mobile, ITunes, private radio networks or Microsoft's Radio Tuner the stream that is picked up looks like one (1) listener to the streaming server when in fact it could be thousands.



An age old radio brand name DOES NOT necessarily mean they have more listeners on their Internet stations than an Internet Radio Streaming Station
Listeners have the power of choice and there are hundreds of thousands of web radio stations that reach their targeted niche markets. The terrestrial radio industry has only taken Internet radio seriously over the last 2 years. 24/7 Internet radio networks that have been in place for 7 or more years have more loyal listeners since they have been in the marketplace longer.



Internet audio listening has surpassed AM/FM terrestrial radio
According to the Arbitron/Edison 'The Infinite Dial 2010' study Internet Audio listening has now surpassed AM/FM terrestrial listening.



Podcasts DO NOT provide more listeners than an Internet Radio Stream
Internet streaming radio usually provides 24/7 live coverage with multiple hosts, on demand listening and podcast download options. Internet Radio's average listener TSL (Time Spent Listening) is 2-3 hours. A podcast is usually a 1-hour audio program available live at a specific time and/or made available as a download. A podcast's average TSL (Time Spent Listening) is 3-5 minutes at the top of the hour.



Outside audience measuring services DO NOT provide real numbers
Based on a small audience sampling for a particular genre, usually music, extrapolated numbers are used to determine audience measurement. For example, the Arbitron/Edison 'The Infinite Dial 2010' did a telephone survey of 1,753 people, age 12 and older, to provide statistical information about the radio music industry. Internet Radio has an audience of over 72 million Internet listeners a week. This sampling is much less than even 1% of the audience size and has no relevance for talk radio at all.



Demographics are suspect unless station has its own collection method
Each station is usually unique because of niche market genres and marketing efforts. Real demographic info comes from a station's data gleaned from their own data collection methods such as a clubhouse, surveys or polls.



Pre-roll video ads on Internet radio stations lose listeners
Americans and Internet Users - WANT IT NOW!. There are 100s of 1000s of Internet radio sites to choose from. If a radio station website delays the audio stream, which is what the listener wants and is expecting to hear, the listener will usually click off the site or get frustrated waiting for the audio stream to start, resulting in an unsatisfactory user experience.


Check out HealthyLife.Net Radio Network's Demographics Here: http://www.healthylife.net/RadioShow/CHSR Demographics.htm