Personal data discovery

The first data map is built from what people remember. The second is built from what the systems contain, and the difference between the two is where privacy incidents come from.

The short version

Personal data discovery is the technical search for personal and sensitive data in the places an organization stores it, to find what the inventory and the interviews missed. Inventories built from interviews miss copies: exports, test databases, spreadsheets, logs and backups are where personal data most often turns up unrecorded.

Discovery complements data mapping: the map says what should be there, and discovery says what is.

Where we look

Personal data hides in layers, and the review works down the same list for each estate: databases, cloud storage, SaaS applications, files and email, then logs and backups.

Where personal data is usually found Five places, top to bottom: databases, cloud storage, SaaS applications, files and email, and logs and backups. Databases Production, test and reporting copies. Cloud storage Buckets, shares and exports. SaaS applications CRM, support, HR and marketing tools. Files and email Spreadsheets and attachments. Logs and backups Copies nobody lists.
Each layer is searched read-only and reported by location and data type.

The technical review of the places themselves is covered in database security review and cloud security review; discovery asks the different question of what personal data is in them.

What you get

  • A findings register. Each location of personal or sensitive data, its data types, its apparent owner and whether the inventory already lists it.
  • Unrecorded copies. Test databases, exports, shared folders and logs holding personal data the inventory does not mention.
  • Handling recommendations. Delete, restrict, mask or keep, with the reason and an owner for each.
  • Inputs for the data map. The register feeds the records of processing directly.

How the work runs

Discovery runs read-only, under written authorization, and findings report where data sits and what kind it is rather than copying the data out. Scope, systems and access are agreed in writing first, because the work reads production data stores. Searching is done with scripts and the classification features of your own platforms where they exist. Both. Scripts and the native classification features of your own platforms cover most estates, and a commercial scanner is recommended only where volume or change rate justifies it.

Findings are written for two readers: the system owner who has to act, and the person keeping the records of processing, who needs the location and category, not the data.

Who it is for

Organizations with many systems and no reliable inventory, those preparing for a migration, an acquisition or a deletion request they cannot yet answer, and security teams who need to know where the sensitive data is before deciding where to spend on controls. The procurement view of discovery software is in privacy tool selection.

How the work is bounded

The scope is agreed in writing before work starts, and the engagement is quoted in writing with it.

SecHB does not issue certifications, attestations or audit opinions: those come from accredited certification bodies, CPA firms and QSAs. The work here is what an organization does to be ready for them.

Nothing here is legal advice. Where a question turns on the law, the work is done alongside the client’s counsel, not instead of them.

Questions we are asked

What is personal data discovery?

Personal data discovery is the technical search for personal and sensitive data in the places an organization stores it, to find what the inventory and the interviews missed.

Where do inventories usually miss personal data?

Inventories built from interviews miss copies: exports, test databases, spreadsheets, logs and backups are where personal data most often turns up unrecorded.

Does discovery copy our data out?

Discovery runs read-only, under written authorization, and findings report where data sits and what kind it is rather than copying the data out.

Do you use a scanning product?

Both. Scripts and the native classification features of your own platforms cover most estates, and a commercial scanner is recommended only where volume or change rate justifies it.

Tell us where you think the data is

Background reading: where personal data hides. Name the systems and platforms in scope and what prompted the question. The reply says what discovery would cover. See also all privacy services.

Discuss a scope