Glossary. DSGM.
What is DSGM (data security governance and management)?
DSGM is a findings list plus the record of what wasn’t checked and who decided: the half of an audit answer that says how you know.
By Dipal Panchal, co-founder and CEO · Updated November 7, 2026
Findings, plus the record around them
Data security governance and management (DSGM) is the term Sylox uses for finding sensitive data and its exposure, with a record of what was checked, what wasn’t and who decided. Unlike data security posture management (DSPM), which is built to find and reduce data risk, DSGM is built around that record. Some DSPM tools keep part of it, so ask any tool to show it. Without the record, a team can say where its data is and still can’t say how it knows the list is complete.
DSGM isn’t an analyst category as of November 2026. Analysts and vendors write about data security governance, meaning the policies that protect data. DSGM names the record a team needs when an auditor, regulator or customer asks it to prove where its sensitive data is. Under GDPR Article 5(2) and its UK GDPR counterpart, a controller must be able to demonstrate compliance with the data protection principles, and a record of what was checked is part of that evidence.
One audit question, answered two ways
PCI DSS v4.0.1 requirement 12.5.2 has an entity confirm its scope, including every location where account data is stored, processed and transmitted. The second question is ours: how do you know the list is complete?
| What the auditor needs | From a findings list alone | From findings plus a coverage record |
|---|---|---|
| Where account data sits | The locations where card numbers were found, ranked by risk | The same locations, each with its source, scan date and the reviewer’s decision |
| Whether the list is complete | Can’t say: the list shows what was found and is silent on where nobody looked | Every source in scope, with its status: scanned, partial, failed, excluded or not connected |
| A source that failed to scan | Invisible, because it contributes no findings | Listed as failed, with the reason and an owner to fix the connection |
| A source the owner excluded | Invisible | Listed as excluded, with the owner’s written reason |
| What changed after a fix | A later list without the finding | A dated re-scan showing the finding gone, beside the decision that closed it |
The coverage record behind that answer
The team’s record for one estate, built from per-source scan results and owners’ entries, following card numbers only. Three rows hold no findings for reasons a findings list can’t show.
| Source | Status | Reason | Owner |
|---|---|---|---|
| Billing database (PostgreSQL) | Scanned, October 28, 2026 | The tables the scan account can see, a sample of each; card numbers in billing.payment_card | Payments engineering |
| Settlement exports (cloud storage) | Partial, October 28, 2026 | Card numbers in one CSV export; two password-protected exports listed, not read | Finance operations, to supply the two exports by November 13, 2026 |
| Payment gateway logs (SaaS) | Failed | Credentials expired on October 20, 2026; nothing was read | Platform team, to reconnect by November 13, 2026 |
| Legacy billing (Oracle) | Excluded | Decommissioned June 30, 2026; the purge record covers the instance, its backups and exports | Payments engineering; exclusion approved by the CISO on July 2, 2026 |
| Support desk (SaaS) | Not connected | No connection set up, so card numbers pasted into tickets are unknown | Support lead, to decide by November 30, 2026 |
Why the failed and unconnected rows matter
An auditor testing completeness picks a place the card data could be and checks the record covers it. With a findings list alone, the gateway logs and the support desk look identical to clean sources. In the coverage record they are two open items with owners and dates. A source row covers one instance, so replicas, snapshots and extracts of it are rows of their own.
The excluded row needs the owner’s evidence beside it. “Decommissioned” with a purge record is a reason; “out of scope” with nothing attached is a gap with a label on it. Each scan supports the date it ran, so a period needs the series of dated scans.
When a DSPM tool is the better fit
If your main goal is to find and reduce data risk across a large cloud and SaaS estate, and audit evidence is produced another way, a DSPM tool built for that scale fits better. If an auditor, regulator or customer needs to see what was checked, with the gaps on record, the DSGM record is the part that answers them. Mapping shows relevance; your auditors judge whether a control is met.
A team can run both: DSPM to reduce risk, and the coverage record for the audit or privacy team to hand over. A DSPM tool that already records what it couldn’t read, and keeps owners’ decisions, produces most of a DSGM record. To test one, ask for its output on a source you know it failed to read.
A practice with a tool inside it
DSGM is an operating capability, not a purchase, with four parts: people and skills, policy and process, technology, and assurance. A tool can produce the findings and the scan statuses; people own the exclusions, the decisions and the in-scope list itself.
Our rule for a DSGM record: every source in scope has a status and an owner, every finding has a decision, and every exclusion has the owner’s evidence. A record missing any of the three answers “where is our data?” and leaves out “how do you know?”
IRIS records findings and source coverage from dated point-in-time scans of the sources you connect, scheduled or on demand. Capabilities vary by edition.
Findings with their source and scan date: IRIS produces part of this, in the sources you connect.
A scan status for each source in a scan, with a reason for anything it couldn’t read: IRIS produces this.
SaaS sources nobody connected: Not produced by IRIS.
A reviewer’s decision on each database finding: IRIS produces part of this, a person decides; IRIS records the decision.
Owners, due dates and reminders for open items: Not produced by IRIS.
A dated re-scan after your team makes the fix in the source system: IRIS produces this.
Explore audit evidenceCheck the second half of the question
Take the question your auditor asked in your last audit and check your record answers both halves: where the data is, and how you know the list is complete. If it answers only the first, add a status, a reason and an owner for every source in scope before the next fieldwork. 12.5.2 asks for the scope check at least every 12 months and after significant change, and every six months for service providers under 12.5.2.1.
The status column is defined in what is scan coverage, and how to build a source coverage register builds the full record.
Where to get the PCI DSS text
- PCI DSS v4.0.1: requirements 12.5.2 and 12.5.2.1 (free after accepting the PCI SSC license), PCI Security Standards Council. Opened November 7, 2026.
The scan before the record, evidence after
What is sensitive data discovery?
The step every DSGM record starts from: the scan that finds where sensitive data sits.
What is DSPM?
The category next to DSGM: what DSPM tools find, and what they don’t replace.
What is audit evidence for data security?
The next step: turning findings and coverage into evidence items an auditor can retest.
DSGM next to GRC tools and DLP
Underneath it. The GRC tool holds the control framework, the risk register and the audit workflow; the DSGM record is the system-sourced evidence those controls point to.
Teams that must show auditors, regulators or enterprise customers where sensitive data lives and how they know the list is complete.
No. Data loss prevention inspects data in use, in motion and at rest, and can block a transfer. A DSGM record says where sensitive data sits, what was checked and who decided, and it blocks nothing.
Protect your brand. Start with your data.
See how IRIS records what was checked, what wasn’t and why, with evidence you can show an auditor.