Your Managers See Only Charts. Is Your Employee Survey Really Anonymous?

Your Managers See Only Charts. Is Your Employee Survey Really Anonymous?

Last updated: 2026-11-16

In a fictional five-person support team, everyone gets a personal survey link. After the survey closes, the manager sees a bar chart with the team’s average score. In the provider’s system, each response still carries a code that can be matched to the invitation list. The same survey gives the manager a group result and leaves the provider able to identify respondents.

Before you call an employee survey anonymous, follow the data from invitation to deletion. Check who can link an answer to a person, including through provider records, exports and information the manager already knows.

What DHL AG reported

DHL AG, acting as data controller, reported unauthorized access to a data analysis and reporting environment operated by a processor it used for centrally run employee surveys. The breach was detected on 2 November 2026. According to the notification, the affected data categories were identity, contact and personnel data, and the incident affected the data of 10,622 employees in Türkiye.

On 11 November 2026, Türkiye’s Personal Data Protection Board decided to publish the notification under decision 2026/2483. The notice said the review was ongoing. As of 16 November 2026, the notice was still listed in the Authority’s register and no later public update was found. The notice leaves open whether any survey answers were affected, which provider operated the environment and how the survey handled anonymity.

Follow the records from invitation to report

  • Invitations. Find out who uploads employee names, work emails, IDs or team assignments. If the tool uses individual links or single sign-on, ask what connects the link or login to each employee. Check whether a completion record used to send reminders can also be matched to a response.
  • Response records. Submit test answers using dummy employee accounts. Have the survey administrator and provider show which fields are saved with each response, including partial responses, timestamps and invite codes. Ask for a sample export and check what the provider retains outside that export.
  • Provider analytics. Ask whether the provider combines responses with staff lists or team attributes to build reports. Check which people in HR, IT, provider support and any subcontractors can view, match or export records. Include analytics workspaces, logs and backups in the check.

‍

What can each see? — Fictional Role Visibility Diagram A responsive three-card isometric 3D diagram of a fictional setup: Employee sees a personal invitation and their own answers; Provider holds response codes and the invitation records that link them to employees; Manager sees the team average. FICTIONAL EXAMPLE SETUP What can each see? Role-based visibility boundaries in the fictional survey architecture 01 / EMPLOYEE Employee Sees a personal invitation and their own answers. Sees a personal invitation and their own answers. Personal Invitation Own Answers Personal Invitation Own Answers 02 / PROVIDER Provider Holds response codes and the invitation records that link them to employees. Holds response codes and the invitation records that link them to employees. Response Codes Invitation Records Response Codes Invitation Records 03 / MANAGER Manager Sees the team average. Team Average No Individual Data Team Average No Individual Data Fictional setup: Provider separates individual identity from Manager reporting

‍

Under Article 10 of Türkiye’s regulation on deletion, destruction and anonymisation of personal data, anonymised data must be impossible to associate with an identified or identifiable person, even when matched with other data. The test covers what the controller and recipients can do using techniques suited to the data environment and the activity, including reversing the anonymisation or matching records. If the provider can use an invitation list to connect a response code to an employee, those response records are still personal data.

Test what someone could infer

Small groups. Test a scenario where all five employees respond. Compare the full team average with a filtered average for the four day-shift employees. If both are exact averages of the same question and the response counts are known, the two charts reveal the fifth employee’s score. Check overlapping filters, reports across survey rounds and exported charts. Set a minimum response count for each breakdown and suppress or combine categories where comparisons can still isolate someone.

Free text. A comment such as “I was the only person on call during the Sunday outage” can identify its author even after names are removed. Keep verbatim comments out of manager reports where a person could be identified. Share reviewed themes or disable free text if you cannot review it safely.

Metadata. Try matching submission times to link-click or login records. Check whether IP addresses, device information or invitation codes are logged alongside responses or can be matched across systems. Turn off fields you don’t need and test again with the launch settings.

Set access, export and retention rules before launch

  • Access. Give managers the approved group reports. Limit access to identifiable records to specific roles with a documented need. Sign in with each test role and check what it can see, including provider support access where applicable.
  • Exports. Disable individual-response downloads by default. If an export is needed, specify who can approve it, where it may be stored and when it will be removed. Check that the same rules cover scheduled reports and API access. Where the platform supports it, log and review who exported what.
  • Retention. Set a deletion date or trigger for each set of records, including invitation mappings, raw answers, comments, logs and downloaded copies. Confirm when backup copies expire and what happens after a restore. Ask for evidence that the provider applies the agreed schedule.

Use wording your survey design supports

If the provider can connect answers to employees and managers receive group results only, describe those limits in the invitation.

The survey provider can link your response to your personal invitation. Your line manager receives group results only. We combine results from small groups with a larger group before sharing them.

If HR can also access individual responses, name that role and explain why. Describe how comments are handled. If identifiable responses are collected and anonymised later, state when that happens and who has access before then. Use “anonymous responses” only for the stage and dataset where your checks support that claim. Keep the description consistent across the invitation, survey and privacy notice.

Before you send the first invitation, put this question to your provider in writing. What employee-level records do you keep, which ones can be linked to survey responses, and who can export them? Ask for a sample export and verify the answer against your settings.

If you need help reviewing provider evidence and tracking access and retention controls across survey cycles, explore Kooch’s ongoing compliance support.

Sources

‍

Masoud Salmani