How Ooyala IQ Works

During playback, video players send event-based metrics that are recorded along various dimensions, summarized, and aggregated for reporting and analysis. After presenting some key concepts, this overview describes the major components of Ooyala IQ.

Key Concepts

Three ideas are essential to understanding Ooyala IQ: 1) facts, 2) events, and 3) dimensions. These points are described below.

  1. All information measured by Ooyala IQ is tied to an asset (usually a video or live stream). In traditional analytics terminology, the video is called a fact. Every video has a unique identifying number, called an identifier.
  2. As a video is played, video players record events (such as “display”, “playStarted”, or “complete”) over time in the life of the video. Your videos (facts) and their associated events are the basis for all metrics.
  3. Every event/video combination also has predefined aspects called dimensions (or attributes) that relate to the fact, such as traffic source, player, geography, and device.

Ooyala IQ records data from events over time along the following predefined dimensions during playback.




What kinds of devices the video was played on, such as mobile, tablet, iPhone, iPod, and more.


Where the video was played, such as country, state, and designated marketing area.


Which video players it was played on. You can define these players' names yourself.

Traffic Source

Where the video was served from.


Time is a “hidden” dimension that represents when events occur. All events have an explicit time stamp recorded by the system.

Three Steps of Ooyala IQ

Ooyala IQ has three primary parts. You do parts 1 and 3 and Ooyala IQ does part 2.

  1. Recording Metric Data
  2. Aggregating and Summarizing Data
  3. Reporting, Analyses, and Data Visualization

Each step is discussed in the remainder of this section.

Recording Metric Data

In the first step, to collect metrics, certain events that happen during playback are recorded by the video player as they occur, such as "playStarted", "pause", and others. These events are "data ticks" sent via a Representational State Transfer (REST) application programming interface (API) over the Hypertext Transport Protocol (HTTP). A single event or group of events sent in a single request is sometimes called a ping.

Events are recorded in a database called the event store.

Aggregating and Summarizing Data

In the second step, throughout the day, Ooyala IQ summarizes and aggregates these event-driven metrics into various time “chunks” (or segments), such as daily, monthly, or yearly.

"Aggregation" means that the intersections among the various dimensions are precomputed so reports are faster. For instance, aggregation anticipates that you will want to ask questions like “Which devices were used in which geographies?” or  “Which players were used on which devices?” These “overlaps” are the overlapping parts of the diagram above.

Other data transformations can also occur, such as calculating other metrics based on the raw data.

The end result of aggregation, summarization, and other transformations is datacubes that store these precomputed intersections.

Reporting, Analyses, and Data Visualization

In the third step, after the metrics have been gathered, summarized, and aggregated, and the datacubes created, the data is ready for you to use.

So you can zero in on the particular relationship of the data that you are most interested in, you can access analytics data in several different ways:

  1. The GUI, which includes the following:
    • The Dashboard for near real-time analysis.

    • The Business Intelligence page, where data can be filtered by the various dimensions or manipulated in other ways.

    • The Video Details page, where you can look at detailed metric breakdowns for a specific video.

  2. A REST-over-HTTP API, which can be used for data downloading or for feeding into your own reporting systems.

You can create multidimensional analyses to visualize intersections of the data. Multidimensional means more than one dimension. As mentioned in the previous section, you might want to answer the question “Which devices were used in which geographies?” or  “Which players were used on which devices?” Both questions are answered with multi-dimensional analyses.

Analyses with dimensions rely on the precomputed datacubes described above.

You can also constrain the data by filters (such filtering is sometimes known as slicing and dicing). You can apply filters such as date, geography, and so on, so that the amount of data you view becomes more and more sharply focused. You can thus view only the subsets of the data that interest you.

Data Retention

Ooyala will only retain up to 37 months of your data. This is calculated as the current month along with the past 36 months. Data will be available for dates on or after January 1, 2014. Ooyala will only present and allow queries on the most recent 37 months of valid data for you in UI and API. The rest of your data will be archived. For information on how much historical data will be migrated from v2 Analytics to Ooyala IQ, please see Data Migration (Deprecated).