diff --git a/docs/guides/observability/querying-the-supply-chain-graph.md b/docs/guides/observability/querying-the-supply-chain-graph.md
index 657aa2b..b4acaa8 100644
--- a/docs/guides/observability/querying-the-supply-chain-graph.md
+++ b/docs/guides/observability/querying-the-supply-chain-graph.md
@@ -50,6 +50,35 @@ run every command verbatim.
 
 ---
 
+## Where the data lives
+
+The `supply-chain-graph/` folder on `main` ships the **schema**, the
+**`examples/` fixtures** used in this guide, and any hand-authored records. The
+**real, workflow-generated events are not on `main`** — each producer stages one or
+more immutable event files per run and the `record-graph-events` collector commits
+them (append-only) to a dedicated **`supply-chain-graph-data`** branch.
+
+To query your pipeline's actual events, fetch that branch and index it:
+
+```bash
+git fetch origin supply-chain-graph-data
+git worktree add --detach /tmp/scg-data origin/supply-chain-graph-data
+cssc-graph index /tmp/scg-data/supply-chain-graph -d /tmp/scg.db --rebuild
+```
+
+The branch ships its own `schema/`; from
+[#203](https://github.com/toddysm/cssc-framework/issues/203) onward the collector
+keeps it in step with the code, so the command above needs no `--schema-dir`. (If
+you hit an `unknown kind` error against an older branch snapshot, pass
+`--schema-dir` pointing at a current `supply-chain-graph/schema`.) In the cluster,
+`graph-service` gets the same data through its optional git-sync init container,
+which syncs `supply-chain-graph-data` into `DATA_ROOT`.
+
+The `examples/` walkthrough below uses the in-repo fixtures instead, so every
+command runs verbatim on a fresh clone.
+
+---
+
 ## Experience 1 — the `cssc-graph` CLI
 
 ### Install