Triton Digital performs census-based, digital, on-demand and podcast audio audience measurement. Apart from the Podcast Metrics Demos+ service, no samples, surveys, or panels are used in the collection, transformation, or display processes and procedures described herein. In each case, Triton Digital obtains data on each file request, including the content, source, timing, and bytes requested for each download.
Triton Digital uses a single method for collecting data for Podcast Metrics services. Through raw log files or events of on-demand audio activity collected daily from either the Content Delivery Network (CDN) or the Content Management System (CMS) hosting the network’s podcast and on-demand audio content.
Data Collection
Triton Digital uses multiple methods for collecting log level data for Podcast Metrics services. We accept raw files coming from either the CDN access logs, or CMS/CDN session logs. If other options are not available, we may accept access logs from redirect services. We then apply appropriate normalization and enrichment in order to provide equivalent/comparable measurement.
The audio activity is collected daily from one of the sources mentioned above.
When a publisher or network delivers on-demand audio, it is delivered via a content delivery network or a podcast hosting platform, which then distributes the files via a network of geographically disbursed delivery points (servers), and then logs the transactional activity. These can be recorded in either log files, databases, or data warehouse systems. Individual log events are made available by the CDN or CMS, or redirect service, to Triton Digital. Logs contain the details of all listener file access for a given day. With respect to the scope of this document, the key data points within the log files are IP address (or hashed version), user agent, date/time, transferred bytes, request method, request bytes, and the URL or ID of the file.
Networks approve and arrange for the release of streaming log files directly and automatically from their CDN or CMS of choice to Triton Digital by SFTP, or by S3 bucket in a predetermined format. In all cases, log files are generated and controlled by the CDN or CMS, and not the networks themselves. Triton Digital’s quality control procedures apply to every client, regardless of how the log files are delivered.
Covered Devices and Platforms
Our Podcast Metrics census-based log file methodology supports measurement across all devices and distribution platforms, so long as Triton receives valid server or client access or session logs. This includes, but is not limited to, web browsers and apps, smart phones and tablets, smart speakers, and other devices across a wide variety of operating systems and brands capable of downloading on-demand audio files.
Association of Log Data to Podcasts
Podcast Metrics associates each file download to a specific episode title within a podcast. Triton uses either episodes ID or URL to match log data with episode information within a set RSS feed or data (csv) file provided by the network.
Measurement Limitations
HLS and MPEG DASH
Triton Digital’s Podcast Metrics doesn’t support raw access log information from HLS or MPEG DASH streaming of podcast / on-demand content. To be accepted, HLS or MPEG DASH streams shall report session logs instead of file access logs inclusive of duration and episode information.
Caching from Different Platforms
Some podcast listening platforms have the ability to cache podcast episode files in their own delivery network and as a result, may not provide server logs. A lack of server logs can negatively impact the accuracy of measurement as well as the listening experience as it relates to dynamically inserted advertisements. In many cases, networks can contact distribution platforms and ask to avoid caching.