A data security posture management (DSPM) platform can look busy and still miss the data that matters, because coverage gaps, stale scans and unreviewed findings do not show up on a dashboard of open alerts. This health check is a recurring, measured review of whether the DSPM platform is doing its job: finding sensitive data across the estate, classifying it correctly, getting exposure findings fixed, and staying safe itself while it holds read access to your most sensitive stores. It is for data security leads, cloud security engineers and the internal audit or risk teams who sign off the results each quarter.
It is not a hardening guide. How to lock down the platform's admin roles, connector identities, scanners and stored samples is covered in Securing the DSPM platform itself, and what DSPM is and how to deploy it is in the Data Security Posture Management (DSPM) guide. This page turns both into numbers you can track over time, with formulas, thresholds and a findings log you can reuse every quarter.
What this health check covers
The review measures four things, each with its own group of metrics:
- Coverage: how much of the data estate the platform can actually see, and how recently it looked.
- Accuracy: whether classification results are right often enough to act on, and whether sensitive stores have accountable owners.
- Outcomes: whether exposure findings (public access, broad access, shadow copies) are fixed within agreed times, and whether suppressions are controlled.
- Platform safety: whether the platform's own console accounts, connector identities and retained data samples stay within policy.
Who runs it: the DSPM platform owner collects the data; a data security lead who does not administer the platform reviews it; the CISO or a delegated risk owner signs it off. What it is not: a test of the DSPM product's detection engine against a vendor benchmark, or a privacy compliance audit. It answers one question: is the platform we already run giving us an accurate, current and safe view of sensitive data, and are we acting on it?
Use it as a DSPM audit checklist when you take over a platform, after a major cloud migration or acquisition, and every quarter as a standing control review. The metrics map to NIST CSF 2.0 ID.AM-07 (inventories of data and their metadata are maintained) and PR.DS-01 (data at rest is protected), and to NIST SP 800-53 Rev. 5 PM-5(1), the inventory of personally identifiable information.
Key takeaways
- Coverage comes first. A DSPM that sees 60 percent of data stores produces a confident picture of the wrong 60 percent; measure coverage against an independent inventory, not against the platform's own list.
- Freshness is a metric, not a setting. Scans that silently fail or stall leave findings months out of date; track the share of stores scanned inside the agreed window.
- Sample the classifier. A small quarterly sample of findings checked by a person tells you whether labels are trustworthy enough to drive access decisions.
- Measure fixes, not findings. The number of open findings says little; the age of critical exposure findings and the share closed within service level shows whether the programme works.
- Audit the auditor. Connector identities with read access everywhere, console accounts and retained data samples are themselves sensitive and belong in every quarterly review.
How to use this template
Copy the metrics table and the findings log into your own tracker, then run this DSPM platform health check at the cadence set out below. Collect each figure from the data sources described, score it against the green, amber and red thresholds, and record every amber or red result as a finding with an owner and a due date. Keep each completed review so you can compare results quarter on quarter: the trend matters as much as any single score.
Cadence and sign-off
| Review | Frequency | Owner | Approver | Evidence kept |
|---|---|---|---|---|
| Scan health and connector errors | Weekly | DSPM platform owner | Data security lead | Scan status export, error tickets |
| Critical exposure findings ageing | Monthly | Data security lead | Head of security operations or CISO delegate | Findings export with open and close dates, exception register |
| Full health check (all metrics) | Quarterly | DSPM platform owner | CISO or risk owner | Metrics table, findings log, classifier sample worksheet |
| Connector identity and console access review | Quarterly | Identity team | Data security lead | Role assignment export, credential age report |
| Coverage reconciliation against asset inventory | Quarterly, and after any merger, migration or new cloud account | Cloud security engineering | Data security lead | Reconciliation worksheet, list of unconnected stores with reasons |
| Independent assurance | Annually | Internal audit | Audit committee | Audit report, management responses |
Metrics
NHI Mgmt Group recommended thresholds; adjust to your risk appetite. "Sensitive store" means a data store holding data in your top one or two classification levels, as defined in your Data Classification Policy Template or equivalent.
| ID | Metric | Formula | Data source type | Green | Amber | Red |
|---|---|---|---|---|---|---|
| DSPM-01 | Data store coverage | Data stores connected and scanned at least once / data stores in the independent cloud, database and SaaS inventory | Cloud provider resource inventory, CMDB, DSPM connector list | ≥ 95% | 85% to below 95% | < 85% |
| DSPM-02 | Sensitive store coverage | Sensitive stores connected / stores tagged sensitive in the asset register or by data owners | Asset register tags, DSPM connector list | 100% | 95% to below 100% | < 95% |
| DSPM-03 | Scan freshness | Connected stores with a completed scan inside the agreed window (for example 7 days) / connected stores | DSPM scan history export | ≥ 95% | 85% to below 95% | < 85% |
| DSPM-04 | Scan failure rate | Scans ending in error or partial completion in the period / scans started in the period | DSPM scan history export, scanner logs | < 2% | 2 to 5% | > 5% |
| DSPM-05 | Classification precision (sampled) | Sampled sensitive-data findings confirmed correct by a reviewer / findings sampled (minimum 50 per quarter) | Reviewer sample worksheet | ≥ 90% | 75% to below 90% | < 75% |
| DSPM-06 | Sensitive store ownership | Sensitive stores with a named, current business owner / sensitive stores found | DSPM inventory export joined to HR or directory data | ≥ 95% | 80% to below 95% | < 80% |
| DSPM-07 | Critical exposure findings within service level | Critical findings closed within the agreed time (for example 14 days) / critical findings due in the period | Findings export, ticketing system | ≥ 90% | 70% to below 90% | < 70% |
| DSPM-08 | Publicly exposed sensitive stores | Sensitive stores with public, anonymous or link-based access / sensitive stores | DSPM exposure findings, cloud storage access policy export | 0% | Any, each with an approved exception | Any without an exception |
| DSPM-09 | Over-broad access to sensitive stores | Sensitive stores reachable by external principals, all-users groups or more identities than the agreed ceiling / sensitive stores | DSPM access graph export, IAM policy export | < 5% | 5 to 15% | > 15% |
| DSPM-10 | Shadow sensitive data | Sensitive data locations outside approved stores (snapshots, exports, test copies, personal shares) / all sensitive data locations | DSPM inventory export, approved store list | < 5% | 5 to 10% | > 10% |
| DSPM-11 | Controlled suppressions | Suppressed or accepted findings with an owner, reason and expiry date / all suppressed findings | Findings export, exception register | 100% | 90% to below 100% | < 90% |
| DSPM-12 | Connector identity hygiene | Connector identities using federation or credentials rotated within policy, with no unapproved write permission / connector identities | Cloud IAM role and key age export, DSPM connector configuration | 100% | 90% to below 100% | < 90% |
| DSPM-13 | Privileged console access | Console admin accounts that are named, behind SSO with phishing-resistant MFA and reviewed this quarter / console admin accounts | DSPM role assignment export, IdP sign-in and audit logs | 100% | 95% to below 100% | < 95% |
Two companion counts belong next to the table without a threshold: the number of sensitive stores found for the first time this quarter (rapid growth often means a new team or pipeline), and the volume of stored data samples held in the platform against your retention rule. Track all thirteen metrics as a trend: one red quarter is a finding, three amber quarters in a row is also a finding.
How to collect the data
Pull every figure from a source the platform administrator cannot quietly edit, and keep the raw exports with the quarter's evidence. The data source types below are generic; map each to the equivalent export in your own tooling.
- Build the denominator first (DSPM-01, DSPM-02). Export the resource inventory from each cloud provider's own asset inventory or configuration service, the database list from the CMDB, and the SaaS application list from the identity provider's application catalogue. De-duplicate, then compare with the DSPM connector list. Every store not connected goes on a reconciliation worksheet with a reason: out of scope, not supported, pending, or unknown. "Unknown" is a finding.
- Read scan history, not scan schedules (DSPM-03, DSPM-04). Export completed scan records with start time, end time, status and objects scanned. A scheduled scan that ran but read zero objects counts as a failure. Cross-check a few stores against the scanner's own logs or the data store's access logs to confirm the scanner actually read data.
- Run the classifier sample (DSPM-05). Draw a random sample of at least 50 sensitive-data findings across data types and store types. A reviewer with authority to see the data confirms or rejects each one, recording the reason for any false positive. Repeat with a small sample of stores the platform rated as not sensitive to catch false negatives; report that count alongside, as it has no reliable denominator.
- Join ownership data (DSPM-06). Join the DSPM inventory's owner field, or the store's owner tag, to the directory. An owner who has left or moved role counts as no owner.
- Export findings with dates (DSPM-07 to DSPM-11). Export findings with severity, first seen, status, close date and suppression fields. Match critical findings to tickets in the ticketing system to confirm the close reason is a real fix, not an automatic close because the scan stopped seeing the store.
- Review the platform's own identities (DSPM-12, DSPM-13). From cloud IAM, export the roles and keys used by DSPM connectors with their permissions and credential ages. From the identity provider, export sign-in and audit logs for the DSPM application and the platform's own role assignments. The identity team, not the platform owner, should run this export.
Keep the method stable between quarters. If you change the inventory source or the sampling method, record the change and re-baseline, or trends will mislead. The Security Metrics Dictionary Template is a good place to record each metric's definition.
Red flags that need action now
- A data store that holds regulated data (payment, health or special category personal data) is not connected to the platform, and no one can say why.
- Any sensitive store is publicly or anonymously readable without an approved, dated exception.
- Scan freshness has fallen below 85 percent, or a whole cloud account or subscription has not been scanned for more than a month.
- Critical findings are closing because the store disappeared from scan results, not because access was fixed.
- A connector identity has write or delete permission on data stores that was never approved for automated remediation.
- A connector key or secret is older than your rotation policy, or is shared between environments.
- Console admin accounts include shared, local or non-SSO logins, or former staff.
- Suppressions have no expiry, or one person both suppresses findings and approves the suppressions.
- Data samples or full content copies are retained in the platform longer than the agreed retention period.
Treat any of these as an incident to triage, not an item for next quarter. If you suspect the platform or its connectors have been misused, follow If your DSPM platform is compromised.
Findings log
Record each threshold breach or red flag as a finding. Copy this table into your tracker and keep it with the quarter's evidence.
| ID | Finding | Metric | Severity | Owner | Due date | Status |
|---|---|---|---|---|---|---|
| HC-DSPM-01 | [e.g. 14 production database instances in the new region are not connected] | DSPM-01 | [High] | [Cloud security engineering] | [date] | [Open] |
| HC-DSPM-02 | [e.g. classifier precision 68% for national identifier data type] | DSPM-05 | [Medium] | [DSPM platform owner] | [date] | [Open] |
| HC-DSPM-03 | [e.g. 3 critical public-exposure findings older than 30 days] | DSPM-07, DSPM-08 | [Critical] | [Data owner] | [date] | [Open] |
| HC-DSPM-04 | [e.g. connector key last rotated 400 days ago] | DSPM-12 | [High] | [Identity team] | [date] | [Open] |
Use severity consistently: Critical for regulated or sensitive data exposed now, High for blind spots over sensitive data or platform identities outside policy, Medium for accuracy and process gaps, Low for record-keeping gaps.
Remediation map
| Finding type | Where the fix is described |
|---|---|
| Coverage gaps, unsupported store types, deployment model | DSPM Implementation Playbook |
| Classification errors, missing or unclear sensitivity levels | Data Classification Guide |
| Connector identities, credentials and write permissions | Securing the DSPM platform itself, connector identities and remediation guardrails sections |
| Console admin roles, MFA, stored samples and logging | Securing the DSPM platform itself, admin roles and logging sections |
| Over-broad access to sensitive stores | Least Privilege Techniques Guide |
| Missing data owners, process and maturity gaps | Data Security Maturity Model |
| Capability gaps that need a different product | DSPM Buyer's Guide |
Related guides
- Securing the DSPM platform itself: the hardening baseline this health check measures.
- Data Security Posture Management (DSPM): what DSPM does and how it fits with other data controls.
- Data Security Maturity Model: where these metrics sit in a wider data security programme.
- Security Metrics and KPIs Programme Guide: how to report metrics like these to leadership.
- If your DSPM platform is compromised: the response plan when a red flag turns out to be misuse.
References
- NIST: Cybersecurity Framework 2.0
- NIST: SP 800-53 Rev. 5, Security and Privacy Controls for Information Systems and Organizations
- NIST: SP 1800-28, Data Confidentiality: Identifying and Protecting Assets Against Data Breaches
- CIS: Control 3, Data Protection
- CIS: Control 5, Account Management
- CISA and NSA: Top Ten Cybersecurity Misconfigurations (AA23-278A)
Related NHI Mgmt Group resources: Securing the DSPM platform itself, DSPM Implementation Playbook, Data Classification Guide and Endpoint privilege management (EPM) health check.