You ran the engagement survey. The company-wide numbers are in, and now every manager is asking the same thing: “How did my team do?”
It’s the right question. Company averages hide the teams that are struggling and the ones doing something worth copying. A manager can act on their own team’s results. They can’t do much with a company average.
But splitting results by team is where anonymity usually breaks. If a team has four people and three of them answered, “anonymous” doesn’t mean much. Here is how to give each leader their team’s results without putting anyone at risk.
Problem 1: small teams
The obvious rule is a minimum group size: don’t show a team’s results unless enough people answered. Five is the usual floor for scores. Below that, an average is close to a single person’s answer.
Comments need a higher bar than scores. A score of 3 out of 5 could be anyone. A comment that mentions “the Tuesday stand-up” or “since I moved from Leeds” points to one person. A sensible rule is five responses for scores and ten for comments. A team between the two sees its scores but not its comments.
What to do: set both minimums before the survey goes out, and tell people what they are. “Your team’s results are only shown if at least five people answer” is a reason to answer, not a reason to hold back.
Problem 2: subtraction
This is the one people miss. Say the Sales department has three teams: North (12 people), South (9) and West (3). You hide West because it’s under five. Good.
Then you show Sales as a whole, plus North and South. Anyone with a calculator can take the Sales total, subtract North and South, and get West’s results anyway.
The fix is to hide one more group. When one team under a department is hidden, also hide the next-smallest team beside it, so the subtraction no longer works. The hidden teams still count in the department and company totals. Nobody’s answers are thrown away, they’re just never shown on their own.
The same thing happens one level up: if a department is hidden and has nobody in it directly, its teams can be added up to rebuild it, so enough of them need hiding that the sum no longer works.
What to do: before you share anything, check every hidden team. Could it be worked out from what you are showing? If yes, hide one more group.
Problem 3: managers who see more than their team
A manager should see their own team’s report. They shouldn’t be able to filter the full results down to one team member’s group. That means two kinds of access: the people running the survey, who see everything above the minimums, and team leaders, who see one report each.
The cleanest way is a private link per leader, one report, no login, that expires. Nothing for them to filter, and nothing to forward that shows more than their team.
What a useful team report contains
Raw numbers don’t help a manager much. A good team report answers three questions.
- How are we doing compared with everyone else? Each score beside the company and beside the department above it, with the difference.
- What should we keep doing, and what should we fix? The team’s two or three highest and lowest questions. Not ten. A list of ten is a list nobody acts on.
- What did people actually say? The team’s comments, if the team is big enough to show them.
Group questions into a few themes, such as “Engagement”, “Wellbeing” and “Growth”, and give each a score. A manager can talk about three themes in a team meeting. They can’t talk about forty questions.
Two scoring details catch people out:
- Negatively worded questions (“I often feel overloaded”) need reversing, so a high answer counts as a low score. Otherwise they drag a theme in the wrong direction.
- “Not sure” and “N/A” answers shouldn’t count as a point on the scale.
Read the comments before the leaders do
Before reports go out, someone should read each team’s comments. Not to remove criticism, but to catch the one comment that names a colleague or describes a situation only one person could have been in. Hide that comment from the team reports. It still counts in the overall results.
No minimum is a guarantee
Be honest with yourself and with your people about this. Minimums and hiding make it much harder to identify someone. They don’t make it impossible. Two views that differ by one person can still give that person away, and someone who knows the team well can sometimes guess. That’s why the limits matter, and why you should resist “just this once” requests to see a smaller group.
How SurveyRock does it
We built team reports to work this way. You load your org structure from one CSV file: teams, who leads them, and who belongs to each. Responses are tagged with their team as they come in, including on anonymous surveys, where a response records which team it came from and still stores no contact, IP address or device.
Team results follow two minimums, five for scores and ten for comments by default, and you can raise either. When a small team is hidden, the next-smallest team beside it is hidden too, so the small one can’t be worked out by subtraction. By default, nobody can read results below those minimums. The one exception is Staff analysis access, a setting for owners, admins and editors who need response-level data for small groups: it’s off unless you turn it on, and every view below a minimum is logged. Leader reports stay bound by the minimums either way. You set up themes (“dimensions”) with mean or % favourable scoring and reversed questions, review each team’s comments before release, then send each leader a private link to their report. Reports come in the survey’s language, and leaders can download them as a PDF. Run the survey again next year and each report can show how the team has changed since the earlier wave.
Team reports are available on the Growth plan and above. See how team reports work.