T-Shirt Sizing is where the data team provides an educated guesstimate of how long an Information Product may take to deliver based on a relative sense of effort.
Unlike traditional estimation techniques that aim for precision, this approach is designed for speed and simplicity, enabling data teams to provide a useful estimate in minutes, not hours.
The data team considers several key factors to assign a T-Shirt size:
Data Availability, is the required data already collected, or does it need to be sourced?
Business Logic Complexity, does it require straightforward rules, or complex transformations, aggregations, or calculations?
Standardisation, does it need to combine data from multiple disparate sources?
Delivery Type Complexity, is it delivering a simple dashboard, a data feed or a complex multi-layered report?
Data Sync Rate, does data need to be refreshed based on a periodic batch or a near real-time data stream?
The data team uses T-Shirt sizes (Small, Medium, Large, etc.) to categorise the expected amount of time required. These sizes allow them to quickly compare the relative complexity of different Information Products without getting lost in unnecessary details.
The data team applies sizing ratios to ensure consistency across Information Products.
A Medium-sized Information Product takes roughly one and a half the time to deliver compared to a Small.
A Large takes one and a half the time to deliver compared to a Medium.
A Extra Large takes one and a half the time to deliver compared to a Large.
Only the data team can size an Information Product. Beware of an anti-pattern where others assign sizes and expect the data team to deliver to those estimates.
The T-Shirt size assigned by the data team is not an estimate, not a promise, and not a deadline. It is simply a relative best guess to help guide prioritisation.