Question 1

Will my team actually say anything useful with their manager in the room?

Accepted Answer

Only if they can choose not to be named, so make that the default for the topics that need it. Anonymity is set on the meeting, not on the account, so a post-mortem can run fully anonymous while next week's planning retro runs with names. The setting only ever moves toward more anonymity once a session starts, and people can mark an individual idea or comment anonymous even in a named meeting. Run a psychological safety check first if you are not sure where the team stands. It gives you a rated baseline rather than a guess, and it tells you which topics still need cover.

Question 2

How do I stop the same incident happening a third time?

Accepted Answer

Run a post-mortem the team can be honest in, and make the output an action with an owner and a date rather than a document. Separate the timeline from the causes so the discussion lands on what let the failure through, not on who typed the command. Because open actions are put back on screen at the start of the next meeting, a fix that quietly did not happen becomes visible in weeks rather than at the next outage.

Question 3

How do I make technical debt visible to people outside engineering?

Accepted Answer

Score it instead of arguing it. A technical debt retrospective gets the debt out of individual heads and onto a board the whole team ranks, and an engineering excellence check rates debt visibility, debt reduction and impact awareness on the same scale every quarter. A rated trend line that has moved the wrong way for three quarters is a far easier conversation with a product lead than a list of grievances, because it is the team's own number rather than your opinion.

Question 4

Can I find the delivery bottleneck without standing over the team?

Accepted Answer

Read the aggregate rather than the individuals. An agile delivery maturity check rates work-in-progress management, cycle time predictability and release cadence, so the low score points at the stage that is actually stalling. Insights then show which themes keep coming back across meetings and how many agreed actions are closing, which tells you whether the team already knows the problem and cannot get to it. Neither view names individuals, so you get the diagnosis without the surveillance.

Question 5

How do I shorten the time it takes a new engineer to be useful?

Accepted Answer

Ask the people who just went through it, while they still remember. The engineering excellence model rates knowledge sharing, developer enablement and skill breadth, which is usually where onboarding drag really lives: undocumented setup, one person who knows the deploy, a codebase nobody has mapped. Turn the two lowest dimensions into actions with owners, then re-run the check a quarter later. The next hire inherits the fixes rather than the folklore.

Question 6

How do I keep several squads consistent without adding process?

Accepted Answer

Standardize the template, not the meeting. Publish an organization template so every squad runs the same post-mortem and the same health model, then leave each team to run it when and how it suits them. Because the questions match, the results roll up: you can compare squads on the same dimensions and see which pattern is local and which is systemic. Nobody gains a meeting, and you gain a read across all of them.

Cut repeat incidents, rework and onboarding drag

Tailored templates for teams

Cross-team collaboration

Create actionable outcomes

Data-driven insights

Amazing features designed for Engineering Managers

Organization-wide templates

Power up with AI

Team health and sentiment

Track actions and agreements

Explore word clouds

Data driven insights

Frequently asked questions