Structure performance reviews — fair evaluation for everyone

Classic performance reviews are based on the view of a single supervisor. This blinds essential perspectives — colleagues, direct reports, internal clients. A 360-degree feedback expands the assessment with these views and delivers a significantly more complete picture. In the form this means: per person several assessment forms from different roles, with partly the same, partly role-specific questions.

Keep the number of feedback givers per person at four to eight. Fewer does not deliver a robust picture, more overwhelms the evaluation. Mix the roles deliberately: supervisors, peers, direct reports and internal customers. Per role, 8 to 12 questions are enough — for example on communication, technical competence, results orientation and collaboration. A 5-point scale with clear behavioral anchors delivers comparable data. Add an optional free text field per area for concrete examples — this qualitative depth makes the difference between usable feedback and a pure numbers report.

A central design decision is whether feedback is given anonymously or with full names. Both models have their merit. Anonymous feedback increases honesty, especially with critical statements toward supervisors. Named feedback allows queries and deepening. Companies often combine both: peer and direct report feedback anonymous, self-assessment and supervisor evaluation named.

Communicate the model crystal clearly before the start. A status line at the beginning of the form — "Your answers are evaluated anonymously, the recipient cannot recognize who gave which rating" — creates trust. Pay attention to structural anonymity: if an employee has only one direct report, their feedback is not anonymous no matter what you promise. Set a minimum number of feedback givers per role (typically three), and suppress the evaluation if this threshold is not reached. This is honest and protects the raters.

Performance reviews process particularly sensitive employee data. The legal basis is usually Art. 88 GDPR in conjunction with § 26 BDSG — processing in the employment context, possibly supplemented by works agreements. Bring the works council on board early if the company has one; without co-determination, digital evaluation systems are legally vulnerable.

Minimize the collected fields. It does not belong in every review whether someone has children or which illnesses exist. Store the data only as long as it is relevant for personnel decisions, and delete old reviews after three to five years — this time horizon is enough for development trajectories without ancient evaluations overshadowing careers. Ensure that only authorized people have access: HR, the responsible supervisor and the person themselves. Logging all accesses is mandatory. Also give employees the right to request information about the stored evaluations — that is a GDPR matter of course that many HR teams underestimate.

The real effect of a performance review does not unfold through the capture but through the evaluation. Aggregate the ratings per competence area and visualize them as a radar or bar chart — a picture says more than a table full of numbers. Compare self-assessment and external view; the most exciting insights usually lie in the discrepancies.

The calculation engine can take over the aggregation automatically once all feedback has been received. Automatically trigger an evaluation PDF to the affected person and the supervisor; this saves manual preparation. Build concrete action items into the evaluation: top 3 strengths, top 3 development areas, three suggested next steps. This structure turns a review into a development conversation and prevents the data from disappearing in a drawer. Follow up the review after 6 months with a short check-in — what was implemented, where is support needed. This follow-up is in the end the most effective lever of the entire measure.