Content systems

When the job is a thousand of something rather than one of something, the interesting question stops being quality in the abstract and becomes quality that holds at volume.

Start a project

Quoted per pipeline. Audit first.

What you get

  • A pipeline that produces consistently, not one good example and a promise
  • Costs known per item before you commit to a run, not discovered on the invoice
  • Human review at the point where a mistake would be expensive, and nowhere it would just be annoying
  • Output checked against rules you set, and rejected rather than shipped when it fails them
  • Everything recorded, so a bad batch can be traced to what produced it

Cost per item is the whole design problem

A video pipeline built here produces long form episodes from a script. Working through it established that the shot ceiling forces roughly 330 separate visuals per episode, which sets the cost at about £65 an episode, and that a single setting quietly routing scenes to a premium generator doubled it.

Knowing that number before a run is the difference between a system you can budget for and one you switch off after the first bill.

Where humans stay in it

Volume work goes wrong in batches rather than one at a time. The review step is placed where a mistake costs money, and deliberately not placed where it would only cost patience, because a system that asks for approval on everything gets approved on everything without being read.

Tell me what is not working.

Twenty minutes, no charge, and a straight answer about what it would take and what it would cost.

Start a project