For works councils · DTHgroup
The works council
is not the brake.
It wants to say yes.
It protects the workforce – that is its job, and it takes it seriously. The instrument has to make agreement possible. This page sets out what HSflow collects, what is technically ruled out, and what you should examine before you agree.
✗ What you are right to warn about
Systems that make individuals assessable. Analysis down to team level. Psychological profiles. Data from which performance comparisons can later be built.
↓
✓ How HSflow is built
No personal data, no individual analysis, no psychodiagnostics. Monitoring of performance or conduct is technically impossible – not merely not intended.
0personal profiles
GDPRlegally reviewed
A note on jurisdiction
This page describes German co-determination law under the Works Constitution Act (BetrVG). Employee representation exists in other countries under different names and with different powers – the CSE in France, the RSU and the safety representative RLS in Italy. The design principles described here (no personal data, no individual analysis, minimum group sizes) apply everywhere; the statutory references do not. Written for organisations with German operations.
German law
Your legal position
Two provisions,
and how we handle them.
We set this out openly, because you will examine it anyway – and because a vendor who plays down co-determination is not a good vendor.
§ 87 (1) no. 6 BetrVG
Co-determination on the introduction and use of technical devices designed to monitor the conduct or performance of employees. Digital survey instruments regularly fall within this under prevailing opinion – the objective capability to monitor is sufficient, regardless of intent.
Technical device
§ 87 (1) no. 7 BetrVG
Co-determination on health and safety arrangements within the framework of statutory provisions. Where a survey serves the risk assessment under § 5 ArbSchG, this right applies broadly – on substance, not merely on form.
Health protection
On the case law
Germany’s Federal Labour Court has held that purely voluntary and fully anonymous surveys which do not constitute a risk assessment do not trigger co-determination. That exception is narrow: it requires that no inference whatsoever about individuals or small groups is possible. HSflow is built so that the usual point of contention does not arise in the first place: no personal data is collected, and there is no analysis from which the conduct or performance of individual employees could be derived. The GDPR neutrality of the architecture has been legally reviewed.
Built to be uncontentious
Checklist
Six questions
you should put
to every vendor.
Not only to us. These questions determine whether a measurement instrument is safe for employees.
01
Are identities stored?
If so, a key exists – and what exists can be demanded. Ask about the technical implementation, not about the assurance.
HSflow: no identities
02
From what group size is analysis performed?
Without a minimum group size, a department of four effectively allows inference about individuals. Ask for that threshold in writing.
HSflow: defined minimums
03
Who sees the raw data?
The vendor itself is part of the answer. Clarify processing arrangements, storage location and deletion periods.
Contractually governed
04
Does a psychological diagnosis result?
Instruments that assess the personality or mental health of individuals are highly problematic under employment law. Measuring the strain of a work system is something else entirely.
HSflow: no psychodiagnostics
05
Is participation genuinely voluntary?
And is non-participation detectable? If managers can see who did not take part, voluntariness is effectively removed.
No participation lists
06
What happens with the results?
A measurement without an agreed consequence creates frustration. Record in the works agreement what happens above which value.
Agree an obligation to act
The overlooked cost
The loss is not
the failed project.
It is the time in between.
Presentation, pilot, objections, end. Then the next vendor. In many organisations that is the cycle – three times a year. What happens in the meantime appears in no report.
Months without data
While terms are being negotiated, it would already be possible to see where strain is building. Absences that announce themselves. Patterns that become visible before someone drops out or resigns.
Negotiation period
When speed matters
Restructuring, turnaround, post-acquisition integration: precisely then strain rises most sharply, and precisely then there is no time for months of negotiation. Under § 1 StaRUG a crisis early-warning system is required in any case – psychosocial risk is an early indicator that can be measured before it surfaces as turnover and absence.
§ 1 StaRUG
Costs with no ledger line
Germany’s Competence Centre for Employee Retention puts a single turnover case at a minimum of 43,069 euros. Those amounts appear in no balance sheet position – they accrue quietly.
from €43,069 per case
Two legitimate interests
Someone brought in from outside is there to change something – for them, speed counts. The works council protects the workforce – that is its duty. Both are right. It still does not have to fail.
Not a contradiction
Works agreement
What belongs in it –
and what is ruled out.
We provide the documentation you need for the negotiation: process description, data flow diagram, processing agreement and the data protection assessment.
Should be governed
Purpose limitation · minimum group size for every analysis · circle of recipients · retention and deletion · procedure where values are conspicuous · involvement of the works council in the analysis · repetition interval
Contents
Should be excluded
Analysis below the minimum group size · linkage with personnel data, absence records or performance data · disclosure to line managers at team level · use in employment law proceedings · subsequent change of purpose
Exclusions
The other perspective
Without figures,
strain stays opinion.
Works councils know the cases: the department where someone is always off sick. The shift nobody wants to transfer into. What is missing is the evidence – and without it, it remains a series of individual conversations.
An anonymous, methodically documented measurement shifts the burden of proof. It makes visible what employees have long known, in a form that holds up in a discussion with the board.
The data argues for acting
In the first half of 2026, mental health conditions were for the first time the leading cause of absence in Germany – 184 days lost per 100 insured employees, up nine percent year on year. Average case duration is around 40 days.
DAK analysis H1 2026
Conversation
Ask us
before you agree.
We are glad to speak directly with works councils – including without the employer side present. Documentation for your examination is provided in advance.