The setup is simple on purpose. You need one real, low-stakes problem your organization actually has, split evenly across two small groups, and one invented rule that has nothing to do with the problem's real limitations. The invented rule is the entire point. It is there to force a decision, not to solve a business constraint you already have plenty of.

1
Set the Same Problem, Two Boxes
One real problem. One group gets a rule. One gets a blank page.
30 min
2
≠
Compare What Came Out
Outputs side by side. Name the difference before explaining it away.
30 min
3
Commit to Importing One Constraint
One person. One real project. One invented rule for two weeks.
20 min

Step 1: Set the Same Problem, Two Boxes  30 min

Pick a real problem that's genuinely open and low enough stakes that a rough answer from a workshop won't become policy by accident: redesign the internal wiki homepage, name the new internal tool, draft the first line of the onboarding email, sketch a better layout for the team's weekly update. It has to be something your organization would plausibly work on, or the comparison at the end won't feel like it means anything.

Split into two small groups of three to five and give them the identical brief, with one difference:

Pick a Constraint That's Arbitrary, Not Meaningless
Good options: the design may use no more than three colors, the pitch must fit on one index card, the email may not use the word "welcome," the layout gets exactly one number on it. Bad options: anything that quietly restates a limitation the team already has, like a real budget cap or a real deadline. The rule has to come from nowhere, on purpose, so its only job is forcing a choice, not modeling a real-world tradeoff.

Group A gets the brief and nothing else: solve the problem, present whatever you come up with. Group B gets the same brief plus the one invented rule, non-negotiable, applied strictly. Both groups get the same amount of time, no more than 20 minutes of working time inside the 30-minute step, so neither group can quietly out-effort the other.

Do not tell either group what the workshop is testing. Frame it as two independent teams tackling the same brief, full stop. If Group B knows the rule is "the point," they'll perform insight instead of actually working inside the constraint, and the comparison in Step 2 will be measuring performance, not the effect you're actually there to show.

Enforce the rule for Group B literally, even when it produces something odd. The value of the exercise depends on the constraint actually biting. A facilitator who lets Group B quietly bend the rule when it gets uncomfortable has removed the entire mechanism the workshop is trying to demonstrate.

Step 2: Compare What Came Out  30 min

Bring both groups back together and have each present in under three minutes, no defending, no explaining what they would have done with more time. Just the output, as it stands.

What tends to happen, reliably enough to build a workshop around, is that Group A's answer looks competent and safe: the first reasonable idea that came to mind, polished a bit, with no strong reason to have picked it over three other equally reasonable ideas nobody has to choose between because nothing forced the choice. Group B's answer looks stranger, more specific, and more clearly theirs, because eliminating the obvious move forced the group to search past it. That is the entire mechanism in one sentence: removing the easy option is what makes people find the interesting one.

Group A — No Constraint
Group B — One Invented Rule
Converges fast on the first workable idea. Polished, safe, hard to object to.
Rules out the obvious move immediately. Spends real time searching past it.
Multiple reasonable directions survive, unreconciled, because nothing forced a pick.
One direction, clearly chosen, because the rule ruled the others out.
Sounds like it could have come from most teams doing this brief.
Sounds specific to this group, this rule, this room.

Ask both groups directly: "Which output do you think is more distinct? Which would you have predicted, before seeing it?" Most rooms land on Group B without much prompting, and the more useful conversation is the one right after: why. Get the group to name that it wasn't more talent or more effort at the B table, it was the rule doing work the A table didn't have anything doing for them.

Sometimes Group A's output genuinely wins, and that's a real result too, not a failed workshop. The lesson isn't "constraint always produces better work." It's "constraint reliably produces more distinct work," and distinct is what you want when the safe, expected answer is the thing you're actually trying to get away from.

Step 3: Commit to Importing One Constraint  20 min

The mistake to avoid here is the same one that sinks most workshop closings: asking people to "apply this thinking going forward." That produces nothing, because nothing specific was committed to. Instead, each person names one real project on their desk right now and one small, artificial constraint they will impose on themselves for the next two weeks, on purpose, with no one asking them to.

1
The project: [name a real, current piece of work].
2
The constraint I'm importing, starting now, that nobody assigned me: [one page, one color, one sentence, one day, one meeting instead of three].
3
I'll know it worked if, two weeks from now, I can point to a specific decision the constraint forced that I wouldn't have made otherwise.

Have people read their commitment aloud before the room breaks. Saying it in front of the group who just watched Group B outperform Group A on distinctiveness gives the commitment a kind of social weight that writing it down privately doesn't.

Facilitator Notes

The "That's Not How Real Work Works" Objection

Someone will point out that real projects don't come with an arbitrary rule attached, so the exercise is artificial. It is, on purpose. The response: "You're right, nobody handed you this rule. That's exactly the point. The rule you're missing on your real project isn't missing because none exists, it's missing because nobody, including you, has picked one yet." Don't argue further. Let Step 3 make the point instead.

When a Group Tries to Negotiate the Rule Mid-Exercise

Group B will often ask, ten minutes in, if they can bend the rule "just for this one part." Say no, and say it before the exercise starts, not in the moment, so it isn't a judgment call you're making about their specific idea. The discomfort of not being able to route around the rule is the exercise. Removing it removes the lesson.

Picking a Rule with Real Teeth

A constraint that doesn't actually change anyone's first instinct isn't doing its job. Before running the workshop, sanity-check the rule yourself: would it have ruled out your own first idea for this brief? If not, pick a sharper one.

The workshop doesn't teach anyone to work under constraints they didn't choose. It teaches them to notice when no one has set a constraint at all, and to stop waiting for one to show up before doing their sharpest work.

Create the Conditions. Start at the Table.

Tech CoLab's games are built inside tight, deliberate constraints for exactly this reason. Run the exercise. Run the debrief. See what your team makes when the easy option isn't on the table.

Read: Constraints Are the Medium → Culture Audit Workshop →