May 2020 Summaries

In the current crisis, businesses are turning to online video streaming as a way to reach their audiences who are now confined at home. Monetization strategies for video content have become crucial in this context. Some common misconceptions around cost include thinking that serving video online is cheap and that free platforms like YouTube can be used instead of owning your video workflow. However, these platforms cut into the profits of original content creators and limit control over product design, quality assurance, and monetization strategy. There are several effective strategies for monetizing video content: Freemium, Base plus Overage, Ad Supported, Revshare, Transactional Video on Demand (TVOD), and Cost Plus. Each of these methods has its own pros and cons, and businesses should choose the one that best suits their needs and audience preferences. A well-executed monetization strategy can lead to a profitable and sustainable business in the world of online video streaming.

The Mux Data platform is a vital tool used by broadcasters to monitor video streaming experiences of their end users. It processes billions of video views every month and allows real-time querying of views as data is being collected. Previously, the storage of video views was done using sharded Postgres databases, which necessitated aggregation for acceptable performance on queries spanning more than a few hours. However, this came with limitations such as limited filtering, operational fragility and difficulty in making changes. Mux Data transitioned to ClickHouse, an open-source analytics database designed at Yandex, known for its speed. It uses columnar compression and sparse indices to achieve efficiency. The new architecture involves a single table called views with no aggregation or stored procedures. This resulted in significant performance improvements, including faster query times and the ability to handle complex queries that would often timeout client-side when sent to Postgres. Despite being column-oriented and using sparse indices, ClickHouse is still faster than sharded Postgres at retrieving a single row of data. The use of materialized indices was instrumental in reducing average single-row query latency from a few seconds to just 75ms. Additionally, the CollapsingMergeTree table engine allowed for "mutating" data without SQL UPDATE statements, which is not supported by ClickHouse. The transition to ClickHouse has been beneficial and Mux Data continues to explore ways to take advantage of its features, such as using Materialized Views, replacing the rollup process with AggregatingMergeTrees, and architecting an efficient real-time query layer.

The Mux Data Real-Time Dashboard is a valuable tool for monitoring the performance and quality of video experiences, offering real-time metrics such as concurrent viewers, rebuffering percentage, failures, video start times, and exits before video start. It updates every few seconds to quickly identify issues and minimize disruptions. The new Real-Time Dashboard API allows users to access all data displayed in the dashboard with real-time latency and five-second aggregation. This API is available to all customers on a Media plan, and can be used to integrate real-time video reporting into existing network operations interfaces or central points of video reporting.

Mux Data has improved its reporting flexibility to better match the complexity of video experience issues faced by users. The filtering capabilities have been enhanced with higher limits, OR queries, and exclusion options. Additionally, a persistent filter bar has been implemented for easier navigation between metrics and data types. These changes aim to help customers be more efficient in monitoring and solving complex streaming problems.

Due to the cancellation of NAB, Mux plans to share new features and improvements through a series of blog posts focused on enhancing video stream visibility. These updates aim at improving developer experience, increasing operational efficiency, and boosting transparency in observability. Observability is crucial for understanding technical workflows via detailed logging, tracing, and monitoring across services. Video observability encompasses various aspects such as end-user quality of experience measurements, CDN logs, packager metrics, origin and storage performance, encoder quality monitoring, CDN switching, engagement metrics, and churn prediction. Monitoring tracks the performance of a resource over time while observability provides detailed information about user experiences and system functioning for specific viewers. Mux Data allows developers to monitor stream performance in real-time and observe individual viewer sessions, helping diagnose issues and improve the video experience.

Mux has been recognized by the San Francisco Business Times and Silicon Valley Business Journal as one of the Best Places to Work in 2020. The company's six core values - Care Obsessively, Be Human, Turn Customers into Fans, Be Clear and Direct, Bias for Action, and Do a Lot with a Little - have played a significant role in maintaining team morale during the pandemic. These values are reflected in Mux's vision to power online video, its commitment to listening and adapting, and fostering rhythms that encourage both work-life balance and team cohesion. The company continues to prioritize learning and growth as it expands its operations.

The author reflects on their experience redesigning a corporate blog, emphasizing the importance of enhancing the reading experience and applying a mobile-first approach. In their redesign, they prioritized content over distracting elements like CTAs and social sharing icons, focused on typography instead of images, and ensured the blog design was guided by editorial aesthetics rather than corporate standards. The team aimed to balance business goals with an improved reading journey by strategically placing actions like visiting the main site or subscribing to a newsletter throughout the reader's experience. They also streamlined the writing process for contributors by simplifying tasks such as adding cover images, previewing posts, and integrating coding features. The collaborative prototyping process, largely conducted on the front end, allowed for real-time adjustments and improvements to the blog's design, ultimately leading to a more accessible and engaging platform for both readers and writers.