Tell me if you've heard this one before... "your data is available as a CSV export." Pretty common, right? CSV exports are basically required for any SaaS product: they’re straightforward to use, they’re supported as an import format in a ton of tools, and you don’t need a lot of experience to get started. All good things.
But technology moves forward, and eventually better things come along. This is one of those times. Let's say hello to something better: Streaming Exports.
Streaming Exports use event streaming services, such as Amazon Kinesis or Google Pub/Sub, to send data to you as it happens, rather than once per day in a batch download. Today we’re announcing general availability of Streaming Exports. We’re starting with Mux Data video views, to stream data of individual views as they complete.
This more granular method of data export provides a lot of advantages that translate to greater business value.
We are using the protobuf format, which allows for a richer data schema definition than CSV. With this richer object representation, we can now provide the events collection in the view export, which is available in the Video View API responses but not CSV exports.
The View events collection contains all of the major events tracked by Mux Data for the view: play, rebuffer start/end, seeking, rendition change, and others.
If you want to get even more sophisticated, you can use the events in the collection to gain additional insights into viewer behavior and use them to inform your business. For example, you can analyze view data to build a heat map of the most watched portions of your videos, which you can use to inform content creation and to better showcase the content that is performing well in your app.
Individual Video Views are exported as each view completes rather than batched once per day. This more granular export provides a lot of flexibility in how you can process and use your data.
- It allows you to create daily or hourly reports more quickly — no more waiting for the daily export batch job to finish! Your reports can have up-to-the-minute freshness, if you need that level of detail.
- You can aggregate data on a local time zone in a timely manner rather than waiting for a scheduled job at a standard time. Your time zone may not align well with once-per-day exports based on a UTC schedule. Daily schedules are no longer a limitation for you and your stakeholders.
Video data can be used not only to inform your business, but also to build engaging customer-facing features. We’re seeing developers build cool applications on top of the data they get with streaming exports.
- Users are building dashboards that features consistently updated data, such as video views. These show valuable information for both internal audiences and end users of a service.
- Flike, a recommendation engine powered by machine learning, built video recommendations by ingesting data through Google Pub/Sub. Their engine is trained on data about a user’s viewing activity throughout a watch session, and serves up video based on that behavior.
Streaming data services are the method of communication for cloud-based architectures. We wanted to make sure Mux Data fits naturally in with the way are building, in order for the data to be as easily accessible as possible.
You can take the streamed data and send it to almost every data and processing service. Store data in Amazon Redshift, Google Big Query, Snowflake, or Splunk; monitor your video experience with New Relic or Datadog; process data with AWS Lambda or Google Cloud Functions. It's kind of hard to overstate how many services can take data directly from Kinesis and Pub/Sub.
Currently, this feature is available to customers on a Media plan or to those who subscribe to Streaming Exports. If you’re using Mux Data and don’t see this feature available, please contact us if you’d like access.
Streaming Exports configuration can be accessed from the Streaming Exports tab on the Settings page in the Mux Dashboard. This page allows you to add, view, and administer streaming exports.
To add a new streaming export, click the New streaming export button. Choose the service you want to export the views to (Kinesis or Pub/Sub) and follow the instructions.
Once the configuration is completed on the AWS or Google side, complete the configuration in the Mux Dashboard. Almost immediately, the service will start exporting your video views as they complete.
We're so excited about this feature that we’re also using it ourselves. We use Mux Data to get deep operational insight into the quality of experience for Mux Video viewers. To extract Mux Video views from Mux Data into a warehouse for secondary analysis and storage, we use this same streaming data functionality that we are making available to customers — so you can be sure that the functionality supports high volume and has the detailed view information necessary to be useful.
We love to talk about video and data, so if you have any questions or creative ideas on how to use all this flexibility, please feel free to drop us a line.