Both scales are trying to stop fake precision
Fibonacci cards and T-shirt sizes look different, but they solve the same problem: teams should not pretend uncertain work can be estimated with tiny precision. A rough scale keeps the discussion focused on meaningful differences. The question is not whether thirteen is mathematically perfect. The question is whether the team sees enough uncertainty to treat the item differently from an eight.
Use Fibonacci when the team can compare stories
Fibonacci works well after the team has enough shared history to compare backlog items. The widening gaps make larger estimates feel intentionally coarse, which helps people avoid false confidence. It is also easier to connect Fibonacci-style points to velocity trends, as long as the team remembers that velocity is a planning signal, not a personal productivity scoreboard.
Use T-shirt sizing when the conversation is earlier
T-shirt sizing is often better before detailed refinement. Small, medium, large, and extra large are easy for product, design, and engineering stakeholders to understand together. That makes the scale useful for discovery, roadmap slicing, or triage. Once the team is closer to sprint planning, they can translate the most relevant work into a more specific point scale.
Do not switch scales silently
The scale matters less than shared meaning. If a team changes from T-shirt sizes to Fibonacci, write down what changed and recalibrate with a few known examples. Silent scale changes make historical estimates hard to compare and can turn a useful ritual into a debate about labels. A scale is only useful when the team can explain it consistently.