Streaming Video Encoding & Packaging Pipeline  ·  View 03 of 22  ·  People and journeys

Actors and Their Core Journeys

Eight actors, their goals in their own words, and which journeys get their own map.

Editable source SVG draw.io All views
People who watch Viewer on a big screen assumed 40 M MAU Goal — I press play on the thing everyone is talking about and it looks like television, immediately, with no spinner and no soft first ten seconds. Core journeys Watch a new release premiere night Resume something old Viewer in the long tail 72% of catalogue hours Goal — I picked a 1974 documentary nobody else is watching tonight, and I would rather not be punished for having unusual taste. Core journeys Play a cold-stored title rendition not resident People who ship the content Content operations assumed 24 staff Goal — I need tomorrow's premiere playable on every device before the embargo lifts, and I need to know hours in advance if it will not be. Core journeys Get a title to air 1,800 versions/day Waive a narrow QC failure Video quality engineer owns the gate Goal — I want the gate to fail the things viewers would notice, and I want to know which encoder build caused a regression without bisecting the catalogue. Core journeys Tune a ladder from telemetry Roll out a new encoder build Media FinOps cost per delivered hour Goal — I want to see what a codec or a rung costs before it is switched on for 400,000 titles, not in next month's invoice. Core journeys Price a codec addition Partners, systems and machines Rights holder assumed 310 suppliers Goal — I hand over one master in the format we agreed and get told within the hour whether you accepted it, not three days later. Core journeys Deliver and redeliver a master DRM licence service 3 systems Goal — Issue a licence for a key the pipeline never stored itself. Core journeys Serve a playback licence Platform SRE on call Goal — When 60% of the fleet is taken away at 03:00 I want the queue to drain anyway and nobody to page me. Core journeys Survive a reclaim storm Actors and Their Core Journeys Goals are in the actor's voice. Journeys with an id get their own map; the rest are named so the set cannot pretend they do not exist. v 1.0 · owner Media Platform Architecture

The two viewers are different people

  • The premiere viewer and the long-tail viewer want the same thing and cost wildly different amounts to serve. Splitting them is what makes demand-driven generation arguable at all.
  • The long-tail viewer is the actor the critical design decision is paid for by — their journey is the one with a cache miss in it.

Who can override the machine

  • Content operations can waive a QC failure; the video quality engineer owns the floor it waives against. Separating them is deliberate: the person under deadline pressure is not the person who sets the bar.
  • Media FinOps is an actor rather than a report because a codec addition is a four-hundred-thousand-title decision and needs a price before it is switched on.

What is not here

  • No subtitle author, no metadata editor, no recommendations team — all real, all upstream or downstream of this boundary.
  • Platform SRE appears because the fleet's reclaim behaviour is an architectural property, not an operational inconvenience.
◀  High-Level Architecture All views Journey — Premiere Night  ▶
Solution Architecture v1.0 · Google Cloud · Media Platform Architecture · 2026-10 · 22 views · 15 architecture decision records03-actors-and-journeys