Dynamic Ad Insertion Isn’t Kryptonite; It’s Rocket Fuel.

Dynamic insertion has caught the eyes of investors, entrepreneurs, and advertisers for its potential to change everything about podcast advertising. If the system can become sustainable, revenues, content, and audiences will certainly follow; changing podcast advertising from a niche media investment vehicle into a lucrative, mainstream advertising channel that could propel podcast advertising into a multi-billion dollar industry.

There have been countless articles claiming that Dynamic Ad Insertion, or DAI, is the omnipotent ad-serving mechanism of podcast advertising’s future (mostly written by DAI platforms or platform partners); and almost as many have been written about how terrible DAI is for marketers in the podcast space. This won’t be either of those. Instead, I aim to show where DAI and its associated technologies, publishers, agencies & advertisers, customers, and investors all intersect, and ultimately, who benefits when.

First, a quick overview of the technology itself. DAI describes the injection of an advertisement, usually a 15, 30, or 60-second audio spot, into a podcast upon being accessed. Whether a podcast is streamed (e.g., via Stitcher) or downloaded (e.g., via iTunes), the host file is accessed, and an ad is placed based on targeting parameters. This method of delivery stands in contrast to traditional “live,” host read endorsements that are indistinguishable (and more importantly, unremoved) from the content of the show itself. We call these “embedded” advertisements.

So why DAI? Put simply; it’s a function of scale, control, and ultimately revenue. With over 400,000 podcasts on iTunes, it would be impossible to manage podcast advertising in the same way as Joe Rogan, Pod Save America, or Ben Shapiro. Since these shows have significant (relative) scale, buying them is often done on a one-off basis, requiring phone calls, network relationships, negotiations, host onboarding, buy placement, manual verification, and buy maintenance. They are prone to multiple errors that birth inefficiencies, like hosts misreading the CTA, lack of timely proof that the spot airs (in the form of receiving an aircheck), or a scheduled drop simply not happening and having to be rescheduled or canceled.

So, while it can make sense for an agency, advertiser, and publisher to engage in these procesesses when the shows are at scale, it is operationally infeasible for the same parties to execute in this way for 10,000 placements in shows that would cost, by themselves, $2 each. There has to be a better way if the industry is to scale in both ways: the big shows will continue to get bigger (corresponding to mainstream audience growth) and be purchased manually, while the hundreds of thousands of smaller shows will depend on bourgeoning technologies like DAI to monetize. 

Because DAI will support smaller, niche shows in their monetization efforts, it will play a crucial role in expanding the diversity and volume of all types of content and genres. This will, in turn, bring in new niche audiences to the space, which is a major draw for advertisers: using podcasts as a vehicle for exposing audiences who would otherwise remain unexposed using traditional media vehicles like digital, TV, radio. Therefore, DAI will support the entire ecosystem: content producers and publishers by injecting funding for diverse content development; audiences by giving them more tailored, niche content types to seek out in podcasts (or a reason to seek out podcasts in general); and advertisers and agencies by creating an efficient system of buying placements that can scale.

Certain advertisers will benefit more than others. As it stands, DAI functions very well for brand advertisers. It allows for content control by using wholly pre-recorded and potentially reviewable spots, whereas creative control is often a liability when it comes to embedded, native reads. Secondly, it allows for more precise campaign parameters, like DMA/Zip Code/Country fencing, which is important if an advertiser needs to run region-specific marketing campaigns. Additionally, certain advanced forms of DAI serving, like Panoply’s platform Megaphone, utilize third-party data providers (in Panoply’s case, Nielsen), which allow for the data-driven behavioral targeting that feels familiar to social and digital marketers.

The future is also bright for brand advertisers as the technologies that underpin these platforms advance. With new ways of identifying listeners, a reach number that is actually verifiable isn’t too far off. Correspondingly, the frequency of users across an industry standard, unique, TBD identifier may be possible if the different platforms can agree to share this data through some sort of independent third-party. This would complete the brand advertising puzzle: creative control, targeting parameters, reach and frequency (as a natural byproduct of verified impression delivery).

Sadly, DAI is not currently a successful ad delivery type for performance advertisers. Reasons why are only theories, but there are a few factors that could be inhibiting performance.

First, embedded ads are priced on a CPM basis, which is based on estimated downloads (impressions) a given show will deliver, rooted in a 30-day per episode average. Impression delivery is therefore uncapped, and over-delivery is a mainstay of successful podcast campaigns. Because DAI is strictly impression-based serving and is also based on a CPM, your $20 CPM will remain a $20 CPM, whereas a $20 CPM on an embedded read could easily become $10-$15 due to over-delivery. Therefore, you can gain 50% more efficiency, simply by buying embedded reads.

There is an easy fix to this, which creates a pricing conundrum for publishers and platforms. In order to participate in DAI, publishers often have hefty bandwidth fees that are due to the DAI platform. This creates upward pricing pressure and eliminates the possibility of lowering CPMs to a level where they would perform. So, until bandwidth becomes cheaper (either through technology or through platform competition), the pricing question will remain open.

Another possibility for the relatively low DR response rates from DAI delivery could be the likelihood that if you’re placing advertisements via DAI, an advertiser may not be accessing a host-read endorsement (as you may be running across multiple shows with the same spot, thereby prohibiting a unified ‘host’ as a voice). Moreover, DAI advertisements may fit neatly inside beginning and end ad markers, whereas embedded ads often go over the purchased time (a :60 turning into a 1:35, for example), which could significantly impact performance. One of the major forces at work in podcast advertising is a personalized endorsement, so by not utilizing the connection that the host has to his or her audience, an advertiser may be losing out on the key value of the medium. In limited testing at Oxford Road, we haven’t seen this play a decisive role when using host-read vs. non-voiced spots as a variable.

Because of its importance to the podcast ecosystem, there has to be hope for successful DR response via DAI, and it’s only a matter of time before someone cracks the code. A few things might happen that could catalyze the marketplace for performance advertisers. For example, if brand advertisers find that DAI is a safe alternative to embedded reads, more dollars could be invested in DAI placements, which could reset what looks like, at times, a sellers’ market. This could push prices on both DAI and embedded placements down for DR advertisers, as brands will be able to afford higher prices on the more efficient (from a publisher-profit perspective) placements (DAI). Meanwhile, DR advertisers will be able to continue their stronghold in the embedded space while achieving efficient pricing on DAI placements that won’t drive publishers into a loss.

Additionally, in thinking about how DAI will indirectly provide more fiscal flexibility to afford a direct response DAI marketplace, it is possible that organically growing audiences coupled with increased revenue from brand advertisers utilizing DAI could further stimulate an already excited investor market. This could lead to even more venture capital flowing into the marketplace. The cascading benefits would run full circle as new capital means more content, more new audiences, and better technology, which would lead to more capital ad infinitum.

In sum, DAI is a good thing. It’s a necessary tool for monetizing a growing space that is as diverse as the internet since literally anyone with a smartphone can record and upload a podcast. It’s not perfect yet, but it’s already proving to be a pivotal element to the ecosystem. As performance advertisers, we may have to wait until pricing is more efficient and the tech more sophisticated, but we’re not far away. In the meantime, as it pertains to brand advertisers, publishers, content producers, and audiences, DAI is the rocket fuel that will take podcast advertising even further off the ground.

Leave a Reply