Methodology for Podtrac's Free Third Party Measurement of Podcasts

Background

This provides a technical summary of the technology, process, and methodologies used by Podtrac for its free third party measurement services available to any podcast. Podtrac has developed its podcast measurement methodologies working with a team of professionals who have developed:

• Internet data collection methods currently used by comScore
• some of the first advanced technologies for measuring Internet advertising effectiveness
• measurement systems for cable and digital television
• improvements to methodologies for meter, diary, and People Meter services for Arbitron

Podcasts Defined

Podtrac defines a podcast as a recorded file or series of recorded files made available to Internet users through a distribution protocol known as Really Simple Syndication, or RSS. To be measured by Podtrac, a podcast must have an associated RSS feed.

Podcasts are sometimes referred to as podcast series. Individually recorded files, editions, or releases from a podcaster within a single podcast series are generally referred to as podcast episodes.

Measurement Technology and Process

Podtrac is a complete measurement and reporting system, providing podcasters and advertisers with information about thousands of podcasts. The system begins with Podtrac providing the podcaster a method for uniquely identifying the podcast so that Podtrac can collect data. The system continues with processing and categorizing collected data, and it finishes with analysis and reporting of the processed data.

Podtrac collects data using technology which encodes the "enclosure" XML of an RSS feed and the direct URL links on podcaster websites. Whenever a listener or viewer downloads a podcast episode via one of these encoding methods, Podtrac collects information about that transaction as an integral part of the fulfillment of the request for the actual episode file resident on the podcaster's server. This operation is instantaneous and transparent to the end user.

Podtrac's measurement of podcast usage is at the lowest (most granular) level, the individual podcast episode. This enables usage to be measured whenever it occurs. Once data is collected, it is then analyzed using Podtrac's proprietary nested podcast analysis technologies. This provides advertisers and podcasters with the most reliable data set for better understanding user activity on a per podcast episode basis.

Data Collected by Podtrac

Podtrac gathers the following information about each Podtrac-encoded podcast episode:

• The date and time of the request to start the download
• The source of the podcast download
• The unique podcast ID, as assigned by Podtrac collection technology
• Whether the request came through a URL or an RSS feed

Podtrac uses these data points, along with proprietary information about podcast episodes, series, and sources to summarize and aggregate podcast usage.

Measurement Analysis Provided by Podtrac

Podtrac's analysis technology produces the following information on both a daily and monthly basis about each encoded podcast:

Downloads

This is a measure of the number of times a specific podcast episode file is requested during a given period of time. Downloads include file requests coming from podcatcher software, podcast directories, and podcaster websites. The Downloads metric recognizes that, unlike some other media types, podcasts can be enjoyed online without the user waiting for the entire file to download. Therefore, if ads or sponsorship messages are placed near the beginning of a podcast, users can be exposed to these messages without downloading or listening to the entire podcast episode.

Unique Downloads

Unique Downloads is an unduplicated count of individually identified machines that began downloading a specific podcast episode during a given analysis period. Podtrac defines Unique Downloads through a combination of cookies and algorithms based on IP address.

Download Source

Download Source is a measure of the client software used to download a podcast being counted, or the web site or server the user last accessed prior to downloading the podcast episode.

Downloads by Country of Origin

Downloads by Country of Origin is a measure of Downloads for a podcast originating from various countries during a given analysis period. Podtrac utilizes best of breed databases to infer country origination from podcast download request data.

Limitations of the Data Collection and Analysis Methodology

Every measurement system has its limits, and Podtrac's is no exception. The Podtrac measurement methodology has the following limitations that should be considered by users of the data:

Downloads

• Downloads are tracked to specific computers through the use of proprietary methodology utilizing IP addresses, cookies, Internet directories, and specialized databases and algorithms. This methodology maintains the privacy and identity of the individual users that are downloading a podcast episode.

• Podtrac measures requests for episodes from podcaster servers. If podcast aggregators bypasses podcaster servers by caching content on their servers, these requests are not counted.

• Since podcasts can be downloaded and stored, some percentage of downloads may not be played. Podtrac does not currently have a way to determine the exact number that fall into this category.

Unique Downloads

• Podtrac uses a proprietary method utilizing IP addresses, cookies, Internet directories, and specialized databases and algorithms to differentiate Unique IP addresses. IP addresses and cookies may not necessarily represent a single user, as more than one user may share a machine, or a user may download the same podcast from multiple machines.

• Since the Podtrac method uses cookies as part of the algorithm to identify unique users, an undetermined number of users may be incorrectly identified as unique because the user has deleted cookies. A March 2005 survey from Jupiter Research found that as many as 39 percent of online users delete cookies from their primary computer each month: 12 percent delete monthly, 17 percent weekly, and 10 percent daily. The extent to which these numbers reflect the habits of podcast listeners and viewers is unknown at this time. Additionally users who do not allow the use of cookies, or whose cookies have been lost because the computer was upgraded or browser-related software was reinstalled may be identified as unique, when they are actually returning downloaders. The extent to which either of these situations occurs cannot be determined by the Podtrac methods.

• IP addresses can change frequently, especially for mobile wireless users. To the extent that the same podcast is downloaded multiple times through different IP addresses by the same user, Unique Downloads will be overstated.

• IP addresses are assigned by Internet Service Providers (ISPs). For large ISPs, IP addresses often bear little relationship to the actual location of the end user. All North American AOL users, for instance, appear to come from AOL headquarters in Virginia.

• Corporations with field-based personnel often provide remote connectivity through in-house proxies, so the IP address belongs to the corporate headquarters while the actual user could live and work somewhere else entirely.

• Some users may simultaneously use multiple modems, each with a different IP address, to accomplish the download. The occurrence of this is low, and expected to go lower as broadband high-speed access grows.

Download Source

Certain podcatcher software packages do not provide a referrer, in which case the client software being used cannot be identified. (We are working with podcatcher software developers on an ongoing basis to minimize this issue.)