Audience Segments
Segments create collections of viewers based on viewing behaviors to help you analyze viewership based on custom viewer groups. Options for behavioral viewer segmentation include consumed content, consumption attributes, and specific behaviors.
This topic contains conceptual information about segments in Conviva Audience.
Behavior Segments
Based on data science analysis techniques, including clustering, and other methods of distribution analysis, Audience provides a set of advanced behavioral segments to drive a better understanding of your audience behavior. Segment data is computed daily, based on a 90-day trailing window, and sets thresholds based on offline machine learning, clustering analysis, and usage patterns.
You can also use the segmentation options to define custom segments based on filter settings, logical AND/OR operations, and customized criteria. Segments are viewer-based and identify a set of viewers based on a specific set of attributes of the viewers, for example people who watched a given episode, in a particular DMA, on a particular device. Because the segments are viewer-based, the data set is inclusive of the viewer consumption. As a result, the viewer data also includes data for consumption outside of the prescribed set attributes (episode, DMA, and device in this example).
Note: For enhanced data sets, data can be joined with external data to create advanced data collections. For more details, see Complex Segments and Filters and Segments.
| Behavioral Segment | Package | Segment Components | Statistical Usage | Contains users that |
|---|---|---|---|---|
| Series Consumption | Advanced | Binger | Top 20% of playing time over a period, typically seven days, watched over a single series with less than 10 minutes of interruption. | Consumed at least 80% of the overall playing time watching a single series during a single sitting. |
| Super Binger | Top 50% of playing time over a period, typically seven days, watched over multiple series with less than 10 minutes of interruption. | Consumed at least 50% of the overall playing time watching multiple series during a single sitting. | ||
| Streaming Consumption | Standard | Heavy Streamers | Top 5th percentile of users based on playing time per day | Consumed the most content per day over the last time period. |
| Medium Streamers | Between 50th percentile and 95th percentile based on playing time per day | Consumed a typical / median amount of content per day. | ||
| Low Streamers | Less than 50th percentile based on playing time per day | Consumed the least amount of content per day. | ||
| Loyalty Segments (Recency, Frequency, Engagement) | Advanced | Highly Engaged Viewers |
Recency: Recency less than 80% quantile Frequency: More than 60% percentile of play hours Engagement: More than 50th percentile of play hours |
Watched more hours of content than the typical (median) users
Watched content very frequently Watched content in the recent past. |
| Occasional Viewers |
Recency: Recency less than 80% quantile
Frequency: Between 30% to 60% percentile of play hours Engagement: More than 50th percentile of play hours |
Watched more hours of content than typical / median users
Watched content less frequently Watched content in the recent past. |
||
| Infrequent Viewers | Recency: Recency less than 80% quantile
Frequency: Less than or equal to 30% percentile of play hours Engagement: More than 50th percentile of play hours |
Watched more hours of content than the typical / median users
Did not watch content frequently at all Watched content in the recent past. |
||
| Lapsed Viewers | Recency: More than or equal to 80% quantile
Frequency: --- Engagement: Less than 50th percentile of play hours |
Watched less hours of content than the typical / median users
Have not watched content in the recent past |
||
| In-active Viewers | Recency: More than or equal to 80% quantile
Frequency: --- Engagement: Less than 10th percentile of play hours Watched very few hours of content Have not watched content in the recent past |
--- | ||
| High Life Time Value | Advanced | HLTV-6 | Time frame A are viewers who were watching 6 to 8 months ago. So in a 60 day time range they have watched at least one session. Time frame B are viewers who have watched at least 1 session in the last 60 days. The intersection of time frames A and B is High LTV (6 months). |
Watched at least one session in 60 days prior to the last 6 months and at least one session again in past 60 days. At least 8 months of data is required for data calculation. Insufficient status appears if less than 8 months of data is available. |
| HLTV-3 | Time frame A are viewers who were watching 3 to 5 months ago. So in a 60 day time range they have watched at least one session. Time frame B are viewers who have watched at least 1 session in the last 60 days. The intersection of time frames A and B is High LTV (3 months). |
Watched at least one session in 60 days prior to the last 3 months and at least one session again in past 60 days. At least 5 months of data is required for data calculation. Insufficient status appears if less than 5 months of data is available. |
||
| Device Type | Standard | Multiple Devices | At least 10% or more of content plays on multiple devices. | Streamed content on multiple devices. |
| Pure Single Devices | 100% content plays on a single device. | Streamed content on only a single device. | ||
| Day Segments | Advanced | All Days of the Week Watcher | Every day of the week. | Watched every day of the week. |
| Any Day of the Week | At least one weekday and one weekend day, but not every day. | Watched at least one day of the week and one day of the weekend, but not every day. | ||
| Weekend Only Watcher | At least one day of weekend watching. No weekday watching. | Watched at least one day of the weekend without watching on any of the weekdays. | ||
| Weekday Only Watcher | At least one weekday watching. No weekend watching | Watched at least one day of the week without watching on any of the weekend days. | ||
| Day Part Segments (A viewer only belongs to the most major segment) | Advanced | Day Time Watcher | Weekdays between 9:00 am and 6:00 pm | Primarily watched between 9:00 am and 6:00 pm on any weekday. |
| Early Morning Watcher | Weekdays between 5:00 am and 9:00 am | Primarily watched between 5:00 am and 9:00 am on any weekday. | ||
| News Time Watcher | Weekdays between 6:00 pm and 10:30 pm | Primarily watched between 6:00 pm and 10:30 pm on any weekday. | ||
| Late Night Watcher | Weekdays between 11:30 pm and 5:00 am | Primarily watched between 11:30 pm and 5:00 am on any weekday. | ||
| Saturday Night Watcher | After 7:00 pm on Saturday | Primarily watched after 7:00 pm on Saturday. | ||
| Sunday Night Watcher | After 7:00 pm on Sunday | Primarily watched after 7:00 pm on Sunday. | ||
| Weekend Day Time Watcher | Weekends between 12:00 pm and 6:00 pm | Primarily watched between 12:00 pm and 6:00 pm on any weekend day. | ||
| Weekend Morning Watcher | Weekends between 7:00 am and 12:00 pm | Primarily watched between 7:00 am and 12:00 pm on any weekend day. | ||
| Password Sharing | Advanced | Non-Primary Shared Subscription HHs | For a subscription that is shared with multiple households, the primary household is considered the one with the highest total play time and therefore excluded from this segment as the ‘legitimate paying household’. | Households that are potentially leveraging a shared subscription and therefore can be converted to a paying subscription. |
| All Shared Subscription HHs | All households associated with a subscription that is used in multiple households. | Are associated with a subscription that is being used across multiple households. |
Complex Segments
Audience supports advanced data segmentation and joining of behavior segments to help you query complex viewership behaviors. The combination of segments can be applied to refine the viewership focus and can also be exported out of Audience for offline use with other segments or Data Management Platforms.
Use-Case
Let's assume, Episode 1 of the series, The Big Bang Theory, has the most viewership and for episode 2, the viewership has dropped. The goal is to identify the number of viewers that were engaged for episode 1 but did not engage for episode 2. We can figure this out by analyzing the viewer engagement across episodes.
For detailed steps on how to analyze this use-case, refer to Complex Segments topic.
Segment Overview Dashboard
The Segment Overview dashboard helps you understand your audience based on consumption or behavioral patterns that identify which segments have the highest reach (size of audience) as well as emerging segment trends.
Segment Management Dashboard
The Segment Management dashboard displays the list of all the segments created for your C3 account. You can also create and edit segments on this page.