Webcast Metrics reports on “census based” Internet audio streaming traffic and audience measurement. It does not rely on samples, surveys, or panels in the collection, transformation, or processing of data. Instead, it uses client-side measurement to collect the data directly from the web-based player or mobile device, which we refer to as the "LT" or "listener tracking" method.
Triton Digital obtains data on each stream including the station, individual stream start time, individual stream duration, and listener identifier. This methodology is based on all recorded session activity, subject to filtration as described in Filtration for Accuracy.
The items below describe how Listener Tracking works and provide a brief outline of the Webcast Metrics and Streaming Metrics reporting methodology. For a fuller description, see the Description of Methodology (WCM + WCML).
How Listener Tracking Works
The listener tracking functions monitor and report the initiation, continuation, pausing and resuming, and termination of the stream. Triton Digital uses a combination of third party cookies and publisher supplied unique user ID to report a CUME figure based on unique cookies or VID.
Upon initiation of the stream, an initial start event is sent to the measurement servers identifying the stream (e.g., Station ID) and any additional optional information the station includes in the event string. The tracking system then continues to send "ping" events every sixty seconds as long as the stream is active. If the user pauses the stream, the ongoing event is stopped until the listener resumes the stream. In cases where the pause event was less than three minutes, the ongoing event is resumed and counted, otherwise, a new session is started. Finally, if the user ends the stream, the ongoing event is stopped, and if the user subsequently restarts the stream, a new session is started regardless of the time between the end event and the new start event.
If you need more detailed description of the LT methodology, contact your Triton Digital Account Manager.
Filtration for Accuracy
Triton Digital employs several techniques to identify and filter (exclude) invalid activity, including but not limited to known and suspected non-human activity and suspected invalid human activity. Because user identification and intent cannot always be known or discerned by the publisher, advertiser, or their respective agents, it is unlikely that all invalid activity can be identified and excluded from report results. However, repeated testing has confirmed that the filtration protocols described below are effective at removing most invalid activity without significantly affecting the measurement of valid activity.
One Minute Rule
All streaming sessions with a duration of less than one minute are considered invalid and are removed from the collected measurement data. This rule reduces most of the noise from extremely short sessions, robotic activities, and initial connectivity issues, without cutting into real, high-value audience measurement. Therefore, the session is not considered “active” until the first ping event occurs 60 seconds after the “start” event.
Note that Triton Digital has not detected any robotic/spider activity that can consume an audio stream for longer than one minute.
Spiders and Bots Filtering
Listener data is filtered according to the IAB/ABC International Spiders & Bots List to exclude site traffic associated with known robotic activity. The IAB/ABC list is updated every two weeks.
Internal Traffic IP Filtering
All stream session data generated from within Triton Digital is removed from collected measurement data. This includes our VPN traffic and any internal traffic generated by participating stations/publishers.
Activity Filtration
Triton Digital employs activity filters to identify anomalies. All suspicious streaming activity is documented in a monthly report and the publisher is immediately notified. Triton Digital works with publishers to investigate and resolve the issue.
Collection and Reporting Notes
The following notes should help with your understanding and interpretation of the reports in Webcast Metrics and Streaming Metrics.
- Gross vs Net metrics: metrics marked "Gross" represent all sessions, including invalid ones that are normally filtered as per Triton’s Description of Methodology; metrics marked "Net" represent counts that exclude invalid sessions – such as from "bots" and spiders – and is more representative of actual human listenership.
- Total Listening Hours (TLH) is the only metric that contains a decimal, and the decimal only goes one place. (E.g., "2.5" but not "2.5.5" or "2.55.")
- Average Time Spent Listening (ATSL) is displayed in HH:MM format, with rounding applied to the sub-minute fraction. Values of ".5" are rounded down.
- All metrics are displayed as integers (whole numbers), with the exceptions of ATSL and TLH. All metrics are rounded, with ".5" rounded down.
- Raw log files are maintained for a rolling thirty-two days.
- Data aggregate records go back as far as April 2015, or as long as you have been a user of Webcast Metrics/Streaming Metrics (whichever limit is reached first).
- Data is available online for all reporting periods (including the current month). Current month data is available 15 days after the close of the period. The near-line archive database maintains data starting in 2008, and is updated nightly. Data prior to 2008 is stored as offline backups.
- Data confidentiality: Triton Digital will not share data across ownership groups without prior written permission. Proof of written permission shall be maintained at the Triton Digital corporate office.
- Data error disclosure: Triton Digital will reissue data whenever an error or omission is found that affects any reported metric, for any station, by more than 2%, or would lead to a change in the ranking of affected station(s). In the event of such an error, Triton Digital will communicate the above error or omission to you via email.