A DSPM scanner crawls every storage asset it can reach — declared buckets and databases, plus "shadow data": undocumented backups, test dumps and forgotten clones nobody registered. For each object it runs pattern-matching classifiers (credit-card checksums, SSN-shaped digit groups, medical record codes) to tag PII, PCI or PHI content, then folds that into a risk score.
risk = sensitivity_weight
× (public_access ? 2 : 1)
× (encrypted ? 0.4 : 1)
- Sensitivity — none / PII / PCI / PHI, discovered by the scanner's pattern classifiers as it visits each store.
- Shadow data — dim, unlabeled nodes with no owner metadata; the riskiest surprises tend to hide here.
- Access exposure — a red access ring means the store is reachable without internal credentials.
- Encryption — an open lock icon means data at rest is unencrypted, multiplying any sensitivity risk.
Real-world relevance: DSPM tools (a defined cloud-security category) exist because sensitive data sprawls across accounts faster than manual inventories can track it — the highest-value finding is usually a forgotten copy nobody knew was exposed.