Plan requirement
| Subscription | Any plan |
| Access | Admin |
Custom objects are easy to forget in a retention review, because they were built for a purpose that may have ended. They also hold personal data as readily as anything else: an object listing customers' devices, addresses or contract details falls under the same obligations as a ticket.
Check the relationships before you set anything
How a record relates to others decides what a schedule removes, and the two kinds behave very differently.
| Relationship | What happens when a record is deleted |
| Parent and child | Deleting a parent record also permanently deletes all associated child records. |
| Lookup | The related record stands on its own. The lookup field keeps its stored value but shows as empty in the agent workspace. |
The first row is the one that catches people. A schedule set on a parent object is not removing one kind of record, it is removing a tree, and nothing warns you at the moment you set the age.
Create one
- Open Admin Center > Account > Security > Deletion schedules.
- Choose the custom object.
- Set the conditions. Last updated is required on every schedule; you can add Created at and your own custom fields alongside it.
- Check what depends on those records, then enable it.
Conditions combine with All. With the Advanced Data Privacy and Protection add-on you can use All or Any, with no limit on how many.
Each object can have its own period
You can create up to ten schedules per object and run one of them at a time, so different objects can each have an active schedule with a different age. That matters, because retention here is not one number: an asset record probably outlives the tickets about it, and a temporary registration for an event does not.
With the add-on you can activate up to ten schedules per object at once.
Objects that stopped being used
Almost every account has one built for a project that ended. Those records serve nothing and are still your responsibility, and a retention review is a good moment to notice them.
Test on a small object first
Run it on something low-stakes to see the effect, then apply it more widely. The mechanism is the same and the confidence transfers, and it is a great deal cheaper than discovering the parent-child behaviour on your asset register.
Comments
0 comments
Article is closed for comments.