T-Shirt Sizing

Purpose

T-Shirt Sizing trades precision for speed. By mapping items to simple size labels (XS, S, M, L, XL) the team can estimate a large backlog in a fraction of the time card-based estimation would take, making it ideal for roadmap-level planning.

At a glance

How to run

1. Define sizes (5 min) — Agree on what each size means using reference stories the team knows well. XS might be a few hours, XL might be multiple Sprints.
2. Present items (2 min per batch) — Read each Backlog item title and brief description.
3. Silent placement (5 min) — Each person writes the size they think fits on a sticky or selects it digitally.
4. Reveal and cluster (10 min) — Show all sizes. For items with disagreement, allow 2 minutes of discussion, then take the majority size.
5. Record (5 min) — Capture the agreed size on each item.

Materials needed

• Backlog items (titles and brief descriptions)
• T-Shirt size cards or labels (XS, S, M, L, XL)
• Reference stories for calibration
• Whiteboard or digital board with size columns

Pitfalls & common mistakes

• Sizes become absolute — Remind the team these are relative; they are not commitments.
• Too many sizes — Keep it to 5 labels max. Adding XXS and XXL muddies the simplicity.
• Skipping calibration — Without reference stories, everyone has a different mental model of what M means.
• Using T-shirt sizes for Sprint commitment — They are too coarse; convert to story points or use card-based estimation when precision matters.

Inclusion & accessibility

• Use both colour and text labels (not colour alone) for each size.
• Allow silent voting so quieter team members are not influenced.
• Provide reference stories in writing beforehand.

Variations

• Dot voting on sizes — Each person places a dot in the size column they think fits; majority wins.
• Fibonacci mapping — After T-shirt sizing, map sizes to Fibonacci numbers for velocity tracking.
• Relative mass valuation — Combine sizing with business value to prioritise.

When NOT to use this

• When precise estimates are needed for Sprint-level commitment — use card-based estimation.
• When the team has only 3–5 items to estimate — card-based estimation is fast enough.

Source & further reading

• Cohn, M. (2005). *Agile Estimating and Planning*. Prentice Hall.
• Leffingwell, D. (2011). *Agile Software Requirements*. Addison-Wesley.