Check the evidence and definitions
Case metrics need an explicit sample, time window, question set, model version, retrieval settings and calculation method. An anonymous narrative cannot replace auditable data or guarantee results for a new project.
Keep measurement layers separate
Mentions, recommendations, citations, visits and conversions measure different things. Before-and-after comparisons require comparable conditions and disclosure of model changes, channel changes and sampling uncertainty.
Correlation is not causation
An increase after publication does not establish that the content caused growth. Business attribution requires additional study design. This page does not repeat growth figures that have not been independently verified.
Project scope, delivery and fees follow the formal agreement.