Weekly retro template.

The Retro That Produces Decisions, Not Feelings

Most retros are venting sessions. Forty-five minutes of complaints, a few comments about what went well, the team leaves feeling slightly heard and nothing structural changes. The retro is performative. The cadence becomes optional. Within a quarter, the retro has been retired or has continued in a degraded form that nobody takes seriously.

The Weekly Retro Template is built on a hard rule learned from watching dozens of retros run badly: most retros produce sentiment, not action. The template fixes that by replacing open-ended prompts with structured ones that force the conversation toward decisions.

The template is opinionated. Three prompts kept. Four prompts banned. No exceptions. The opinionation is the point. Retros that try to be everything to everyone become venting sessions; retros that have a structural discipline produce numbered decision lists that survive the week.

This essay is about why three prompts is the load-bearing number, what the banned prompts produce that the kept ones don't, and how the structural-escalation channel converts the retro from a complaints log into a feeder of structural change. The kit guide covers the mechanics; this is the operator narrative.

The Three Prompts That Were Kept

The retro uses three prompts. Same three, every week. The structure is the value.

Prompt 1. What is working that we should keep doing? Ten minutes. The team names specific practices, decisions, or patterns from the week that produced good outcomes. The discipline is to name the specific thing — "the new working session agenda" rather than "we're communicating better." The output is a short list of validated practices. The list compounds; over a quarter, the team builds a catalog of what works that becomes part of the operating system.

Prompt 2. What got in our way this week? Twenty minutes. The team surfaces specific obstacles — process friction, missing information, decisions that didn't get made, dependencies that didn't resolve. The discipline is to name the obstacle precisely enough that someone could act on it. For each obstacle, the team commits to one of three responses: accept it as a known constraint, propose a tactical change for the next week, or flag it as a candidate for structural change requiring leadership input.

Prompt 3. What decision do we need this week? Fifteen minutes. The team names the decisions that need to be made in the coming week, with the owner and the timing for each. The retro becomes the decision queue for the working session.

That is the entire retro. Three prompts. 45 minutes. Run by a rotating facilitator from the team, not by a manager.

The Four Prompts That Were Banned

Four prompts are explicitly banned from this template because they reliably produce sentiment without action.

"How do we feel about the week?" Produces vague emotional check-ins that don't connect to specific operating changes. Feelings matter — they are signal — but they are not what the retro is for. The retro produces decisions; one-on-ones and skip-levels produce the space for feelings.

"What should we stop doing?" Sounds productive, almost always produces a list of pet peeves with no commitment to actually stop. The right framing is in Prompt 2 — "what got in our way" — which surfaces the same content but routes it toward a concrete next-step commitment.

"What did we learn?" Produces post-hoc rationalizations that feel insightful and rarely change behavior. The team agrees that lessons were learned, then proceeds to repeat the patterns. The retro is for forward decisions, not for retrospective interpretation.

"What are our wins?" Produces a celebration ritual that, while pleasant, displaces operating discussion. Celebrations belong in the working session opener or in a separate cadence. The retro's 45 minutes are too precious to spend on celebration.

The banned prompts are not banned because they are bad. They are banned because they crowd out the work the retro is uniquely good at. Other cadences exist for the work those prompts produce.

The Structural Escalation Channel

Prompt 2 produces three classifications: accept, tactical change, structural escalation. The third one — structural escalation — needs a defined channel.

When an obstacle is flagged as structural, it means the team has identified a problem they cannot fix within their own authority. It requires leadership input, cross-functional change, or organizational design work. The structural-escalation list goes to the leadership team's working session, where it's reviewed in a standing 10-minute slot.

The discipline matters because structural problems that have nowhere to go become culture problems. The team learns that some obstacles cannot be fixed, so they stop flagging them. Over a quarter, the retro loses its diagnostic power.

A defined escalation channel keeps the retro honest. Leadership commits to either addressing the structural item, declining it with reasoning, or scheduling it for a future quarter. Any of the three is acceptable; silence is not.

Operations that install the escalation channel see retro yield (the fraction of surfaced issues producing structural change) rise from 30-40% in the first quarter to 60%+ by the second. The yield rise is what converts the retro from a complaints log into a feeder of structural improvement.

Why a Rotating Facilitator, Not a Manager

The retro is run by a rotating facilitator from the team, not by a manager. This matters for two reasons.

The facilitator role builds operating muscle. Facilitating a 45-minute structured discussion is a skill. Rotating the role gives every team member practice. By the end of a quarter, every team member has run the retro at least once. The cumulative effect on the team's meeting-facilitation capability is substantial.

Manager-led retros produce filtered content. Even the best managers create a power dynamic that filters what gets surfaced. Team-led retros, with the manager present as a participant, produce honest content because the team owns the conversation.

The manager's job during the retro is to participate as a peer in Prompts 1 and 2, and to take notes on the structural-escalation list for the working session. The manager does not facilitate; the manager does not control the agenda; the manager does not adjudicate disagreements. The discipline reinforces that the retro is the team's process, not the manager's.

What Compounding Retros Look Like

Operations that run the retro disciplined for a quarter see three patterns emerge.

The "what is working" list builds. Over 12 weeks, the team has accumulated 30-40 validated practices. The list becomes a real catalog of what the operation has learned works in their specific environment. The catalog gets referenced when onboarding new team members or when designing new processes. The compound knowledge is substantial.

The same items stop appearing in Prompt 2. Obstacles that get tactically addressed don't recur. Obstacles that get structurally escalated and addressed don't recur. The remaining recurring items are either accepted constraints (which the team has stopped trying to fix) or genuinely hard problems (which need leadership attention). The signal-to-noise improves.

The structural escalation list shapes the leadership agenda. Items the team has been quietly working around surface in the working session as candidates for structural change. The leadership team gets visibility into operating friction it would not otherwise see. The cumulative effect on operating health is measurable.

Operations that run the retro inconsistently produce none of these compounds. The retro happens or doesn't happen depending on the week; the format varies; the structural channel doesn't exist or doesn't get honored. The retro becomes a habit nobody believes in.

What to Do This Week

The install is one retro. 45 minutes. Same team that will run the cadence going forward.

Identify the facilitator. Pick someone from the team, not a manager. Could be different person each week (recommended) or fixed for the first month and then rotated.

Open the retro with the three prompts visible. Write them on the wall or share screen. The structure is visible the whole time.

Run the three prompts in order with the timeboxes. 10 minutes, 20 minutes, 15 minutes. Total 45 minutes. End on time even if the conversation isn't finished — the timeboxing is part of the discipline.

Capture the three lists. Kept practices, classified obstacles, decision queue. Post them in the team's working channel within an hour of the retro ending.

Install the escalation channel. Define where the structural-escalation list goes (typically the next leadership working session). Commit to leadership responding to every item — action, context, or acknowledgment. Silence is not an option.

The kit guide at /playbooks/weekly-retro-template covers the structural detail. This essay is the operator narrative for why the three-prompt discipline is the load-bearing element. If Accountability or Operating Cadence is in your top three Ops Check risks, the retro is one of the lowest-cost installs that produces structural improvement.

Three prompts. 45 minutes. Team-led. Structural-escalation channel honored. The continuous-improvement loop that converts weekly operating noise into structural operating gain.