Privacy and preparation
Consolidated legacy archives across portfolio companies: access and security controls
By SourceX Editorial · Updated
Short answer
Legacy data archive security in a holding group rests on four controls: a named owner for each acquired company's records, segregation so one entity's archive never mixes with another's, scheduled access reviews, and a tracked retention clock for every record set. Consolidate the storage if it saves effort, but keep ownership, permissions and deletion rules separate per entity.
Key takeaways
- Consolidate storage, not ownership: each archive belongs to the legal entity that collected the records.
- Separate containers and encryption keys per entity let the group answer a deletion request, a breach question or a licensing scope for one company without touching the others.
- Legacy archives need access reviews precisely because nobody opens them daily, so stale permissions go unnoticed.
- Retention clocks keep running after a system is retired, including vendor deletion windows that start when a subscription is cancelled.
- An archive register kept current is the one document that answers who owns, who can open and when to delete each record set.
What controls does a consolidated legacy archive need?#
A consolidated legacy archive needs four core controls: entity ownership, segregation, access review and retention tracking, backed by logging and encryption. Holding groups usually build a shared archive for sound reasons, such as retiring a duplicate help desk or an acquired company's on-premises servers, and then find it has become the least governed system in the group.
Each control should produce evidence a group COO can show a board, an acquirer's diligence team or a prospective data licensee without reconstructing history from memory.
| Control | What it means in practice | Evidence it is working |
|---|---|---|
| Entity ownership | Each record set is tagged to the entity that collected it and to a named business owner | An archive register with entity, system, date range and owner |
| Segregation | Each company's archive has its own container, permissions and encryption keys | Storage and key inventories that map one to one to entities |
| Access review | Only named custodians hold standing access, reviewed on a fixed calendar and after role changes | Signed review records and a list of removed accounts |
| Retention clocks | Each record set has its own keep-until and delete-after dates, plus any legal hold | A retention schedule with triggers, holds and completed deletions |
| Logging | Reads, exports and permission changes are logged where archive admins cannot edit them | Exportable access and export logs |
Who owns an archive after the acquired company is merged or retired?#
An archive belongs to the legal entity that collected the records, even after that company's systems, staff and brand have been folded into the group. Shared services can be the custodian that runs the storage, but custody is not ownership, and the difference matters when the subsidiary signed the contracts and published the privacy notices.
Mergers blur the line. When one subsidiary is merged into another, its records may pass to the surviving entity together with the customer agreements, data processing addendums and privacy promises that limited how they could be used. Those obligations generally travel with the records rather than resetting at closing, so ask counsel how they apply to each record set.
Record the answer in an archive register, with one entry per record set rather than per company, because a single acquisition often leaves several systems behind.
- Original legal entity, and the successor entity if it was merged
- Source system, such as an old Zendesk instance, a self-hosted GitLab server or an on-premises SQL Server database
- Record families and the date range actually present in the export
- Governing customer agreements, DPAs, privacy notice versions and vendor terms
- Business owner who approves use, and custodian who runs the storage
- Retention rule, deletion trigger and any active legal hold
- 01Your work records
- 02Personal details removed
- 03You approve the release
- 04Licensed to AI developers
How should archives from different portfolio companies be segregated?#
Archives from different portfolio companies should be segregated so each entity's records can be found, restricted, released or deleted without touching another company's records. Segregation lets a group answer a narrow question narrowly: one customer's deletion request, one subsidiary's breach analysis or one company's licensing scope.
Shared storage is fine; shared boundaries are not. The common failure is one cloud bucket with a folder per acquisition and one admin group over all of them. A permission granted at the top flows down to every company at once, and a search index built across the share quietly joins records that different contracts governed.
| Model | How it is built | When it fits |
|---|---|---|
| Separate account or tenant per entity | Own storage account, admin roles and encryption keys | Default for entities with customer data, distinct contracts or a likely future sale |
| Separate container per entity in a shared account | One account, with container-level permissions, per-entity keys and no inherited access from the top | Workable when shared services run the platform and permissions are tested |
| Shared folders under one permission set | Folders by company, one admin group, one key | Avoid for anything beyond non-sensitive internal documents |
| Cross-entity search index | One index over several companies' archives | Only for a defined, approved purpose, built on prepared copies rather than raw archives |
How do you review access to archives nobody opens every day?#
Access to a legacy archive should be reviewed on a fixed calendar and after every role change, because archives are where stale permissions survive longest. A live help desk shows its users every morning; an archived export in cold storage can keep a departed engineer's access for years without anyone noticing.
Default to no standing access. Give named custodians the minimum they need, route other requests through a ticket with a business reason and an end date, and keep a logged break-glass account. Close shared admin logins inherited from an acquired company's IT contractor during integration, not at the next audit.
- Export the current permission list for each entity's archive, including service accounts and API keys.
- Send each business owner the list for their entity only, to confirm or remove every entry.
- Remove anything not confirmed, then check that the removal took effect.
- Compare access logs with the confirmed list to catch unexplained use.
- File the signed review with the archive register and set the next review date.
Which retention clocks keep running after a system is retired?#
Retention clocks keep running on three fronts after a system is retired: the group's own legal and contractual obligations, the vendor's deletion window on any cancelled subscription, and deletion rules configured inside the old system. Missing one can destroy records the group still needs, or keep records it promised to delete.
Vendor windows are the easiest to miss because they start when someone cancels a subscription to save money. Atlassian states that after a cancelled subscription period ends, the site stays accessible for 15 more days and is then deactivated, with data retained for 60 days on Free, Standard, Premium and Enterprise plans (15 days for trials) before permanent deletion. Microsoft says it keeps customer data in a limited-function account for 90 days after a cloud subscription ends so the customer can export or renew.
Old systems can also carry their own deletion automation. Zendesk admins can create ticket deletion schedules that keep deleting archived tickets matching their criteria, and deleted tickets cannot be restored, so check for active schedules before assuming the full history was exported. Confirm every vendor window against current documentation for the plan.
| Clock | Who sets it | What to record in the register |
|---|---|---|
| Statutory and tax retention | Laws that apply to the record type, as advised by counsel and accountants | Minimum keep-until date per record set |
| Contractual deletion | Customer agreements and DPAs the acquired company signed | Deletion triggers such as contract end, and proof of completion |
| Vendor deletion window | The SaaS vendor's terms for cancelled or expired subscriptions | Cancellation date, export date and a completeness check |
| In-system automation | Deletion schedules or retention policies set in the old tool | Whether rules were active, paused or carried into the archive copy |
| Legal hold | Counsel, when litigation or an investigation is expected | Hold scope, start and release; a hold overrides deletion |
Illustrative: a software holding group untangles four acquisitions#
Illustrative: a fictional group that buys vertical market software companies had acquired a self-storage management platform, a field service tool for pest control operators, a fleet inspection app and an accounting add-on for construction firms. During integration, shared services exported each company's old help desk, issue tracker and CRM into one cloud storage account, with a folder per company and a single admin group drawn from group IT.
The trigger was a routine question. The fleet inspection CEO wanted to explore licensing its support and engineering history, and the group COO could not say which folders held only that company's records, who had opened them since closing, or whether a deletion schedule had run before export.
The group split the account into a container per entity with separate keys, named an owner and a custodian for each, revoked the shared admin group and ran a first access review. The fleet inspection company could then scope a licensing review to its own records alone, and a later deletion request for the accounting add-on was handled inside one container.
Preparing an archive for licensing without weakening its controls#
Preparing an archive for data licensing should happen on a copy in a separate, time-limited workspace, so the archive stays the controlled system of record. Preparation work, such as removing personal details or sampling records for review, needs broader tool access than an archive should grant.
Keep the entity boundary here too. A preparation workspace should hold one company's records for one approved scope, with its own access list, and be destroyed or sealed when the work ends. Log which record set was copied and under whose approval, who used the workspace, what left it and when it was closed.
How SourceX works with holding groups on archived records#
SourceX treats each portfolio company's archive as its own supply, even when the group runs the storage centrally. In the SourceX five-step transaction, Supply and Rights are scoped per entity, Preparation runs on copies the supplier controls, and the supplier approves the prepared output before Delivery. Large archives stay in the group's own storage or ship on encrypted drives.
The archive register, access reviews and retention schedule feed the SourceX Evidence Packet, which records provenance, licensing rights, permitted use, the privacy record and release authorization for each package. The initial fit check collects only metadata, such as systems, date ranges and record families.
Frequently asked questions
Is it better to keep a retired system running than to archive it?
Sometimes. A read-only license on the old system preserves search, links and attachments that exports often lose, but it keeps a vendor relationship, user accounts and a deletion window alive. Many groups keep the old system only until the archive export is verified complete against the source.
Can the parent company's IT team be the custodian for every subsidiary?
Yes, if custody is documented as a service to each entity rather than ownership. The subsidiary's business owner still approves who gets access and how records are used, and the custodian acts on those approvals. Check that intercompany agreements cover archived records as well as live systems.
What happens to an archive when the group sells a portfolio company?
The purchase agreement usually decides, so keep the archive register ready before a sale process starts. Common outcomes are moving the archive to the buyer, keeping a copy under a transition services agreement for a defined purpose, or destroying the group's copies. Per-entity segregation makes each option practical.
Should archived records be pseudonymized before they are stored?
Not by default. The archive is usually the authoritative copy, needed for disputes, audits and customer requests, so altering it can create problems. Strong access controls protect the archive itself, and personal details are removed on prepared copies made for a specific approved use, such as a licensing package or an analytics project.
Do privacy laws apply to records in a cold archive?
Privacy laws may apply to personal information wherever it is stored, including archives nobody opens. Access and deletion requests can reach archived records, and breach obligations can cover them. Which laws apply depends on whose data the records hold and where those people are, so assess each entity's archive with counsel.
Sources
- Atlassian states that after a cancelled subscription period ends, the site stays accessible for 15 more days and is then deactivated; data is retained for 15 days for trials and 60 days for Free, Standard, Premium or Enterprise plans, then permanently deleted and unrecoverable. Source
- Microsoft's Trust Center says that when a cloud subscription ends or expires (free trials excluded), Microsoft keeps customer data in a limited-function account for 90 days so the customer can export it or renew. Source
- Zendesk admins can create ticket deletion schedules that delete archived tickets after a set period. Deleted tickets cannot be restored, and the schedules keep deleting any tickets that match their criteria. Source
Related resources
See if your company qualifies
A short company assessment. No data uploads are needed.