Why retired fields keep showing up
A field can be marked retired in the schema and still appear on a live form, a saved report, or an automation that writes to it. The retirement decision often lives in a ticket, while the surfaces that use the field live in several tools with different owners. This test checks whether retired fields are actually gone from the places where people and systems still touch them. It is a dependency audit, not a judgment about whether the retirement was correct.
Define the retirement record
Start with the list of fields the team has formally retired during a stated period. For each one, record the object, field API name, retirement date, approver, and the replacement field if any. If the retirement exists only in a conversation, write that down as a finding rather than treating the field as officially retired. The test needs a clear population; a vague list of "old fields" will produce a vague result. Keep the list fixed before you inspect the surfaces.
Inventory the live surfaces
List the forms, reports, dashboards, list views, workflows, integrations, and exports that could reference the field. Include surfaces owned by sales, marketing, operations, and finance, because a field retired by one team may still be used by another. Record the owner and the last modified date for each surface. This inventory does not need to be exhaustive on the first pass, but it should cover every system that can read or write the object. A surface that is not in the inventory cannot be tested.
Search by API name, not by label
Search each surface for the field's API name and internal value, not only its display label. Labels change, and a retired field may still carry an old label in a saved report. Include partial matches where the tool allows them, but review the results so a similarly named active field is not counted as a hit. Record the exact location of each match and whether the surface is currently active or archived. Archived surfaces are recorded separately because they do not affect daily work.
Test forms with a real submission
For each active form that references a retired field, submit a test record and inspect what the CRM stores. A hidden field can still submit a value, and a visible field can be present in the form builder but removed from the published version. Record whether the field appears to the user, whether it writes a value, and where that value lands. Delete the test record or mark it clearly. The rendered form is the evidence, not the builder preview.
Test reports and dashboards
Open each report and dashboard that references the field and check whether it still renders, filters, or groups correctly. A report can keep working while silently dropping the field from a grouping, which changes the totals without an error. Record the report owner, the last run date, and whether the output changed after retirement. If a report cannot be opened, record that as a separate finding rather than assuming it is safe. A dashboard that nobody has opened in a year is still a live dependency until its owner confirms otherwise.
Test automations and integrations
Review workflows, triggers, and integrations that write to the field. A retired field may still be populated by an automation that no one monitors, which keeps stale data alive and can mislead a future migration. Record the automation name, owner, trigger, and whether it would fail or silently skip if the field were removed. Test in a sandbox where possible. If a production test is required, use a clearly labeled test record and document the cleanup.
Classify each finding
Group the results into active use, orphaned reference, archived reference, and no reference. Active use means a current surface still depends on the field and needs a decision. Orphaned reference means the surface exists but no one owns or uses it. Archived reference is documented and out of scope for daily work. This classification prevents a long list of matches from being treated as equally urgent. It also gives each finding a clear next owner.
Quantify the gap without inventing a benchmark
Report the number of retired fields tested, the number with at least one active reference, and the number of active surfaces involved. Report counts alongside percentages. Do not describe the result as an industry benchmark or a target. It describes this set of retirements at this time. If the team wants a target, it should set one after seeing the baseline and agreeing which surfaces count as active.
Choose the repair that removes the dependency
For each active reference, decide whether to remove the field from the surface, keep the field and reverse the retirement, or replace it with the successor field. Record the decision and the owner. Prefer removing the reference at the surface rather than re-enabling the field everywhere, because that keeps the retirement decision intact. If a report depends on historical values, keep a snapshot before changing the report. The repair should leave the surface working, not just remove a name.
Retest and document
After the repairs, rerun the same inventory and search with the same definitions. Compare the new result with the baseline and report what changed. Record any surface that was missed in the first inventory so the method improves. Keep the evidence with the retirement record. A test that is repeated after each cleanup becomes a control; a one-time audit becomes a story nobody can verify later.
Limits and data handling
The test cannot find references in tools the team does not know about, and it cannot prove that a surface is unused just because no one remembers it. State those limits in the method note. The searches may expose customer data in report outputs, so restrict access and avoid copying record-level values into the working file. Delete test records and extracts when the review is complete. The goal is a repeatable check that respects both the data and the people who depend on it.
Sources
- https://www.nist.gov/cyberframework
- https://knowledge.hubspot.com/records/understand-original-source-properties
- https://support.google.com/analytics/answer/1033863
- https://www.w3.org/TR/NOTE-datetime
- https://www.ftc.gov/business-guidance/privacy-security