Lightning Decision Jam

What it is

LDJ is published by AJ&Smart, the workshop-design agency behind several compressed-format variants (design sprints, remote design sprints, and LDJ itself). Jonathan Courtney wrote the canonical public reference — The Lightning Decision Jam: a design exercise that will change your daily life for the better in UX Collective (2018) — and the format has spread widely through the product and agile communities.

The structural innovation is to collapse the standard retrospective-plus-action-planning flow into eight time-boxed steps, most of them silent:

1. What’s working (silent, ~4 min)
2. Challenges / issues (silent, ~7 min)
3. Dot vote on top problems (~2 min)
4. Reframe the top problem as “How might we…?” (~3 min)
5. Solution ideation — silent, individual (~7 min)
6. Dot vote on solutions (~2 min)
7. Prioritize on a 2×2 of effort versus impact (~5 min)
8. Commit — pick solutions to ship, assign owners (~5 min)

Total: ~35 minutes of active work, ~10 minutes of transitions and framing. The format is tightly prescriptive — the facilitator’s job is to enforce the clock, not moderate discussion. Teams new to the format often complain that there isn’t enough talking. That’s the point. Silent work prevents loud voices from anchoring the group, silent voting surfaces honest priorities, and the effort × impact grid forces trade-offs instead of accepting everything.

LDJ is not a replacement for a full retrospective or for the 3-hour workshop guides. Those formats are designed for diffuse problems — when waste is spread across multiple categories and the team needs the slower conversational format to disambiguate. LDJ is for the opposite case: one specific, focused problem that the team wants to solve in under an hour.

When to use it

Reach for Lightning Decision Jam when:

• The team has a single, specific problem to solve. “Sprint planning is running too long,” “our PR reviews have become a bottleneck,” “standups have started feeling pointless.” Scoped, observable, committable.
• Full 3-hour workshop bandwidth isn’t available. LDJ fits in a standard retro or standup slot.
• The team has a loud-voice problem and the open retrospective format is being dominated by one or two people. Silent steps neutralize that.
• You want to introduce a new team to structured facilitation without scaring them with a full-day workshop commitment.
• A recent retrospective generated an action list that didn’t get executed. Re-running the same retrospective won’t fix it. Running LDJ on the specific unstuck action often does — the effort × impact prioritization and explicit ownership at the end are where LDJ earns its keep.

Don’t reach for Lightning Decision Jam when:

• The problem is diffuse and multiple categories are in play. A 45-minute session will produce a superficial plan across too many fronts. Reach for the longer workshop guide that matches the dominant category instead.
• The team needs conversation to reach shared understanding. LDJ’s silent structure prevents the conversation that’s sometimes actually the point. If people need to talk through the shape of the problem, do that first in a different format; LDJ is for after the problem is understood and a solution is the output.
• The decision requires deep expertise the team doesn’t have in the room. LDJ’s format compresses decision-making; it doesn’t substitute for missing knowledge.
• The team is in psychological-safety crisis. LDJ’s silent format can mask disengagement — people check out and nobody notices because the format doesn’t require them to speak. Address the safety issue first.
• The action requires cross-team coordination. LDJ produces team-owned actions. If the action depends on another team’s commitment, LDJ’s 5-minute “assign owners” step won’t resolve it; you need a coordination conversation instead.

How to run it

Total time: 45 minutes — non-negotiable. LDJ’s time-boxes are the discipline.

Setup (5 min before the clock starts). Post the eight steps visibly. Confirm the problem scope in one sentence. Make sure everyone has a stack of stickies and a way to dot-vote (three dots each, or three virtual dot stickers). Introduce the silent-work rule — during timed steps, no discussion.

Step 1 — What’s working (4 min, silent). Each participant writes one per sticky, placing on the board as they go. No discussion. The facilitator enforces silence. Ends at 4:00.

Step 2 — Challenges / issues (7 min, silent). Same mechanic, different prompt: what isn’t working, what’s in the way, what’s the problem. Still silent. Ends at 11:00.

Step 3 — Dot vote on problems (2 min). Each participant gets three dots. Place them on whichever challenges/issues stickies feel highest-priority. One dot per sticky per person; three total. Count dots when time’s up. The highest-voted sticky is the problem the rest of the session will tackle.

The facilitator resists the urge to “cluster first.” If the team wants to tackle several adjacent issues, run LDJ again later on the next one. Pick one.

Step 4 — Reframe as “How might we…?” (3 min). The highest-voted problem gets rewritten as a question beginning “how might we…?” — a lightweight framing that shifts the group from complaint into solution-mode. This is the only step with any conversation; the facilitator can guide the reframe out loud, but keep it tight.

The phrasing matters. “Standups are too long” is a complaint. “How might we make standups feel useful without extending them?” is a question that invites solutions. Spend 2 minutes getting the framing right; the ideation will be only as good as the question.

Step 5 — Solution ideation (7 min, silent). Each participant writes solutions, one per sticky, against the “how might we” framing. Silent. Write as many as you can — quantity beats quality at this stage. Place on a new section of the board.

Step 6 — Dot vote on solutions (2 min). Same as Step 3 — three dots per person, one per sticky, place on the solutions that look strongest. Count when time’s up. Keep the top 6–8.

Step 7 — Effort × Impact 2×2 (5 min). Draw a 2×2 grid: effort on one axis, impact on the other. Place the top-voted solutions into quadrants. The goal isn’t precise placement — it’s quadrant assignment. The upper-left (high impact, low effort) is the commit-now quadrant.

Use the 2×2 prioritization matrix rules here — silent individual placement first, then group to converge.

Step 8 — Commit (5 min). Pick 1–3 solutions from the upper-left quadrant. For each:

• One-sentence description of what the solution is.
• Named owner (a person, not a team).
• A deadline or “by when” — usually the next retrospective or next sprint.

Write commitments visibly. Close the session. That’s 45 minutes, end to end.

Worked example

A six-person engineering team has had three retros in a row surface the same complaint: sprint planning runs 90 minutes, people check out, decisions are half-committed. The team agrees to try LDJ rather than running another open-format retrospective. The engineering manager facilitates.

Steps 1–2 (11 min). Silent writing. The board fills with stickies — what’s working (honest pair-programming, good slack channels), what’s challenging (planning runs long, refinement feels rushed, stories aren’t clear, PR reviews pile up, too many meetings). About 22 stickies on the challenges side.

Step 3 (2 min). Dot vote. Winner: “Sprint planning runs 90 minutes and people disengage in the second half.” Seven dots, well ahead of the runner-up.

Step 4 (3 min). Reframe: “How might we make sprint planning’s second half feel as focused as the first?”

Step 5 (7 min). Silent ideation. About 15 solutions land on the board:

• Pre-estimate stories before the meeting
• Split into domain-specific breakout groups
• Move refinement to its own session earlier in the sprint
• Set a hard 45-minute limit
• Only the PO and tech lead attend the full meeting; others drop in as needed
• Use a “story-ready” checklist before planning
• Async pre-read the stories the day before
• Cut the retrospective-at-planning hybrid (they’d merged them some weeks back)
• …and several others

Step 6 (2 min). Dot vote on solutions. Top 6:

• Async pre-read the stories the day before (8 dots)
• “Story-ready” checklist before planning (6 dots)
• Move refinement to its own session earlier in the sprint (5 dots)
• Hard 45-minute time-box (5 dots)
• Domain breakout groups (4 dots)
• Split retro from planning (3 dots)

Step 7 (5 min). Effort × Impact 2×2:

• High impact, low effort: Async pre-read; Hard 45-minute time-box; Split retro from planning.
• High impact, high effort: Move refinement to its own session; Story-ready checklist.
• Lower impact: Domain breakouts (and the other low-voted options).

Step 8 (5 min). Commit. Three upper-left actions, each with an owner:

• Async pre-read the stories the day before planning. Owner: Chris (PO). Ships next sprint.
• 45-minute hard time-box on planning, explicit “park for later” for overflow. Owner: Alex (EM). Starts next sprint.
• Split retrospective from planning into a separate 30-minute slot. Owner: Alex (EM). Calendar change effective next cycle.

Total elapsed time: 45 minutes. The team committed to three actions with owners and deadlines. Next retrospective (now split off into its own slot) will review whether the three actions changed the planning-meeting experience.

Without LDJ, this conversation would likely have taken 90 minutes and generated either a larger list or a mushier commitment. The format’s strictness is what produces the shorter list with named owners.

Common failure modes

• Loose time-boxes. The facilitator lets Step 2 run to 10 minutes because “people were still writing.” The whole format collapses; it becomes a normal retro wearing an LDJ costume. Enforce the clock.
• Open discussion during silent steps. Someone says “oh, that’s interesting” during Step 5 and the silent writing turns into a group conversation. The anchoring problem comes back. The facilitator must interrupt.
• Clustering too much in Step 3. The group wants to combine three related problem-stickies into a single “cluster” and vote on that. This dilutes the format — LDJ picks one specific problem. If several related problems are tied, run LDJ again next time on the next one.
• Accepting all the upper-left solutions. Step 8 says “pick 1–3.” Teams often commit to five or six, which is neither discipline nor decision. Enforce the 1–3 cap.
• Running LDJ on a problem that needs a different format. Diffuse, cross-team, or deeply technical problems don’t fit the 45-minute shape. The result is a superficial action list the team later ignores. Match the tool to the problem.
• Treating commitments as aspirational. The named owner in Step 8 isn’t optional. Without a person (not a role, not a team) owning each commitment, the session’s output evaporates by the next sprint. If no one will own a solution, it doesn’t get committed.

References

In the playbook

• Planning workshop — explicitly mentions LDJ as an alternative format when the full 3-hour workshop isn’t feasible.
• 2×2 prioritization matrix — Step 7 is a compressed 2×2; the matrix guide covers the mechanics in depth.
• Dot voting — Steps 3 and 6 are dot-vote rounds; that guide covers the convergence technique.
• Force Field Analysis — consider running force-field on LDJ’s committed actions for higher-stakes commitments.
• 1-2-4-All — an alternative facilitation pattern for Steps 1–2 and 5 when the team is large and silent stickying alone won’t surface enough ideas.

External references

• Jonathan Courtney, The Lightning Decision Jam: a design exercise that will change your daily life for the better (UX Collective, 2018) — the canonical public write-up.
• AJ&Smart, LDJ on YouTube — facilitator-run video walkthrough.
• AJ&Smart, Pitching Ideas (The LDJ Episode) — the agency’s internal reference.
• Jake Knapp, Sprint: How to Solve Big Problems and Test New Ideas in Just Five Days (Simon & Schuster, 2016) — the design-sprint format AJ&Smart adapted LDJ from.
• Henri Lipmanowicz and Keith McCandless, Liberating Structures — the broader library of compressed facilitation patterns; LDJ sits alongside these philosophically even though it isn’t formally one of them.