As our lives become more and more digitized, podcasts have grown greatly. A study in 2022 showed that 38% of the US population listened to a podcast in the past month, over 3x the number from the decade before.
With such strong penetration, it’s no surprise that podcast advertising has become a must-have for many businesses’ marketing mix.
There are different podcast ad formats that fall under the umbrella of podcast advertising. These range from pre-recorded ads, to host-read sponsorships read by the podcast host, and so on. And then there are the formats: pre-roll, mid-roll, post-roll, or even native ads where hosts discuss the products being featured.
As with all other forms of marketing, measuring performance and understanding how revenue is being attributed across channels is critical for businesses to keep tabs on their marketing efforts. However, the podcast landscape is vastly different from other media. In this piece, we will go into detail on podcast attribution, with a particular focus on pixel-based attribution.
The challenges with Podcast Attribution
It is a common assumption that podcasts are digital or online—you’re probably listening to them on your phone or smart speaker, after all. Which means attribution should be no different than other digital marketing channels right?
It may surprise you to learn, then, that podcasts are actually an offline channel.
They are essentially radio shows that you download from the internet. Once it moves from the server to your device, tracking interactions and engagement post-download is limited.
Once offline, advertisers and platforms lose the ability to track listener activity (such as if they skipped segments, or your ads) post-download. Podcast attribution is thus different and more challenging than other digital channels.
Several methods have been introduced and used to try and track podcast attribution. These methods include coupon codes, vanity URLs, post-purchase surveys, and pixel-based attribution.
ADOPTER did an amazing job covering these in the previous piece on Podcast Attribution 101, but it’s our belief at Podscribe that pixel-based attribution and tracking is the closest we’ll come to tracking spend, impressions, and both top of funnel and bottom of funnel metrics connecting direct ad exposures to activity on the advertisers site or mobile app.
Moreover, pairing pixel-based attribution with the power of coupon codes and post-purchase surveys, is what we believe forms the ideal recipe for attribution.
The benefits of pixel-based attribution
Before we dive into podcast attribution, we must first understand how podcasts work.
Unlike other digital marketing channels and streaming, podcasts are accessed through a Really Simple Syndication (RSS) feed.
What exactly is an RSS feed?
It is a formatted text document that contains all the important information about the show. It's hosted on a server and (usually) has a public URL/link so anyone can view or access its contents.
It contains all the information about your show and episodes. This includes the show’s title, description, episode titles, links to the audio files for the episodes, and more. The RSS feed provides a way for anyone to subscribe to the show, or stream or download your episodes.
Without an RSS feed, podcasts would just be blog posts and audio files. There would be no way for people to subscribe and get new episodes unless they visit their website or get a direct download link.
RSS feeds are essential for the distribution of podcasts. However, this downloads the episode locally to the listener's device, complicating tracking and attribution.
Methodology for pixel-based attribution in podcasting
Pixels are the key to bridge the gap between who listened and who took action. Due to the offline nature of podcasts, one of the few data points available is the user’s Internet Protocol (IP) address as well as some others like the user-agent pairs (but those are less important).
The pixel-based attribution process involves grabbing the unique identifier of the listener (Prefix), and matching that to a unique identifier of a purchaser (on-site Pixel). It identifies listeners and matches them with purchasers, despite the absence of online cookies.
And if you’re concerned about privacy, don't worry. Given the nature of an IP address, it provides no Personally Identifiable Information (PII). As such, users will remain anonymous.
A prefix, a short URL embedded in the podcast's RSS feed, captures the listener's IP and user agent before the episode downloads. This is done through a redirect that pings Podscribe’s server and captures the data. It all happens in milliseconds and is completely unnoticeable.
This facilitates listener identification and is unique to podcast attribution.
Podscribe has the largest prefix coverage of all attribution providers due to the fact that we work with the largest buyers in the space, giving us the leverage and coverage needed to be the leading podcast attribution platform.
An onsite-pixel, similar to your google ads or facebook ads pixel is installed on the advertiser's site or mobile app and allows us to grab the IP of the purchaser or person who engaged with the advertiser. From there, Podscribe is able to match those IPs together and determine who was exposed to X and then take action.
There are some other technical things involved such as filtering out Noisy IPs (IPs where multiple people are on the same network) as well as leveraging a device graph (we partner with tap ad) to do cross device matching, which we will go into detail below.
Generally, it's really just about matching the IP of the listener via prefixes to the IP of the purchaser via onesite pixels.
High-level setup: Pixels & Prefixes
To leverage Podscribe's pixel-based attribution effectively, two critical components must be set up: advertiser-side tracking and publisher/campaign-side tracking.
-
Advertiser-side Tracking: This involves placing Podscribe's pixel on your website or mobile app to capture conversion events. This setup is essential for tracking the actions that listeners take after hearing your ad.
-
Publisher/Campaign-side Tracking: This requires integrating Podscribe's tracking solutions (RSS Prefix or DAI pixel) within the audio content itself. This step ensures every ad impression is tracked, capturing the initial listener engagement with your campaign. These tags essentially get added into either the RSS feed(one and done process) or the ad server itself (unique to each campaign, ad creative, etc)
By implementing both components, advertisers can harness the full power of Podscribe's pixel-based attribution system, transforming raw data into meaningful insights that drive strategic decisions and campaign optimizations
Using podcast attribution data for your business
Matching and modeling for attribution
As mentioned above, upon obtaining the listeners' IPs, they are matched with the purchasers' IPs. This is collected via another pixel embedded on the online storefront—like classic Facebook or Google ad pixels.
The process of modeling includes dealing with "noisy IPs". These are public or shared IPs such as public wifi, cars, and so on. Pixels aren’t perfect, and 40 to 50% of podcast listener IPs are noisy.
As these IPs are shared among multiple people, attribution providers cannot definitively say that the ad-exposed IP matches the purchaser IP given. This makes it tricky. For instance, if someone is listening to an AG1 ad in a coffee shop and someone sitting next to them happens to purchase AG1, that ad shouldn't get credit. That’s because those are 2 different people who just happened to be in the same place at a coincidental time.
As a result, we model the conversions driven from these noisy IPs to look as if we had perfect tracking ability. We take the conversion rate from the perfectly matched 1:1 IPs (residential) and apply it to the noisy IPs to extrapolate what conversions would look like if we could perfectly track those. The thought behind this is that there's no reason why that random subset of users who happen to be on a noisy IP would engage with the brand any differently.
While we strive for accuracy, these numbers (and models!) are not perfect. This is a limitation given the nature of the data. It’s something we face all across the industry, and is not unique to Podscribe.
One way or another, all podcast attribution companies do modeling. Modeling is a fundamental part of any other channel attribution given that there is always some data that can’t be perfectly matched.
Data collection strategy: A mix of data
While pixel-based attribution has its strengths, it cannot be your sole source of data. In fact, no business should rely solely on one source.
In order to measure and capture the full funnel of your podcast advertising, it is important to implement a mix of pixels and first or zero party-data. This will give your business greater clarity into the full funnel.
First-party data is collected directly from customers’ interactions with your brand. You can capture how they are engaging with your website, social media, or app.
Zero-party data is data collected with the consent and willing participation of your customers. When collecting it, brands typically leverage a value exchange for customers’ information, like gated content, exclusive offers, or advanced personalized experiences.
Using pixel data, you can capture unattributed podcast ad conversions. Thereafter, verify them with your first or zero-party data.
Results and analytics
Outside of just capturing all conversions, aggregating them into one centralized location is critical—allowing you to move beyond the spreadsheet jungle. Podscribe provides a dashboard showing all the key metrics. This includes the likes of IAB verified impressions, reach, frequency, and performance metrics such as visitors, visitor rates, conversions, conversion rates, promo codes, etc.
One of the unique things about Podscribe is that all your attribution data, performance metrics, and reporting is consolidated into one centralized dashboard.
With Podscribe, you are in the driver's seat when it comes to controlling how you look, model, and measure your attribution data. Directly in Podscribe’s dashboard, you have the ability to look at results across different attribution models, set your own conversion window, and exclude existing customers or visitors from what is being attributed to your podcast ads.
The platform allows for varied analyses and ways to look at and export data—whatever it is the user needs. It can remove modeled conversions and view multi-touch results for only static data (perfectly matched IPs). This allows you to see the direct order IDs that Podscribe was able to match with ad exposed impressions.
Podscribe’s focus here is transparency, where no proprietary modeling technique or systems is involved in reporting out the data and each user has complete control and flexibility.
Unique challenges and comparisons
Podcast advertising is similar to other offline channels in terms of attribution challenges. It shares similarities with CTV (Connected TV) and linear TV in its approach to attribution.
Connected TV (CTV) refers to devices that are connected to the Internet and allow viewers to stream videos and music, and browse the web. CTVs enable audiences to use online applications like Prime Video or Netflix. Linear TV, often referred to as traditional broadcast TV, encompasses cable and satellite television.
It’s called “linear” because content follows a predetermined programming schedule, unlike on-demand content which the individual viewer decides to watch based on their own preferences and schedule.
As a result, attribution in these channels have had to develop roundabout methods for attribution–just like we’ve done with podcasts.
Need help with your podcast campaigns and attribution? Let Podscribe help.
