The Structural Brand Power Index is the scoring pass. Nightly research updates node properties. The weekly editorial reads those updates as a diff and writes what changed. The composite (client graph plus REA plus BMC) is what makes the diff readable. Without it, nightly research produces a pile of facts. With it, the same facts arrive already typed, already placed on the canvas, already positioned in the exchange ledger.
SBPI is nine divergent rubric families, not one score. The choice of family decides the story a given week tells. That is not a weakness; it is the correct shape for a regulatory-grade instrument. The composite is what makes the rubric choice audit-friendly. Every score walks back to nodes an agent (or a reviewer) can name. Every claim in the weekly report traces to node properties an agent can cite.
Two changes make SBPI a pass over the composite rather than a standalone product. First, node updates from nightly research write into the client bundle's frontmatter and body, so the graph state after Sunday night is the state the Monday editorial reads. Second, the editorial pass reads with REA and BMC typings turned on. A story about a channel shift is not a paragraph about "the channel;" it is a paragraph about a Channels cell and the Economic Events that fired through it this week. The prose gets tighter because the typings do the disambiguation the writer used to do.
How to apply it. Point nightly research at the composite. Log the rubric-family choice on every editorial pass. Trace every claim to a node in the graph. The next post is OKF, the substrate that makes the whole stack portable between agents and teammates.