Mad Tea: Reading the Room in Twelve Minutes Flat

The plan for Q4 was to rewrite the Recap CRM mobile app. Not refactor — rewrite, from a three-year-old hybrid stack to native, with the deal pipeline and activity feed as the first surfaces. I’d written the doc, the doc had numbers in it, and in the planning meeting fourteen people agreed with it in the way people agree with a plan that has numbers in it.

I didn’t believe the agreement — not because I suspected anyone of dishonesty, but because “does anyone have concerns” is the least effective question in product management, and a room that has just read a well-argued doc will not produce its real reservations in response to it.

So before the detailed planning session we ran twelve minutes of Mad Tea.

What it is

Mad Tea is the fastest structure in the Liberating Structures set and the one that looks least like work. Everyone stands in two concentric circles, inner facing outer, so each person has a partner. The facilitator reads a sentence stem aloud. Both partners complete it — fifteen to thirty seconds each, no discussion, no follow-ups. Then the circles rotate a position or two and the next stem is read.

Nine or ten stems, twelve to fifteen minutes, over before anyone settles into a posture. The pace is the point: at thirty seconds you say the first true thing rather than the third careful thing.

The stems carry the design — short, open, provocative enough to require a genuine answer. There is no output artifact, deliberately: capturing turns a fast honest exchange into an on-the-record statement, and the honesty is the product.

Running it before the rewrite kickoff

Fourteen people, two circles of seven, and I read the stems rather than pairing in — with an even headcount a facilitator who joins would break the circles. My stems, written the night before:

1. What excites me about a native rewrite is…
2. What worries me most is…
3. Something nobody has said out loud yet is…
4. In six months I’d be disappointed if…
5. The part I don’t understand is…
6. If we had to cut scope, the first thing I’d cut is…
7. The person or team I’m most worried about in this plan is…
8. What would make this obviously worth it is…
9. What I need in order to do my part well is…

Stem three broke it open. Standing between the circles, I heard three versions of the same thing pass by in thirty seconds: we had never established whether mobile usage was actually growing, and the rewrite doc’s central metric was session count, flat for a year.

Stem six was the other one. “The first thing I’d cut” produced, from at least five people I could hear, the activity feed — one of the two scheduled launch surfaces. Nobody had objected to it in the planning meeting. In thirty seconds with one partner, most of the room independently identified it as the least valuable half of the plan.

And stem seven, which I’d nearly cut for being too pointed, surfaced something I’d been oblivious to: several people named the two-person design team, asked to produce a full native design language while continuing to support web. A staffing problem hiding inside an architecture decision.

Twelve minutes, no notes. Then we went into the planning session and I opened by naming the three things I’d heard standing in that circle: the session-count metric, the activity feed, and design capacity.

That’s the move. The structure surfaces the material; naming it yourself gives everyone permission to discuss what they’ve all just said privately to two or three people. The plan we left with launched pipeline only, added a two-week instrumentation spike, and had a designer contract attached.

Why speed produces honesty

Thirty seconds is too short to construct a position. It’s long enough to say what you already think and not long enough to package it — and packaging is where candor is lost. Give someone three minutes to answer “what worries you” and you get an answer that has already accounted for how it will land. Give them thirty seconds and you get the worry.

The rotation matters as much as the clock. Each stem is answered to a different partner, so nobody builds a case with one person and nobody feels committed to a position. It stays low-stakes, and low-stakes is where people say the inconvenient thing.

The volume makes it legible. Fourteen people times nine stems is a hundred and twenty-six spoken answers in twelve minutes. You cannot record that and you don’t need to: what you notice is repetition.

When to skip it

Skip it when you need a decision or a record. Mad Tea produces neither, by design — use it to open something, never to close one.

Skip it with a group in genuine distress. The pace that makes it candid also makes it glancing, and asking someone to summarize a painful situation in thirty seconds and then rotate is dismissive. Conversation Cafe fits that shape.

Skip it if the stems are bland — “what I hope for this quarter is…” gets you nothing. And it needs a real room, or a video tool that reshuffles pairs fast: if your reshuffle takes ninety seconds the tempo collapses, and you’ve built a slow structure with a fast one’s shallow prompts.