Receiving Feedback and Building a Team Where It Flows

The first two lessons in this course covered what feedback does and how to give it well. Both assume someone is willing to say the thing. In most teams the most valuable feedback flows upward or sideways: a mistake noticed by a peer, a plan a junior person doubts, a process everyone quietly works around. That feedback is the easiest to withhold.

Amy Edmondson named the condition that decides whether it gets said. In a 1999 study in Administrative Science Quarterly of 51 work teams in a manufacturing company, she defined team psychological safety as

Definition: "a shared belief held by members of a team that the team is safe for interpersonal risk taking."

She found psychological safety was associated with learning behaviour, such as asking for feedback, discussing errors and seeking help, and that learning behaviour in turn mediated the link between psychological safety and team performance. The facilitation course on AnyLearn applies the same research to retrospectives; here the focus is the feedback itself.

The idea came out of a result Edmondson did not expect. In an earlier study of patient care teams in two hospitals, published in 1996 in the Journal of Applied Behavioral Science, she was examining drug administration errors, expecting that teams with better leadership and relationships would make fewer mistakes.

The lesson for feedback is uncomfortable. A team with few complaints, no raised risks and smooth one-to-ones may be performing well, or may be a team where nobody feels safe saying otherwise. The two look identical from the manager's chair.

Gotcha: An absence of critical feedback is not evidence that there is nothing to criticise. Treat silence as missing data, not good news.

Google's internal research team reached a similar conclusion in its internal study of team effectiveness, known as Project Aristotle and published as a guide on its re:Work site. It examined 180 teams: 115 project teams in engineering and 65 sales pods, and identified five dynamics that set effective teams apart, listed in order of importance:

Google used Edmondson's definition of psychological safety. The study is a single company's internal analysis rather than peer-reviewed research, so it is best read as corroboration: a large employer, looking at its own teams, found the same factor at the top of the list.

Safety explains whether feedback is given. Whether it is heard is a separate problem, and the lesson one research predicts it: feedback that lands on the self triggers self-protection instead of learning.

Douglas Stone and Sheila Heen of the Harvard Negotiation Project give a practical map of those reactions in their book Thanks for the Feedback (2014). It is a practitioner framework rather than a tested model, but it is a useful diagnostic:

Notice that each fix pulls attention back from the self to the specific content, the same move the giver is meant to make.

Career advice often says to seek feedback proactively. The evidence on that is weaker than the advice suggests. Frederik Anseel and colleagues' meta-analysis of feedback-seeking behaviour (Journal of Management, 2015) found that its correlation with job performance was small, $\rho = 0.07$ across 11 samples, with a credibility interval that included zero.

That does not make asking pointless. It means the asking is not what helps; what you do with the answer is. Generic requests also produce generic answers.

Strong requests are specific, name the standard you are working toward, and make it cheap to answer with one concrete point. They ask the giver for task-level feedback directly.

In practice: After receiving feedback, say what you will change and when. Then tell the person when you have done it. That turns a single exchange into a loop worth repeating.

Formal performance reviews are the most institutional form of feedback, and their core instrument, the rating, has a measurement problem. Steven Scullen, Michael Mount and Maynard Goff (Journal of Applied Psychology, 2000) analysed two large datasets of managers, each rated on three dimensions by seven raters: two bosses, two peers, two subordinates and themselves. They separated how much of the rating variance came from the person being rated and how much from the rater.

Idiosyncratic rater effects accounted for 62% and 53% of the variance, more than twice the combined effect of the ratee's general and dimension-specific performance, 21% and 25%. A rating tells you as much about who wrote it as about who received it.

The rater effect and the feedback research together suggest several design choices for formal reviews.

Gotcha: Removing ratings does not remove judgement. Decisions about pay and promotion still happen; without a visible rating they can simply become harder to audit. Whatever the format, make the evidence behind each decision explicit.

A single conversation is one pass. Culture is whether the loop keeps turning.

The loop in the diagram breaks at two points: when feedback is withheld, and when it is defended against. A manager has the most influence over both, because their reactions are the ones the team watches most closely.

1. Go first. Ask for feedback on your own work, specifically, in front of the team. "What would have made yesterday's planning meeting shorter?"
2. Reward the messenger visibly. When someone raises a problem, thank them for raising it before discussing the problem. Edmondson's hospital teams show what happens when reporting is punished: problems disappear from view, not from reality.
3. Receive the way you want others to. Ask what they saw, separate the message from how it was delivered, and name a next step. Your behaviour sets the norm more than any policy.
4. Close loops in public. "Last month Priya pointed out the handover gap; here is what we changed." This is the acknowledgement step, and it is what makes the next person speak up.
5. Watch for silence. Track whether risks, errors and disagreements are being raised at all, not only how they are handled.

Key idea: People decide whether feedback is safe from how the last piece was received, not from what the manager says about openness.