Anonymity
The anonymity rules protect individual responses, ensuring that no one can see what a single person has answered. Brilliant guarantees that all responses are treated confidentially and remain anonymous.
Basic Rules
- From 3 responses: Displays index value and average per question, incl. result insights in Brilliant Insights.
- From 5 responses: Shows the distribution of positive, neutral, and negative responses.
- From 10 responses: Enables displaying responses on a five-point scale.
Specific Rules
- eNPS: Score is displayed from 3 responses, response distribution from 5 responses, and a detailed graph from 20 responses in index view.
- Open-ended comments: Only shown for group consisting of at least 20 individuals. Not reported at the team level.
- Yes/No questions: At least 5 responses are required to see the response distribution.
- Engagement matrix: From 3 responses, a marker is displayed; from 10 responses, the percentage distribution is shown.
- Equal treatment questions: Displayed for group consisting of at least 30 individuals, with certain sensitive questions only visible to HR.
Text Analysis
Text analysis is available for groups with at least 20 respondents and is not reported at the team level. The feature is permission-based and, according to Brilliant’s standard settings, is only visible to users with an HR role in the platform. Customer-specific variations may apply.
Background Rules
Background questions (in the survey) and/or background variables (from the organizational structure) are used to analyze differences between groups at an overarching level.
Results are only displayed if the group consists of at least 50 employees and there are at least 5 responses per response option or variable. Both criteria must be met for the results to be shown.
"I am the only man in my team—won't my manager see my response?!"
For example, in a team of 7 employees where 1 is a man and the remaining 6 are women, no results would be shown for either men or women. This is because there are fewer than 5 responses from men and fewer than 50 employees in the group.
Combination Rules
To prevent responses from small groups or individuals from being identified by cross-referencing different data points, combination rules are applied.
Combination rules typically affect summarised results. For example, a summarised result might consist of multiple teams of different sizes. If some teams have too few responses to receive individual results (fewer than 3 responses), this impacts the summarised results.
A few Common Scenarios:
Summary - 14 responses
- Team A - 4 responses ✅
- Team B - 3 responses ✅
- Team C - 7 responses ✅
- All teams meet the minimum response requirement (3 responses), and full reporting is available for the summary.
Summary - 18 responses (2 responses excluded)
- Team A - 9 responses ✅
- Team B - 5 responses ✅
- Team C - 4 responses ✅
- Team D - 2 responses 🚫
- Team D is excluded from the summary to protect the individuals in the small team. The summary is based only on Teams A, B, and C (18 responses).
Summary - 6 responses
- Team A - 2 responses 🚫
- Team B - 2 responses 🚫
- Team C - 2 responses 🚫
- None of the individual teams have enough responses to receive team-specific results, eliminating any anonymity risk. Instead, full reporting is provided for their combined results in the summary.
Example of a Combination Rule Applied to Background Data Analysis
"I see results for a background question even though some response options have fewer than 5 responses!"
In this example, employees are spread across five different countries. The response distribution for the question about which country they work in is:
- Norway: 4 responses 🚫
- Sweden: 18 responses ✅
- Denmark: 26 responses ✅
- Finland: 3 responses 🚫
- Iceland: 17 responses ✅
Results for Norway and Finland cannot be shown individually because they each have fewer than 5 responses.
However, in situations where two or more smaller groups (e.g., Norway and Finland) together meet the anonymity threshold (in this case, 5 responses for background information), results for the remaining groups (Sweden, Denmark, and Iceland) can be displayed. It is not possible to determine whether a response originates from Norway or Finland.Prioritization
"Why does this happen for the group I'm analyzing? Why are certain responses excluded?"
The answer is that it's difficult to say 🚩 — There are often as many potential combinations as there are groups in an organisation. If an anonymity conflict arises, the platform prioritises displaying results in the following order:
- Team — Always displayed if the group has at least 3 responses.
- Summary — Displayed as much as possible, but applies combination rules.
- Background information — Not displayed if this would cause issues for the team or summary levels.
🚩 Unfortunately, Brilliant cannot deep dive into individual cases to provide detailed explanations as to why specific results cannot be displayed or why certain responses have been excluded due to anonymity, but refers to the above description.