The challenge of determining KPIs for audio
Updated: Nov 17
Any talk about audio, whether it’s the sheer volume of audio content today or the potential of audio advertising, carries some sort of uncertainty (I was tempted to say stigma but that would be too harsh) because it inevitably reaches the topic of measurement. The lack of standardized KPIs for audio is seen by many as a stepping stone for further advancement of audio.
Even though it’s an old medium, audio in its digital form has only exploded in the past couple of years so measurement or even attribution are still in their early stages. The way audio is measured now, especially the advertising part, is highly inaccurate.
Again – why?
The current situation is such largely because the audio measurement strategy is based on basic metrics from video and display such as viewability (audibility – did the user listen), completion rate, and the all-present CTR. Sometimes I think if the person who “invented” CTR received royalties for it in some way, he/she would be a trillionaire by now.
More like “trillionaire”.
For this reason, many publishers and advertisers are reluctant to get fully on board with audio advertising. Plainly speaking, the buyer and seller aren’t speaking the same language at the moment.
Since this is a fast-developing industry, there are only a few systems/solutions that tend to this issue. The brunt of current technology is extremely limited in terms of what capabilities you can get from measuring audio, and that means the majority is heavily reliant on CTR.
The problem with CTR is that it’s pretty much an antiquated metric. Most of the time, there is a display banner that’s paired with audio, and the only way to get any sort of information as to who listens to what or when is to tag the paired display banner with the audio.
As you can imagine, this is highly inadequate because most of the people likely won’t be clicking. Basically, you are hoping most of your listeners are going to click the banner so the only real measurable part of that equation is the people who do click on the banner, which would be your click-through rate.
The year-over-year percentage of mobile listening is increasing steadily, while listening on the desktop is actually decreasing. Subsequently, two things come to mind. First of all, there are fewer people who are clicking banners on phones. Second of all, when they hear the audio ads, a lot of them aren’t even looking at their phones in the first place because they are in their pockets.
The data I am referring to stems from the eMarketer report. Historically speaking, desktops and laptops have been dominant in accessing digital audio content but the tides are shifting. The July report shows 2018 having a 0.4% drop (73.5% of the total desktop digital audio listeners in the US) for the first time, while 120.3 million mobile listeners in 2017 grew to 127 million in 2018 (63.2%). Both trends are predicted to stay on the same course in the next few years. By 2021, exactly two-thirds of US digital audio listeners will access content via a mobile device, as opposed to 68.8% via desktop.
Setting a new standard
In the sense of digital marketing KPIs, the current measurement practices are problematic for another reason: audio without any paired visual element provides nothing to track. That’s why determining KPIs for audio is so important in allowing more brands to buy into audio.
Roughly how interest in audio looks like most of the times.
Here’s a peek at how we at Trinity deal with this issue. Usually, when a brand chooses to advertise on a podcast or streaming app, the amount of data they get is scarce, meaning there’s very little user data, the type of content, etc. All the data they are getting is based on the platform they are buying from, with minimum ability to use their technology to collect the data.
We offer full transparency in terms of data an advertiser gets, meaning information from display and video is the same for audio (the usual stuff + additional parameters like user data, device, user agent, and many more). This helps us determine the effectiveness of a brand campaign based on the usual audio KPIs crossed with the relevant parameters mentioned above.
Would a standardized set of metrics help speed up and smooth the process? Absolutely.
Right now, there are few attempts at standardizing that largely revolve around podcasting as the audio’s big story. One is IAB’s podcast measurement certification program that focuses on server-side data, with future extension to client side measurement in plans. Another is NPR’s Remote Audio Data (RAD), a method for sharing listening metrics from podcast applications straight back to publishers. A URL is inserted into an audio file that allows players to insert specific points/triggers that ping the server when a listener gets through them (e.g. an ad or sponsorship message).
On a similar basis, there are a few other companies working on setting a new industry standard and providing a new set of KPIs for audio ads. So far, the path has been laid out to defining standard metrics everyone can rely on for accurate information.
These are the latest attempts to figure out a universal set of metrics that everyone can get onboard to measure listening. The idea is to move away from CTR and implement LTR – listen-through rate. The way I see it, that just makes perfect sense as to where the industry is going.
With these significant movements, there’s hope the industry will soon get universally adopted metrics or at least, more universally adopted metrics so there are fewer (ideally none) tradeoffs that publishers face right now. As listening continues to grow, so do advertising investments. In turn, the need for a united contextual insight into who is actually listening and what exactly they are listening to increases. Here’s to making it happen.
Make sure you’re following me on Twitter for ongoing updates, tips, and industry takeaways!
SOUNDS GREAT, RIGHT?!
Subscribe to our newsletter for industry updates and subscriber only content
We promise not to spam
This field is for validation purposes and should be left unchanged.