---
name: workshop-facilitation
description: Use when a PM needs to plan or run a focused, aligned product workshop — pre-work, facilitation modalities, time-boxing, decision-capture, and follow-through — rather than convening yet another open-ended discussion.
role: product-manager
intent: |
  Equips a PM to design and run a workshop that has a clear reason to exist,
  earns its synchronous time through structured pre-work, produces a captured
  decision (not a transcript), and closes with an owner and a deadline. The
  skill is opinionated: if you can make the call without a room, make the
  call without a room.
type: workflow
theme: workshops
best_for:
  - A PM who needs to produce alignment on a contested question (positioning, roadmap priority, solution direction) with 4-12 people in under half a day
  - Running a time-boxed decision-making session that surfaces divergent ideas and forces a convergent choice — without drifting into an open-ended discussion
  - Preparing pre-work that lets the synchronous session start from a shared base rather than repeating context already available async
scenarios:
  - "I need to run a 3-hour roadmap prioritisation workshop with my cross-functional team next Tuesday — what's the structure?"
  - "We've had three meetings on positioning and still no decision — how do I design a session that produces one?"
  - "The leadership team wants a half-day strategy workshop; I've got 10 days to prepare — what pre-work do I need and how do I run it?"
estimated_time: "2-5 days total (pre-work design and distribution: 1-3 days; in-room facilitation: 2-4 hours; synthesis and follow-through: 2-4 hours same day)"
---

# Workshop facilitation

> v0 draft. Primary frameworks: Sam Kaner's divergent/convergent diamond and facilitation modalities (*Facilitator's Guide to Participatory Decision-Making*, 3rd ed., 2014, Jossey-Bass); Priya Parker's purpose-first gathering design (*The Art of Gathering*, 2018, Riverhead); Jake Knapp, John Zeratsky, and Braden Kowitz's time-boxed decision structure (*Sprint*, 2016, Simon & Schuster). Editorial review recommended before production use.

## Purpose

Turn a contested, ambiguous, or multi-stakeholder question into a captured decision with an owner and a deadline — by running a purposefully designed workshop rather than an open-ended meeting.

The workshop is a synchronous investment. It is expensive: it costs every participant's presence simultaneously. The PM's obligation is to justify that cost. Justify it through specificity of purpose (Parker: the best gatherings know why they cannot happen any other way), through pre-work that collapses context-sharing into async, and through facilitation that moves a group through divergent exploration into convergent choice.

This skill is not about running better meetings. It is about deciding whether to run a workshop at all, designing one that earns its time, facilitating it with the modalities that match the group's state, and capturing the decision in a form that survives the room.

## Key Concepts

### Divergent/convergent rhythm — Kaner's diamond

Kaner's diamond is the load-bearing framework: productive group decision-making passes through a divergent zone (many ideas, messy, uncomfortable) before arriving at a convergent choice. The groan zone — the messy middle where options proliferate before they narrow — is not a failure of facilitation; it is the work. Facilitators who collapse the groan zone early produce false alignment: people agree publicly, disagree privately, and re-litigate later.

The PM's job as facilitator is to protect the divergent phase long enough that options are genuinely surfaced, then drive convergence without letting the session extend indefinitely. Time-boxing each phase (Knapp et al.) is the mechanism: a hard stop on divergence forces the group to hold what they have and begin choosing. Without time-boxes, workshops expand into divergence with no convergent pressure.

### The pre-work that earns synchronous time

Parker's principle: the gathering's purpose determines its design, and its design must start before people walk in the room. For a workshop, "starting before people walk in" means sending pre-work that:

1. Frames the question the workshop will answer in one sentence.
2. Shares the context participants need so they arrive with a shared base — not a blank slate you fill the first 40 minutes restating.
3. Specifies the expected output: a decision, a ranked shortlist, a draft principle — not "a discussion."
4. Asks participants to do one concrete task before arriving (read a brief, rank options, answer two questions).

Pre-work is not optional preparation; it is the mechanism that lets the synchronous session start at the groan zone rather than at setup. A session that starts with 30 minutes of "let me catch everyone up" has failed before it started.

### Facilitation modalities — when to apply each

Kaner describes a set of modalities that address different group states. The PM should select modality from group state, not from comfort:

- **Heads-up start**: opens with the question and the output, not with introductions or agenda review. Orients the group immediately. Use at the top of every session.
- **Context-dump bypass**: when participants are unequally informed, front-load context as a shared document rather than an in-room brief. Participants who don't read it get caught up through a 5-minute table comparison, not through a 30-minute presentation.
- **Best-guess mode**: when the group is stuck in "we need more data," ask each participant to commit to a best guess now and label it as a hypothesis. Prevents the groan zone from becoming an infinite evidence-gathering loop.
- **Progress labels**: visible timekeeping on a shared display. Not a countdown timer that creates anxiety, but a visible "we're now at the convergence phase" label. Kaner: groups under-estimate how long divergence takes; progress labels give the facilitator a shared referent to call time without conflict.
- **Interruption handling**: designate a parking lot for off-topic threads. When the room pulls toward a tangent, the PM names it, parks it, and redirects — without dismissing the person who raised it. The parking lot earns its name: a parked item gets a home (the post-session action list or a separate session), not a quiet death.

### Decision-capture vs. discussion-recording

A workshop that ends without a captured decision has not ended. It has paused.

Decision-capture is not meeting minutes. It is a single document produced in the room or within 30 minutes of the room closing, containing:

- The question the workshop was convened to answer.
- The decision made — one sentence, specific enough to be falsifiable.
- What was explicitly ruled out (this is where alignment breaks down if skipped — people remember the options differently).
- The owner of the next action and the deadline.
- What would re-open the decision (Kaner: decisions made under uncertainty need a revisit condition, or they get re-litigated informally).

Discussion notes are supplementary. If decision-capture is absent, the discussion notes become the authoritative record — and they almost always contain contradictory attributions that allow re-litigation.

### Psychological safety as a facilitation precondition

Kaner: inclusive facilitation requires that participants believe their contribution will not be used against them. In PM workshops this surfaces as: junior members stay quiet while senior stakeholders speak; dissent is signalled with body language, not words; the loudest voice in the room wins.

Facilitation techniques that protect safety: round-robins that give every voice a structured slot; silent parallel ideation (Knapp et al.'s sticky-note sprints) that decouples idea quality from social status; explicit invitation for dissent before the decision is logged ("is there anything important we haven't said?").

If the group does not have baseline safety — if two participants have an unresolved conflict, or if hierarchy is so steep that reports will not contradict the most senior person in the room — the workshop will produce theatre. In that case, the right PM move is to meet 1:1 with the key parties before the session and surface the actual disagreement outside the room.

### When NOT to run a workshop

A workshop is the wrong tool when:

- The decision belongs to one person and consultation has already happened. Convening a workshop to "align" on a decision already made is participation theatre.
- The question needs more research before options can be meaningfully evaluated. A workshop without viable options to compare produces divergence with no convergent surface.
- The group is fewer than 4 people. Below this threshold, a structured 1:1 conversation or a short async thread is faster and preserves the same quality.
- The question is sensitive enough that the answer will be politically distorted by group dynamics. Surface this through 1:1 discovery first.

The test: "If I sent a well-framed async document to everyone involved, with a 48-hour deadline for responses, would I get to the same decision?" If yes, don't run the workshop.

## Application

Four phases, PM-led. Total PM time: 2-5 days including prep.

### Phase 1 — Design the workshop (1-3 days before the session)

**Step 1 — Name the decision, not the topic.** Write the question the workshop will answer in one sentence. "Should we prioritise the API redesign or the onboarding flow in Q3?" beats "Q3 roadmap discussion." The question determines the output; if the question is vague the output will be vague.

**Step 2 — Apply the not-a-workshop test.** Would a well-framed async document with a 48-hour response window produce an equivalent decision? If yes, send the document. If no — because the decision requires real-time trade-off negotiation, because the group has unresolved tension that needs surfacing, or because the options are not yet clearly enough articulated for async — proceed.

**Step 3 — Define the expected output and ruling-out criteria.** What does "done" look like when the session ends? A ranked shortlist of 3 options? A yes/no decision on one option? A set of principles the team can apply autonomously? Name what you're ruling in and what you're explicitly ruling out. The ruling-out criteria prevent the groan zone from re-surfacing options the group has already rejected.

**Step 4 — Design the pre-work brief.** Write a document (1-3 pages) that:
- Frames the decision question.
- Provides the context participants need to engage productively (not everything you know — the load-bearing facts).
- Names the expected output.
- Assigns a pre-work task: read and annotate a brief, complete a ranking exercise, answer two questions in writing.
- Specifies the time commitment: "30 minutes before Tuesday."

Send it no later than 2 days before the session. A pre-work brief sent the evening before is a pre-work brief that doesn't get read.

**Step 5 — Design the agenda as time-boxed phases.** Use Kaner's diamond as the structural template:
- **Open + heads-up start** (5-10 min): the question, the output, the rules.
- **Divergent phase** (20-40 min): ideas out, no evaluation, sticky-note sprint or round-robin.
- **Groan zone** (15-30 min): clustering, pattern-naming, challenge and surfacing of tension.
- **Convergent phase** (20-30 min): dot-voting, ranking, or structured debate toward a shortlist.
- **Decision-capture** (10-15 min): the facilitator states the decision aloud; the room confirms or corrects; the PM captures it in the shared doc.
- **Next actions + close** (5-10 min): owner, deadline, parking-lot disposition.

Total: 75-135 minutes. Sessions shorter than 75 minutes can work for simple decisions; sessions longer than 150 minutes degrade in convergence quality. If the question requires more time, break it into two sessions with explicit output from each.

**Step 6 — Prepare the decision-capture template.** Before the session, have the template open in a shared document:

```
Decision: [question]
Answer: [one sentence]
Ruled out: [options explicitly not chosen, with one-line rationale each]
Owner: [name]
Deadline: [date]
Revisit condition: [the specific signal that would re-open this decision]
Open questions (parked): [list]
```

### Phase 2 — Run the session (in-room, 75-135 minutes)

**Step 7 — Open with a heads-up start.** State the question, the output, and the schedule. Confirm the output format aloud: "By the end of this session we will have agreed on X and I will capture it in this shared document." If participants don't share a baseline from the pre-work brief, run the context-dump bypass (5-minute table comparison) before proceeding.

**Step 8 — Protect the divergent phase.** No evaluation during divergence. When participants start evaluating during ideation ("that won't work because..."), redirect: "Park the evaluation — let's get everything on the board first." The silent sticky-note sprint (Knapp et al.: each person writes options individually, then posts them) decouples ideation from social status and produces more divergent input than open discussion.

**Step 9 — Navigate the groan zone explicitly.** When divergence peaks and options are proliferating, name it: "We've got a lot of options — this is the hard part, the groan zone. We're going to start clustering and narrowing." Use affinity mapping or dot-voting to surface which options have support. The groan zone is not a pause; it is active work. The facilitator's job is to keep the group working, not to rescue them from discomfort by collapsing convergence early.

**Step 10 — Drive convergence with time-box discipline.** When the convergent time-box is reached, stop adding options. Knapp et al.'s time-boxing principle: the clock forces a decision from what's on the table. Options still in active debate at time-box expiry go to the decision-maker (usually the PM or the most senior stakeholder present) for a call — not back into discussion.

**Step 11 — Capture the decision in the room.** Before the session closes, read the decision aloud from the template. Ask: "Is this an accurate reflection of what we decided?" Pause for corrections. Log any dissent: "X raised a concern about Y — it's captured in the open questions." Dissent that is named and parked is more durable than dissent that is ignored and resurfaces.

**Step 12 — Handle the parking lot.** For each parked item, assign: file as an open question (owner + deadline) or schedule a separate session. Do not re-open the session to resolve parked items; the parking lot earns its credibility by having items actually move.

### Phase 3 — Post-workshop synthesis (within 30 minutes of session close)

**Step 13 — Distribute the decision-capture document.** Send the completed template to all participants and key stakeholders not in the room, within 30 minutes. Include a 24-hour window for corrections ("if this doesn't accurately reflect the decision, reply by [time] tomorrow"). After the window closes, the document is the authoritative record.

**Step 14 — Log open questions in the team's discovery or action system.** Each parked item becomes an owned action. Torres' OST framing: unresolved questions are opportunities, not loose ends. They should enter the team's standard opportunity-tracking process, not a workshop appendix nobody reads.

**Step 15 — Set the revisit condition in a calendar system.** A revisit condition that isn't tracked is aspirational. If the decision has a meaningful revisit condition ("if we haven't shipped the API redesign by Q3-end and conversion hasn't moved, we re-evaluate the Q4 roadmap"), put a calendar entry on it with the captured decision attached.

### Phase 4 — Follow-through (1 week post-session)

**Step 16 — Confirm action items are moving.** Check with named owners that their actions are tracked and sequenced. The workshop's durability is determined by follow-through in week one. A decision that produces no action in the first week is a decision that will be re-litigated.

**Step 17 — Note what to change for the next workshop.** Was the divergent phase too short? Did the pre-work get read? Did the time-boxes hold? One paragraph in a personal retro doc. Not a team retro — the team's workshop hygiene improves incrementally through the PM's own discipline, not through a meta-workshop about the workshop.

## Examples

### Example 1 — Positioning workshop (full worked plan)

**Context.** A B2B SaaS product team has run three roadmap reviews without agreeing on the positioning for their SMB vs. enterprise tier. The PM suspects the disagreement is actually a targeting disagreement (who is the primary customer?) dressed as a feature disagreement. Three stakeholders (Head of Sales, Head of Product, CEO) hold different mental models. Total workshop time available: half a day.

**Pre-work brief (sent 3 days before):**

> We've been building for two customers at once and it's showing in the roadmap. This 90-minute session will answer: "For the next 4 quarters, are we primarily building for SMB buyers who self-serve, or enterprise buyers who buy through sales?" This is not a features discussion — it's a targeting decision that will determine which features belong in the roadmap at all.
>
> Before arriving, please spend 20 minutes writing answers to: (1) What does our most successful customer look like today — company size, buying process, first use case? (2) What customer type should we be optimising the next 4 quarters for, and what would we stop doing if we made that call? Bring your written answers.

**Agenda (90 minutes):**

| Phase | Time | Activity |
|---|---|---|
| Heads-up start | 5 min | State the question, output, and time-box structure |
| Pre-work compare | 10 min | Each participant shares their written answers (round-robin, no debate yet) |
| Divergent | 20 min | Sticky-note sprint: "What is the primary job our most successful customers hire us for?" — silent, individual, then posted |
| Groan zone | 25 min | Affinity cluster the stickies; name the two or three poles of disagreement explicitly |
| Convergent | 20 min | Dot-vote on customer archetype; structured debate on the top-two; PM calls or puts to the room for a show of hands |
| Decision-capture | 5 min | PM reads decision aloud; room confirms or corrects |
| Parking lot + close | 5 min | Park feature debates that emerged; assign owners; set revisit condition |

**Decision-capture template (completed after session):**

```
Decision: Which customer segment do we optimise for in the next 4 quarters?
Answer: SMB self-serve buyers (≤50 employees, single-seat trial path).
Ruled out:
  - Enterprise co-development path: insufficient revenue signal in current
    pipeline to justify dedicated PM and sales capacity for 4 quarters.
  - "Both equally": produces the split roadmap we've been running;
    ruled out by consensus.
Owner: Head of Product
Deadline: Roadmap locked by 2026-06-01
Revisit condition: If enterprise pipeline grows to ≥15% of new ARR by
  end of Q3, re-evaluate before Q4 planning.
Open questions:
  - What happens to enterprise deals already in flight? (Sales lead, answer
    by 2026-05-25)
  - Which current features are enterprise-only and will be deprioritised?
    (PM, answer by 2026-06-01)
```

**What makes this work:** the pre-work brief named the decision (targeting, not features) before participants walked in; the round-robin on pre-work answers forced everyone to surface their mental model before debate; the affinity cluster in the groan zone named the poles of disagreement explicitly rather than leaving them implicit; the decision-capture ruled out alternatives, not just the chosen option; the revisit condition is specific.

---

### Example 2 — Counter-example: a discussion that masquerades as a workshop

**Setup.** A PM sends a calendar invite: "Roadmap workshop — come ready to discuss Q3." No pre-work. The session opens with 25 minutes of slide-deck context-sharing. A VP dominates the first 40 minutes with a set of feature requests. The PM tries to facilitate but has no agenda structure, no time-boxes, and no decision-capture template. The session ends with "great discussion — let's take next steps offline." No decision. No owner. No deadline.

**What failed and why:**

- **No stated purpose.** "Discuss Q3" is a topic, not a decision. Parker: a gathering without a specific purpose defaults to whoever has the most status driving toward their preferred outcome.
- **No pre-work.** Participants arrived without a shared base; the first 25 minutes repaired that at synchronous cost. The session budget was spent on setup.
- **No time-boxes.** The VP's 40-minute monologue was not the groan zone — it was a status broadcast. With no time structure, there was no mechanism to protect divergence from dominance or to force convergence.
- **No decision-capture.** "Next steps offline" is the tell. If the decision could have been deferred to "offline," the workshop was not needed.
- **No ruling-out.** Options were discussed but not evaluated. Participants left with different mental models of what had been decided.

**Fix:** Any one of the missing elements — pre-work, time-boxes, decision-capture template — would have improved the outcome. The combination of all four missing is not a bad workshop; it is a regularly-scheduled meeting with a different name.

## Common Pitfalls

### Pitfall 1 — Workshop as a substitute for a decision

**Symptom.** The PM calls a workshop when a decision has already been made by a senior stakeholder and the workshop's actual purpose is to generate the appearance of consensus.

**Consequence.** Parker: gathering people under false pretenses damages trust in the convener. Participants who realise the decision was pre-determined disengage from future workshops and from PM-led processes generally.

**Fix.** Apply the test before scheduling: "If every participant in this room said no, would the decision change?" If the answer is no, don't run a workshop. Run a 1:1 alignment conversation with dissenting stakeholders instead.

### Pitfall 2 — Collapsing the groan zone

**Symptom.** The PM, uncomfortable with the group's unresolved tension during divergence, jumps to convergence early — summarising options prematurely, moving to a vote before all views have been surfaced.

**Consequence.** Kaner: false convergence. Participants who didn't get to surface their concern agree in the room and dissent in Slack two days later. The decision requires re-litigation.

**Fix.** Name the groan zone out loud when you're in it. "This is the part where things are messy — that's the process working. We've got 15 more minutes in this phase before we start narrowing." Time-boxing the groan zone gives both the facilitator and the group a shared contract for when it ends.

### Pitfall 3 — Decision-capture skipped or deferred

**Symptom.** The session ends with good energy and no document. The PM sends meeting notes 48 hours later. Participants remember the decision differently.

**Consequence.** The meeting notes become the authoritative record. Meeting notes contain contradictory attributions. The decision gets re-litigated. The workshop's cost — everyone's time — produces no durable artifact.

**Fix.** The decision-capture template is open in the room on a shared screen. It is filled in before the session closes. The PM reads it aloud. Corrections happen in real time, not 48 hours later.

### Pitfall 4 — Pre-work sent too late or skipped

**Symptom.** The pre-work brief goes out the night before. Participants arrive uninformed. The first 30 minutes are context-sharing.

**Consequence.** The synchronous session is used for async work. The groan zone is compressed or missed. The output is proportionally weaker.

**Fix.** Pre-work sent less than 48 hours before a session is not pre-work — it is an agenda email. Design the pre-work brief as a 30-minute read + task. Send it 2-3 days before. If the brief isn't ready 48 hours out, delay the session.

### Pitfall 5 — Parking lot never revisited

**Symptom.** Items parked during the session appear in the parking lot document. Nobody is assigned an owner. The document closes. The items silently die.

**Consequence.** Participants learn that parking an item means discarding it. Future participants stop raising concerns when they can see the parking lot is a graveyard. Facilitation trust erodes.

**Fix.** Every parked item leaves the session with an owner and a deadline. The PM follows up in week one. If an item has no owner who will act on it, surface that explicitly: "This is out of scope for this decision — it belongs in [other process/next session]."

### Pitfall 6 — No ruling-out in the decision-capture

**Symptom.** The decision-capture records what was chosen but not what was ruled out. Participants leave with different understandings of which alternatives are off the table.

**Consequence.** The ruled-out options re-enter conversation at the next meeting, roadmap review, or planning cycle, dressed in slightly different language. The workshop's convergence is undone incrementally.

**Fix.** Ruling-out is mandatory in the decision-capture template — not optional. For each alternative seriously considered and not chosen, write one sentence on why it was set aside. This is the work that separates a workshop from a meeting: explicit ruling-out makes the option space legible.

## References

- **Sam Kaner** — *Facilitator's Guide to Participatory Decision-Making* (3rd ed., 2014, Jossey-Bass). The diamond model (divergent/convergent rhythm), the groan zone, facilitation modalities (heads-up start, context-dump bypass, best-guess mode, progress labels, parking lot), and the inclusive-facilitation principles underpinning psychological safety. The structure of Phases 1 and 2 in Application traces directly to Kaner's facilitation framework.
- **Priya Parker** — *The Art of Gathering: How We Meet and Why It Matters* (2018, Riverhead). Purpose-first gathering design: the best gatherings know why they cannot happen any other way; gatherings without specific purpose default to whoever has the most status. The not-a-workshop test and the counter-example in Examples both apply Parker's framing. The Pitfall 1 fix ("would the decision change if everyone said no?") is a direct application of Parker's host-responsibility framing.
- **Jake Knapp, John Zeratsky, and Braden Kowitz** — *Sprint: How to Solve Big Problems and Test New Ideas in Five Days* (2016, Simon & Schuster). Time-boxed decision structure, the sticky-note sprint (silent parallel ideation), and the forcing function of the clock. The convergent phase time-boxing in Step 10 and the silent sticky-note sprint in Step 8 trace directly to Knapp et al.'s sprint facilitation mechanics.
- **Jake Knapp** — Google Ventures Sprint methodology (2010–2016, foundational work preceding the book). The sprint format's influence on time-boxed product facilitation practice; the sticky-note approach for decoupling ideation from social status.
- **Related Hornero skills:**
  - [`cross-functional-orchestration`](../cross-functional-orchestration/SKILL.md) — when the workshop's output is a product decision that requires multi-role synthesis, orchestration is the downstream step. Workshops surface alignment or tension; orchestration produces the decision recommendation.
  - [`positioning-statement`](../positioning-statement/SKILL.md) — the positioning workshop example in this skill is a common trigger for the positioning-statement skill. The workshop narrows the audience and competitive framing; positioning-statement drafts the artifact from that narrowed frame.
  - [`prd`](../prd/SKILL.md) — workshops that produce a prioritised feature direction or solution hypothesis feed directly into PRD authoring. The decision-capture document is the PM's input to the PRD's problem framing.
  - [`strategy-kernel`](../strategy-kernel/SKILL.md) — a strategy workshop that correctly runs divergent/convergent should produce diagnosis inputs and guiding-policy candidates. The kernel skill formalises those into a Rumelt-compliant strategy artifact.

*Synthesis cue: adapted from Dean Peters' [Product-Manager-Skills](https://github.com/deanpeters/Product-Manager-Skills/tree/main/skills/workshop-facilitation) (CC BY-NC-SA 4.0). Hornero's re-synthesis is MIT.*