One of the most exciting things about working with research on real listeners is the exploring the different kinds of currency you can create, and exploring where you perform within that new space. Anybody that has shuffled up their questionnaire with a new type of question bank or switched consulting companies knows what I’m talking about.
The online world presents us with a plethora of new metrics and KPIs. Not quite as dynamic as they were initially, they are still evolving and changing. You have to be in tune and know what those numbers actually mean. Furthermore, continually review your interpretation of those listening numbers as it is essential to build a successful product. Remember when you used to measure your website by pageviews? That has been replaced by click rates, hits, bounce rates, and so on – rightfully so! We understand more about how online listener behavior works than we used to.
In the same way you shouldn’t be measuring your online audio success simply by number of visits or tune-outs.
Try to dig a little deeper.
Leaning to Count
Historically, radio-industry metrics have either been largely based on image and recall (diary or CATI), or subject to volatile swings due to panel sizes (PPM). Both methods do have something in common. They use a number of clean-up measures to make the listener data more uniform and more reliable.
Online audio, for the first time, gives the radio-industry a chance to measure something directly. Even if it is just a (ever-growing) portion of your audience’s listening minutes. Notice I did not say “a portion of your audience”, because that would be wrong. It’s important to remember that you don’t have “digital listeners” and “FM listeners” because you don’t. You have listeners that use different devices in different situations to listen to your station.
Trim the Trash
Cleaning up the online listening numbers turns out to be just as much of a chore as it was with more traditional research. One of the first things we discovered in the early years of RadioAnalyzer was that not every connection is a real listener, and not even every real listener should be counted when you are trying to measure content impact on the air. Drilling down to the listeners that matter is an essential part of getting a useful read on your listener habits.
A while back a client called me with a question. His online dashboard from his streaming provider was showing him a slightly different listening curve during that day than we were; he wanted to know why.
Since we knew the numbers we were delivering were correct – we double checked – we dug deeper. It turned out that around 50% of what was being reported as listening in his stream dashboard was actually junk.
The Right Tools for the Right Job
The way his dashboard was compiling the numbers also added to the discrepancy. Due to grouping total connections over time in a way that didn’t make a lot of sense for measuring program content. Was the stream provider negligent in reporting the numbers? No, they were compiling them in a way that made sense for them. From the technical perspective of running an audio-streaming server.
At RadioAnalyzer we have worked out filtering methods and algorithms to optimize the data for true usage scenarios by real listeners. Additionally, we continue to refine those methods all the time. It’s an important step in using our tool to improve your station and increase the time your listeners spend there. There are other methods that make sense if you are looking for sales metrics, or a read on the technical performance, and that is fine. Just don’t confuse that with actual listeners, and true listening behavior.
Know Your Toolbox
We’ve come along way in radio research over the past decade. The increased competition and diminished returns are forcing us all to reevaluate the data we use. And how much we are willing to pay for it. Digitally collected data is getting cheaper, and traditional data ever more expensive. It’s forcing us out of our comfort zones into new collection and analysis methods. This is a great thing if you have freedom to experiment and time to do it in… but when was the last time you met a PD that had even one of those, much less two?
The comfortable transition would be to take the new numbers and have them to perform exactly like the old ones, but it doesn’t work that way. I’d say we spend a good 50% of our time in the first few months with a new client helping them understand how the numbers work and where to look to best evaluate their product. We even go so far as to end our first training sessions with “Don’t ask us how to use a function, but come to us with a question on how to evaluate your on-air performance”, because that is what any kind of research should do for you: answer your questions.
TL;DR: pick the right tools for the job at hand, find out how to best use those tools to create the results you want.
What’s in your toolbox and how do you go about making great radio on a day to day basis? I’d love to hear from you, just send me a mail!
You might say the poor guy never had a chance, being born to two radio-crazy ArmedForcesNetwork journalists that met in southeast Asia (think “Good Morning Vietnam” – and no, his dad is not Adrian Cronauer). Since discovering his love for music programming as a teenager, Bill has been obsessed with turning great ideas into numbers you can measure, and measured numbers into actionable programming strategies that make great radio. After touring Central Europe as a Music-Promotion-Programm- Director and a Research&Program Consultant for European and International Consulting firms, he has returned to the radio innovation trenches at RadioAnalyzer, and is loving every minute of it.