Skip to main content
Orion Five Engineering

10 Jalan Kilang #04-05, Singapore 159410
+65 6100 5505

Book a scoping call
← All insights

AI governance

Build data retention rules into the system's data layer

A retention policy in a document depends on people following it. A retention rule built into the database schema is enforced by the system. The bunkering system at PSA Marine purges a face record six months after the account is deactivated. The purge is part of the system, so it needs no procedure and nobody has to remember to run it.

Terence Kok · 2026-08-26 · 4 min read

A brushed steel rack of thin white plates standing on edge like index cards, a red index tab clipped to the front plate and the rearmost plate sliding out of the back of the rack.

Executive summary

6 months

after deactivation, an account and its face record are purged automatically

5 years

the period for which fuel transaction reports are retained

30 days

the period for which CCTV footage is held before it is overwritten

Key conclusions

  • Retention involves three separate clocks for three kinds of data, and each needs a different period. A single policy sentence cannot cover all three.
  • The Retention Limitation Obligation in Singapore's Personal Data Protection Act (PDPA) is met when the system deletes data as soon as the purpose for keeping it ends.
  • When retention is built into the data layer, an auditor can check it with a database query instead of an interview.

Most retention policies are accurate when they are written and inaccurate within a year. A policy depends on somebody running a procedure, and a procedure depends on the person who knows it staying in the job.

The biometric fuel management system at PSA Marine's Brani and West Coast bases has run continuously since 2019. It verifies a launch master's face before fuel flows. The system therefore holds biometric data about identifiable people, in a security application, for a critical-infrastructure operator. That is close to the most demanding retention case a Singapore integrator will meet. The way the system handles it is worth describing, because its retention rules are built into the design instead of written in a policy.

One system applies three retention periods to three kinds of data

  • Identity: a deactivated account and its face record are purged automatically six months after deactivation. The biometric record exists only while the person it authenticates is still authorised.
  • Transactions: fuel transaction reports are retained for five years. They form the operational and financial record, and they outlive the identity record that created them.
  • Surveillance: CCTV footage is held for 30 days. That is long enough to investigate an incident once it is known, and short enough that the footage does not become an archive of everyone who walked past.
  • Across all three: access follows the principle of least privilege, so a person's role decides who can see a record before any retention rule applies.

How long the system keeps each kind of record

CCTV footage
30 days
Face recordafter deactivation
6 months
Fuel transaction reports
5 years
The three retention periods in the PSA Marine biometric fuel management system. Bars are drawn to scale in days.

The PDPA requires deletion once keeping the data no longer serves its purpose

A white mechanical timer with a red pointer beside a short stack of white cards.

The PDPA's Retention Limitation Obligation is simple to state: cease to retain personal data as soon as it is reasonable to assume the purpose for collecting it is no longer served by keeping it. The difficulty is that 'as soon as' requires a timer, and that timer has to run somewhere in the system.

Biometric data raises the stakes. The Personal Data Protection Commission's guidance on the responsible use of biometric data in security applications treats it as data that warrants a higher standard of protection. The reason is that a face, unlike a password, cannot be reissued after it leaks. A six-month purge tied to deactivation meets that standard: the system holds as few faces as it can, for no longer than it needs to.

Retention built into the data layer runs and can be audited without relying on staff

Retention as a policyRetention in the data layer
A document says records are deleted after six monthsA scheduled job deletes them, and logs that it did
Someone runs the purge, when they rememberThe purge runs whether anyone remembers or not
An auditor interviews the administratorAn auditor runs a query for the oldest record and reads the date
A departing employee takes the procedure with themThe procedure is in the schema, and the schema stays

Specify four retention details before the build starts

  1. Every class of personal data the system will hold, and the purpose that justifies holding each one.
  2. For each class, the event that ends the purpose (deactivation, delivery or incident closure) and the retention period after that event.
  3. Where the purge runs, what it logs and how a reviewer confirms that it ran.
  4. The one class that must be kept longer than the others, and why. This is usually the transaction record, which is kept for a financial or regulatory period that the identity record does not need.

Only a retention rule that the system enforces stays accurate after the people who wrote it have left.

Related insights


All insights

Related service


Book a scoping call