Estimating work in t-shirt sizes (XS, S, M, L, XL) sounds simple and fun. But in practice, it often creates more confusion than clarity. The core problem is that people naturally think in time: hours, days, or weeks. When asked to assign a t-shirt size to a task, they must translate their intuitive sense of effort into an abstract label. Later, someone else has to reverse-engineer that label back into time to plan work.
I understand why teams do it. One argument is that t-shirt sizes are relative. Bert’s XL might be Jenny’s L. Or go to Mark and John. But how often are tasks truly interchangeable between developers? It may be politically incorrect to say, but teams quickly learn who the fast and slow developers are. If a task goes to Pietje, everyone already knows it will not be done in a week!
If the team already has a strong intuition about who will do the work, the real question becomes: why hide that intuition behind a t-shirt size? One answer is accountability. Especially in toxic Scrum environments, it feels much safer to say “this is an L” than to commit to “this will take one week.” Abstract sizes create plausible deniability. Nobody can be clearly wrong, and nobody can be clearly held responsible.
Another issue is that t-shirt sizes mean different things across teams. Switch teams, and you have to relearn what “small” or “large” even means. The only time abstractions that truly work across teams are things like “A shitload of time” or “I have no idea”. I've never heard them in sprint plannings however.
So when nearly everything we do is already measured in time (waiting, walking, typing, thinking), why keep pretending we are bad at it? Humans have estimated time reasonably well for thousands of years. Maybe it is time to drop the costume sizes and just say what we actually mean.
Written by Loek van den Ouweland on Jan. 7, 2026.