The Lego City Exercise: Teaching Estimation Hands-On

You can explain story points on a whiteboard for an hour and watch the doubt stay in people's faces. Then you hand them a box of Lego, start a timer, and within two rounds the whole idea clicks — because they've just lived it. The Lego City exercise is the single most reliable way I've found to teach relative estimation to a skeptical team. Here's exactly how I run it.

Why bricks beat slides

Estimation feels abstract until it has consequences you can see. Lego makes the feedback loop physical and immediate: you estimate how much you can build, you build it, and you compare. There's no hiding behind "well, software is different." The gap between what the team predicted and what it actually finished is right there on the table, in plastic.

What you need

• A bin of Lego or building blocks — enough for a few small structures.
• A timer set to 15 minutes per sprint.
• A Planning Poker deck — physical cards or the online tool on everyone's phone.
• A visible board to track estimated vs. actual for each round.

The build backlog

Prepare a list of structures that gets progressively harder. The order matters — you want the team to hit something genuinely large so they experience why splitting is necessary.

1. A simple small house.
2. A house with a garden and a fence.
3. A two-story house.
4. A car.
5. A bus (the deliberately big one).

How a round works

Each 'sprint' is 15 minutes. Before the timer starts, the team estimates — using Planning Poker — how many of the structures they can complete this sprint, or sizes each structure relative to the small house as a baseline of '1'. Then they build. When the timer stops, count what actually got finished to your definition of done (a house missing a roof doesn't count — that conversation alone is worth the exercise).

1. Estimate: size the structures relative to the small house, then predict how much fits in 15 minutes.
2. Build: the team constructs, against the clock.
3. Review: count completed structures. Compare to the estimate. Write both on the board.
4. Repeat: run 2–3 sprints. Watch the estimates get sharper each round.

The aha moments to watch for

• "We way overestimated what we'd finish" — the classic first-sprint optimism, exactly what happens with real backlogs.
• "The bus should've been split" — the team feels why a 13 needs breaking down before it's committed.
• "Our second guess was much closer" — velocity emerging from observation, not from a formula.
• "We argued about what 'done' means" — the definition-of-done conversation, surfaced naturally.

Timing, group size, and agenda

The full workshop fits comfortably in 90 minutes: ten minutes of setup and framing, three 15-minute sprints with five-minute reviews between them, and fifteen minutes of debrief at the end. With more than seven or eight people, split into two competing cities — the rivalry raises the energy, and comparing two teams' estimate-versus-actual boards at the end doubles the learning material. Keep teams at three to five builders; beyond that, people stand around holding bricks.

One logistical note that sounds trivial and isn't: don't provide too much Lego. A generous pile invites gold-plating — someone will spend eight minutes on a decorative chimney. A slightly scarce pile forces the same scope conversations real teams have when capacity is tight, which is exactly the muscle you're training.

Variations for different audiences

• For managers and stakeholders: run the identical exercise, but in the debrief emphasise the forecasting angle — how the third-round estimate was trustworthy in a way the first never could be. This lands the 'velocity needs a few sprints' message better than any slide deck.
• Remote teams: swap bricks for a shared drawing canvas — 'build' houses as drawings with mandated elements (walls, roof, windows, fence). It's less tactile but preserves the estimate-build-review loop, and the Planning Poker voting works exactly the same through the online tool.
• Adding requirement changes: in sprint three, walk in halfway and announce the bus now needs a second deck. The groans are the lesson — mid-sprint scope change has a visible, physical cost nobody argues with afterwards.

Pitfalls I've hit so you don't have to

The master builder: one Lego enthusiast quietly takes over while the rest watch. Counter it by requiring that whoever estimated a structure isn't allowed to build it alone. The skipped review: teams want to jump straight into the next build round — don't let them, because the comparison on the board is the actual teaching moment, not the building. And the trap of leading the debrief: if you announce the lessons yourself, you've turned the workshop back into a lecture. Ask questions and wait out the silence; the team saying "we should have split the bus" out loud is worth ten facilitators saying it for them.

Logistics people always ask about

How much Lego do you actually need? A single medium bin (roughly a shoebox and a half) per team is plenty — remember, scarcity is a feature. Does it work with adults who think Lego is childish? Yes, and faster than you'd expect: the timer converts irony into competitiveness within about ninety seconds. Do you need actual Lego? No — wooden blocks, Duplo, even index-card origami all work; the loop of estimate, build, compare is the exercise, not the brand. And if someone asks whether this scales to leadership offsites: it does, and the estimation lessons land hardest on exactly the people who turn estimates into deadlines.

Bringing it back to real work

Close the workshop by drawing the parallel explicitly: the bus is your epic, 15 minutes is your sprint, and the board is your velocity chart. The team that just felt these ideas will carry them into real refinement far more readily than one that only heard them.

There's a full write-up with the story cards and facilitation notes on the Lego City exercise page, and when you're ready to estimate real work, you can start a free Planning Poker session with your team.