Ad Results Media CEO Kurt Kaufer recently sat down with Ashley Fitzgerald, ARM's Director of Analytics at Ad Results Media, to discuss the current state of how podcast advertising is measured within Media Mix Models (MMM) and why, as an industry, we need to work to improve and standardize our methodology. Podcast advertising has been on an enviable growth trajectory the last few years – in fact, according to the IAB, revenues are expected to double from 2022-2025 to an estimated $4 billion.* While this rapid growth is good news for all of us in the podcast business – it is not without its challenges. The current media mix modeling studies tend to underreport the effectiveness of podcast advertising because of technical nuances and an overall lack of standardization. This type of refinement to ad measurement often comes with time and maturity. Ashely Fitzgerald breaks it all down and also explains how Ad Results Media can help.
Kurt: We have a number of clients that know that podcast advertising is driving significant conversions but when they get the results of their MMM studies back – it looks like podcast advertising is not as effective as they know it to be in reality. Why is that?
Ashley: It all comes down to how podcast consumption is measured. Right now, a podcast “listen” is calculated when a podcast episode is downloaded from the server – not when it was actually listened to. Many people will download a podcast episode and then listen to it days or weeks later. I have spoken to several providers of MMM studies - and they start measuring the impact of podcast ads the minute the ad spend occurs, and some have a standard attribution window of as little as 4 days. Oftentimes the consumer hasn’t even listened to the episode by then – much less completed a purchase.
So what you are saying is that many MMM models start tracking the effectiveness of the ad before it was ever heard?
Yes, that is correct. Podcast advertising is a channel where ad exposure doesn’t trigger the model – only a proxy action like downloading the episode does. I think that it is really important for all of us to use different ways of measuring a “listen” that are more accurate as well as establish longer attribution windows that are more in-line with how podcast advertising drives purchases. Doing these two things will help make the results of MMM more accurate with regards to podcast effectiveness.
What are some other ways that we can help advertisers measure when their ad was heard by their target audience and make the results of these studies reflect real world performance?
The key to this is using measurement methods that guarantee that an individual has been exposed to the ad. Podcast advertising doesn’t have the same type of standardized measurement as digital or social channels where the ad exposure is captured as an impression. Without a standard metric like the impression, we need to do the next best thing. That is where Ad Results Media can really help – by providing different and more accurate data sets to MMM providers running studies. Since the main challenge here is to nail down the timing that a consumer heard the ad – we need to use the tools we have at our disposal like post-purchase “How did you hear about us?” survey data, pixel-based measurement data, promo code redemption and vanity URL traffic. These are all concrete measurement methods that indicate that a consumer was exposed to your ad and made a purchase or visited your site. At Ad Results Media, we work with our clients to set-up this type of measurement and can help pass it back to the MMM providers.
The second update in the models you mentioned is the attribution window. Talk to us more about what that means.
As I mentioned earlier, some MMM providers use a 4-day attribution window with steep degradation over the course of those 4 days. With audio – there is no immediate click to purchase in the same way there is with digital or social. We are asking people to remember the brand or offer and use a promo code or vanity URL – this is not done instantaneously like so many impulse purchases are from Instagram ads for example. Yet, the average MMM attribution window for Meta platforms is 7 days – even longer than that given to podcast ads. So I recommend that we extend that attribution window to 14 -30 days – customized depending on the purchase journey for each brand or product. Longer for more considered purchases and shorter for everyday items.
Well it sounds like there is some work to do but there are tangible and immediate ways to get more accurate with our MMM methodologies.
Agreed! That is what I love about my job. We get to really dig into the data and models and help our clients measure what matters…more accurately.
*U.S. Podcast Advertising Revenue Study 2023.