Circles and Soup (Circle of Influence)

When to use

When retros have become venting sessions — the same complaints about the org, the roadmap, the platform team, sprint after sprint, with nothing to show for it. Run it once as a reset, not on a cadence: the taxonomy sticks, and 'that's soup' becomes team shorthand afterwards. Skip it for a team that already owns its problems — sorting cards you're already acting on is ceremony. And skip it when the complaints are about the team itself; that's a Sailboat or Start/Stop/Continue meeting.

How it runs

1. Draw the three rings

   Inner ring: the team controls it. Middle ring: the team can influence it. Outer ring: the soup — the system the team swims in. Budget cycles, org policy, the market. Two minutes; the labels matter more than the drawing.

2. Write the complaints

   Silent writing, ten minutes. Bring the recurring ones — the cards that have opened the last three retros. If venting is why you booked this meeting, this is where the venting goes: once, in writing.

3. Sort into rings

   In pairs, place each card. The test per ring: control means the team could fix it without asking anyone; influence means there's a specific person who could be asked or persuaded; soup means no realistic path. Contested between two rings? Place it one ring in and move on — an ask that gets refused moves the card out again, cheaply.

4. Write a response per ring

   This is the meeting. Control cards get a direct action. Influence cards get a specific ask — who, for what, by when. Soup cards get a change in the team's own response: buffer against it, adapt to it, or consciously let it go.

5. Name owners and retire the soup

   Actions and asks each get a name and a date. Soup cards get an explicit team decision, said out loud — and then they stop appearing in retros. That deletion is half the value of the meeting.

Why it works

The complaint list doesn't change — who's expected to act on each complaint does. Venting persists because soup-level complaints get written up as team action items, fail, and quietly prove the retro useless. Routing each complaint to a response that can actually succeed breaks that loop. The middle ring does the heaviest lifting: most 'nothing we can do' cards turn out to be 'nobody has asked yet' cards, and turning a complaint into a named ask is the format's best trick. And naming the soup out loud is a release valve — the team stops pretending it will fix the budget cycle, and stops paying ten minutes per retro to re-litigate it.

Variations

• Sort last retro's untouched action items instead of fresh cards. Where they land explains why they're untouched — most dead action items were soup wearing an owner's name.
• Futurespective version: before a project, sort the risks by ring. Soup risks get contingency plans, not ownership arguments. Pairs well with a Pre-mortem.
• Don't collapse it to two rings (control / no control). The middle ring is where the leverage lives; drop it and you lose the asks.
• Re-run after a reorg. Ring membership follows reporting lines — cards migrate when the org chart does.

Facilitator notes

Two failure postures, mirror images. A fatalist room pushes everything to the soup — learned helplessness dressed as realism. The counter is one question, per card: 'who exactly would we ask?' If a name exists, it's influence. A boosterish lead pulls everything into control — which is usually how the team ended up venting in the first place. Hold the line both ways: the rings only help if placement is honest. And cap placement debates at thirty seconds; the response step exposes miscategorised cards anyway.

Pitfalls

• Ending at the sorted board. Sorted complaints are still complaints — the response per ring is the deliverable, and it's the half teams skip when time runs short.
• Letting the soup become the biggest ring. If more than half the cards land there, the team has stopped asking, not run out of options.
• Treating a soup response as defeat. 'Stop spending retro time on it' is a real outcome — it buys back minutes from every future retro.
• Litigating ring placement card by card. Thirty seconds, place it one ring in, move on. The metaphor is a sorting tool, not the meeting.

Remote tips

Concentric circles read badly on a shared whiteboard once cards pile up — three columns labelled Control, Influence, Soup do the same job and survive a 13-inch screen. Seed the board before the call: the recurring complaints are already sitting in your last three retro boards, and starting from them tells the room this meeting is about routing, not re-venting.

Example outputs

• Control: PRs sit two days because nobody's assigned. Response: rotating reviewer named at standup, 24-hour SLA.
• Influence: the platform team's deploy freeze windows. Ask: thirty minutes with their lead this week to get the exception criteria in writing — Priya owns the ask.
• Soup: the annual budget cycle killed the contractor renewal. Response: plan next quarter at current headcount; stop assuming the backfill.
• Soup: the return-to-office policy. Response: let it go — it has opened the last three retros and nobody in this room can move it.

FAQ

Circles and Soup or Sailboat?

Sailboat surfaces friction the team hasn't said out loud. Circles and Soup routes friction the team won't stop saying out loud. If every retro already produces the same complaints, you don't need more surfacing — you need routing. Run this once as a reset, then go back to your cadence format.

Isn't putting something in the soup just giving up?

It's giving up the pretence, which is different. A soup card still gets a response — buffer, adapt, or a deliberate decision to stop spending meeting time on it. Filing the same unownable complaint every sprint is the actual giving up. The gate matters, though: a card only earns the soup if 'who exactly would we ask?' has no answer.

Where does the format come from?

Diana Larsen, 2010, on the Partnerships & Possibilities blog — building on Stephen Covey's circles of concern and influence (via Jim Bullock's control/influence/concern variant), with 'the soup' borrowed from David Schmalz. Larsen's addition is the one that matters here: Covey's version is personal effectiveness; hers sorts a team's complaints and attaches a team response to each ring.

Related activities

Recommended use cases