As part of the VAMP project we explore key areas of the digital audio and especially podcast ecosystem. The first milestone of the project is a comprehensive analysis that looks at the interests and perspectives of publishers as well as relevant service providers in the audio and podcast space.
The following report is based on several months of research conducted through desk research, surveys with service providers as well as interviews and workshops with national and international media companies. It gives an overview of the existing software solutions, compares features and technical standards and reveals in which areas we could only find inadequate technical solutions for publishers. Based on this information publishers can select suitable service providers for their needs.
Our analysis revealed that (1) podcast hosting solutions, (2) dynamic ad insertion and (3) analytical standards are most important for publishers. Thus, we will focus on these areas.
Part 1: Podcast hosting solutions
According to our research, one of the most pressing issues for publishers is to have a stable and flexible audio web player with comprehensive hosting and distribution functions. To date, many publishers have integrated either relatively expensive or inadequate services into their websites and only few executives are satisfied with current solutions.
The market offers a wide range of hosting solutions for podcasts. The spectrum ranges from small, low-cost services for hobby and semi-professional podcasters to all-in-one solutions that primarily have radio stations or podcast networks as customers. We at SPIEGEL, like many other publishers, have a clear focus on text content and a manageable podcast offering compared to public broadcasters, for example. That’s why we see a broader need in our industry between the two extremes mentioned above. It is therefore essential to accurately capture the range of offerings and differences between podcast services.
In our comparison, we surveyed nearly two dozen podcast hosting companies; ten providers responded in detail. Focus of our survey was on a) web players, b) costs, c) monetization and d) analysis.
Links in tables only work in the PDF version: Download complete Podcast-Hosting PDF here
a) Web player
For a website like spiegel.de an embeddable audio player, its functionalities and adaptability to our requirements is central when choosing a potential partner. Depending on podcast series and episode, between 25 and 50 percent of podcast downloads are generated via our own website. We therefore see it as a core task to make our audio playback areas more attractive and practical. Features such as the support of transcripts or chapter marks may increase discoverability in search results and are potentially relevant selection criteria. Therefore, we have extensively queried the feature range of the web players of podcast hosting solutions:
Customizability of the player is important for media brands to ensure a holistic user experience. First, the corporate design should be implementable in the player layout; second, one needs to be able to (de-)activate features in line with user and content requirements.
Our study shows that almost all podcast hosting companies offer an embeddable audio player, but the degree of customizability differs between providers. Some offer two or more player versions, such as a minimal player with only main controls or a full feature player that supports playlists and transcripts.
The pricing models of the smaller providers are very similar and often independent of the required bandwidths — at least if a certain range is not exceeded. With monthly costs between $5 and $20, many providers host a podcast professionally. Higher prices are usually due when the full range of functions, e.g. complete analytics, is booked.
Large providers, which are mainly used by radio stations and podcast networks, usually do not work with fixed price tiers; prices are almost always dependent on the required bandwidth and must be negotiated individually.
Podcast hosting companies differ greatly in the infrastructure they offer for monetizing podcasts. From technical solutions for dynamic insertion of ads to support for paid podcasts and partnerships with podcast marketers.
Getting quality data on the actual use of podcasts and on how listeners deal with audio content is a big challenge. Reasons are discussed in more detail below. We asked what figures the podcast companies can provide and if they support the IAB Podcast Measurement Standard when counting podcast downloads.
Part 2: Dynamic Ad Insertion
How will the advertising market for podcasts and audio on demand develop in the coming years? Will standards, I.e. podcast host read ads, exclusive sponsorship and native ads remain prominent? Or will dynamically inserted, targeted advertising become important in the future, similar to video and display advertising, i.e. will Programmatic Advertising also take over audio on demand?
We discussed this topic extensively with different European publishers at dedicated workshops: Skepticism towards programmatic advertising in podcasts prevails while programmatic is seen as the future in other forms of digital audio content:
- Most publishers are convinced that audio advertising formats produced for radio are generally not suitable for podcasts. Radio commercials spoile the podcast experience because of their brute nature
- Programmatic audio ads especially made and optimized for podcasts are not yet in sight, at least for the German-speaking countries
- CPMs for individual advertisement in podcasts are — for good reasons — very high. Programmatic ads would be lower-priced and might decrease the prize-level for podcasts overall
- In addition, many publishers still lack data and technical solutions to integrate programmatic ads into podcasts
Despite the current arguments against programmatic advertising in podcasts, most workshop participants agree that programmatic ads will play a bigger role in the future, also for audio content. Thus, publishers need to explore the technological opportunities for dynamic ad insertion in audio files.
No matter how the market evolves, Native ads that are permanently integrated into the audio file limit the possibilities of podcast revenue models. A sponsoring of a podcast usually expires, host-read ads only have a certain runtime and become obsolete. Podcasts are often long-term content and could be sold again after an original booking has expired. The subsequent manual insertion of new ads is simply too time-consuming.
We surveyed providers of Dynamic Ad Insertion (DAI) solutions for audio files. The existing offerings differ greatly. While some DAI services act as technical service providers only, others function as selles of advertising space at the same time. Five providers — Voxnest, AdsWizz, Triton Digital, ART19 and ACAST — responded in detail to our market study.
Links in tables only work in the PDF version: Download all tables as PDF here
a) General infomations
b) Costs and share
c) Analytics and more
Part 3: Podcast analytics
Data about podcast usage is often inconsistent and not always comparable. First, the landscape of podcast player solutions is highly fragmented. In addition to podcast platforms and players of large companies such as Apple Podcast, Google Podcast and Spotify, there are many independent podcatchers and player apps. Second, a podcast is an on-demand media format that listeners either listen to directly online or download to listen to later. Both factors complicate the goal of capturing podcast usage data in a meaningful and unified way.
How podcast hosting companies measure downloads was not transparent for a long time. If the audios were delivered via different hosting platforms it was especially hard to compare numbers and success with competitors. For some time now, however, IAB has succeeded in establishing a Podcast Measurement Standard that is generally recognized and implemented by more and more podcast platforms. IAB now also offers certification, first companies are certified (see table above), and many are currently undergoing the certification process.
Some essential counting requirements in the IAB metrics are:
- Pre-load requests from players have to be excluded.
- Each individual downloader may only count once
- Bots and Bogus requests have to be removed from the calculation
- No count if less than 60 seconds of a podcast episode have been loaded
Parallel to the IAB standard, Swedish podcasters, radio stations and publishers have developed another measurement standard, the Poddindex standard. The Swedish Podcast Charts are based on this standard. Internationally, the Poddindex standard may not have the same prominence as the IAB standard, but it is largely comparable.
Downloads vs. Listens
Standards for download counts do not solve the challenge to determine how often or how long an episode has been listened to. Many podcast apps offer their users the option of automatically downloading new episodes of their subscribed podcast as soon as they get published. Publishers usually don’t know whether these episodes, which are counted as downloads, will be listened to later. Even less do podcasters and publishers know how well individual episodes perform:
- What percentage of an episode is listened to on average?
- Are there drop-off points at which a particularly large number of listeners drop out?
- How do listeners behave during advertising breaks (skip rate)?
Publishers need such data in order to identify weaknesses in their content or to systematically eliminate misleading audio elements in podcasts. In addition, the performance of advertising could be better analyzed: Which ads are accepted, which lead to high abandonment rates? Such data would have to be provided by the audio player, because the IAB standard is purely server-based.
Spotify and Apple Podcast already provide performance curves and data for each episode. A comparison of the curves for the same episode of a SPIEGEL podcast at Apple and at Spotify shows that the results are quite similar and probably transferable to user behavior in other app-based players:
However, one can assume that the performance would look worse in a web player. We assume that web-based listeners would exit an episode earlier and expect higher termination rates. In order to measure the performance of a podcast episode uniformly, all relevant player providers and podcast hosting companies in the market would have to agree on a common standard and implement it.
An ambitious concept is currently being developed under the leadership of the US radio network NPR. The Remote Audio Data Method (RAD) is intended to enable analysis “across a wide range of participating providers and platforms” by aggregating data in an analysis center at the publisher’s site. NPR points out that RAD is not intended to replace download statistics, but to help consolidate and extend data standards.
However, the crucial factor to success will be how the measurement method can be reconciled with the data protection guidelines of the respective markets. For publishers, it remains to be seen whether and how quickly RAD or an alternative approach will establish itself in the market.