What Is the Tower Building Challenge?
The tower building challenge gives each team identical materials (typically spaghetti, tape, string, and one marshmallow) and a fixed time limit (usually 18 minutes) to build the tallest freestanding structure with the marshmallow on top. It’s one of the most-studied team building exercises in existence, run with everyone from kindergarten children to business school graduates to corporate executives.
The exercise was popularised by Tom Wujec in a 2010 TED talk that has been watched over seven million times. He ran it with dozens of groups and found a result that should interest every manager: kindergarten children consistently build taller towers than business school graduates. Not because children are better engineers, but because they prototype while adults plan.
That single finding contains the lesson the exercise teaches. Teams that spend 15 minutes planning and 3 minutes building almost always lose to teams that start building immediately and adjust as they go. The exercise compresses an entire project cycle into 18 minutes and shows you, in real time, whether your team plans too much and tests too little.
Materials and Setup
Each team gets exactly the same kit. Consistency is what makes the exercise fair and the comparison meaningful.
Per team, you need:
- 20 sticks of uncooked spaghetti (standard thickness, not thin or thick)
- 1 metre of masking tape
- 1 metre of string
- 1 marshmallow (standard size, not mini or jumbo)
- A pair of scissors
Additional materials for the facilitator:
- A tape measure or ruler for measuring tower height
- A stopwatch or timer visible to all teams
- A flat surface for each team (table or floor)
Set up stations before the teams arrive. Place the materials at each station face down or in a bag so teams don’t start handling them during the briefing. The reveal is part of the exercise.
How to Run the Exercise: Step by Step
The full session takes 30 to 45 minutes: 5 minutes briefing, 18 minutes building, 5 minutes measuring, and 10 to 15 minutes debriefing.
Step 1: Brief the Teams (5 minutes)
Gather everyone together. Explain the rules clearly:
- The goal: Build the tallest freestanding structure. The marshmallow must be on the very top.
- Time limit: 18 minutes exactly. When time is called, hands off.
- Materials: Use as much or as little of the spaghetti, tape, and string as you want. You can break the spaghetti. You can cut the tape and string.
- The marshmallow: Must be whole and on top. Not cut, not eaten, not glued. It must sit on the structure under its own weight when you let go.
- Freestanding: The structure must stand on its own. No holding it, no leaning it against a wall, no taping it to the table.
- Measurement: Height is measured from the table surface to the top of the marshmallow.
Then reveal the materials. Start the timer. Walk away.
Step 2: Build Phase (18 minutes)
Your role during the build is observation, not intervention. Watch for these patterns:
- The planners: Teams that spend the first 10 minutes drawing designs and debating approaches, then rush to build in the last 8 minutes. These teams often fail because they discover structural problems too late to fix them.
- The prototypers: Teams that start building immediately, testing small structures, learning what works, and iterating. These teams usually win.
- The dominator: One person who takes charge and makes all the decisions. Watch whether the rest of the team disengages.
- The marshmallow mistake: Teams that leave the marshmallow until the very last second, place it on top, and watch their entire structure collapse. The marshmallow is heavier than people think. Teams that test with the marshmallow early learn this. Teams that don’t, don’t.
Give time warnings at 10 minutes, 5 minutes, and 1 minute remaining.
Step 3: Measure and Score (5 minutes)
When time is called, everyone steps back. Walk to each tower with the tape measure. Measure from the table surface to the top of the marshmallow. If the tower has collapsed or the marshmallow isn’t on top, the score is zero.
Announce the scores publicly. The reveal is part of the learning. Teams that scored zero need to hear that they scored zero.
Step 4: Debrief (10 to 15 minutes)
This is where the exercise pays for itself. Without a debrief, it’s a fun 18 minutes with spaghetti. With one, it changes how the team approaches their next project.
Debrief Questions That Connect to Real Work
The marshmallow challenge works because it mirrors how teams handle every project: planning versus doing, assumptions versus testing, talking versus building. The debrief makes that connection explicit.
Questions that work:
- “How much time did your team spend planning versus building? If you could redo it, how would you split the 18 minutes?”
- “Did anyone test the marshmallow’s weight on your structure before the final placement? If not, why not?”
- “Who made the key decisions in your team? Did everyone contribute, or did one or two people drive?”
- “Think about your last real project at work. Did it look more like your team’s approach to the tower, or different? How?”
- “The teams that win this exercise are almost always the ones that start building immediately and adjust as they go. What would it take for your team to adopt that approach in real projects?”
The last question is the one that matters. If the team recognises that they spent too long planning and not enough time testing, and if they can name one specific change to make in their next project, the exercise has earned its time.
What the Research Shows
Tom Wujec ran this exercise with dozens of groups across different professions. The results reveal patterns that apply directly to how teams work.
| Group | Average Tower Height | Pattern |
|---|---|---|
| Kindergarten children | ~26 inches | Start building immediately, iterate constantly, test the marshmallow early and often |
| CEO’s | ~22 inches | Prototype like children when given a second chance, plan like business students on the first try |
| Business school graduates | ~10 inches | Spend most of the time planning and debating, build at the last minute, marshmallow collapses the structure |
| Lawyers | ~15 inches | Better than business students because they negotiate roles quickly, but still plan-heavy |
| Engineers and architects | ~39 inches (highest recorded) | Understand structural principles, use triangular bracing, place marshmallow early |
| Executive administrators | ~24 inches | Better than average because they’re used to executing plans, not just making them |
The pattern is consistent across every group Wujec tested: teams that iterate beat teams that plan. The children win because they don’t assume they know what will work. They try something, see if it holds, adjust, and try again. Business school graduates lose because they assume a single perfect plan exists and spend all their time trying to find it.
The implication for your team: if your last three projects ran over deadline or produced a result that didn’t match the brief, the problem probably isn’t the plan. It’s that nobody tested the plan early enough to learn what was wrong with it.
Variations for Different Goals
The same basic exercise can be adapted to teach different lessons depending on what your team needs to learn.
| Variation | Change | What It Teaches |
|---|---|---|
| Limited Communication | Teams can only talk during the first 5 minutes. After that, silent building. | How well the team communicates intent upfront. If the plan isn’t clear enough to execute without discussion, it wasn’t a good plan. |
| Role Assignment | Each person gets a fixed role: architect (designs), builder (constructs), resource manager (handles materials), quality checker (tests stability). No role switching. | Division of labour and handoff quality. The architect can’t touch the building; the builder can’t change the design. |
| Two Rounds | Run the exercise twice with the same teams. Second round uses lessons from the first. | Iteration and learning speed. How much did the team actually learn from failure versus just trying harder? |
| Unequal Resources | Give one team extra spaghetti, another team less tape, a third team no string. | Resource constraints and adaptability. How does the team respond when the playing field isn’t level? |
| Blind Build | One person can see and talk but can’t touch materials. The rest can touch but can’t see the design plan. | Communication precision and trust. The describer must give clear instructions; the builders must trust and execute. |
Common Mistakes Teams Make
Having run this exercise many times, the same mistakes appear in almost every group. Naming them in the debrief helps the team avoid them in real projects.
The Single-Design Trap
The team debates for 12 minutes, agrees on one design, and commits all their materials to building it. When the design fails at minute 15, there are no materials left to try something else. In a real project, this is the team that writes a 50-page specification before writing a line of code.
The Marshmallow Surprise
The marshmallow weighs more than people expect. Teams that leave it until the final seconds watch their structure buckle under a weight they never tested. In a real project, this is the team that discovers the client’s actual requirements don’t match the brief on the day of delivery.
The Tape Problem
Teams use too much tape at the base and not enough at the joints. The structure looks solid but is actually top-heavy and brittle. When the marshmallow goes on, the weak joints fail. In a real project, this is the team that over-invests in the foundation and under-invests in the parts that actually face the user.
The Silent Expert
Someone on the team understands structural principles (triangles are stronger than squares, a wider base supports more weight) but doesn’t speak up because the loudest person has already decided on the approach. The team fails with a design that the quiet person knew wouldn’t work.
Frequently Asked Questions
How long does the marshmallow challenge take?
The build phase is 18 minutes. Allow 30 to 45 minutes total: 5 minutes for the briefing and setup, 18 minutes for building, 5 minutes for measuring, and 10 to 15 minutes for the debrief. Don’t shorten the debrief. The building is the setup. The conversation is where the learning happens.
What materials do you need for the tower building exercise?
Each team needs 20 sticks of uncooked spaghetti, 1 metre of masking tape, 1 metre of string, 1 standard marshmallow, and a pair of scissors. All teams get identical materials. Cost is approximately £1 to £2 per team. The facilitator also needs a tape measure, a visible timer, and flat surfaces for each team.
Why do kindergarten children outperform adults at the marshmallow challenge?
Children start building immediately and iterate based on what works and what doesn’t. Adults spend most of the time planning and debating, then rush to build at the end. When the plan fails, adults have no time to adjust. Children test the marshmallow early and often. Adults assume their plan is correct and discover the marshmallow’s weight only when the structure collapses. The lesson: prototype early, test assumptions, and don’t assume one perfect plan exists.
Can the marshmallow challenge work for remote teams?
Partially. Send each team member an identical materials kit to their home and run the exercise over video call with everyone building simultaneously. The challenge is that team members can’t physically collaborate on a single structure. A better remote adaptation is a digital tower-building tool or a shared whiteboard where teams sketch designs and vote on approaches. The prototyping lesson transfers, but the physical collaboration element is lost.
What’s the tallest marshmallow tower ever built?
The tallest recorded towers in Tom Wujec’s research were built by engineers and architects, reaching approximately 39 inches (about 1 metre). These teams understood structural principles like triangular bracing and wide bases. The average across all groups is approximately 20 inches. Most business school graduate teams build towers around 10 inches, and roughly 20% of all towers collapse when the marshmallow is placed on top.
The Bottom Line
The marshmallow challenge works because it’s a complete project in 18 minutes. The team plans, builds, tests, fails, and either adapts or doesn’t. Every pattern that shows up in the exercise shows up in their real work. The only difference is the stakes.
The exercise isn’t about spaghetti and marshmallows. It’s about whether your team tests its assumptions before committing all its resources, or whether it assumes the plan is right and discovers it isn’t at the last possible moment.
If your team placed the marshmallow at the last second and watched the tower fall, ask them: when was the last time that happened on a real project? The answer tells you everything you need to know about what to change next.