Qualitative vs. quantitative
A dimension holds qualitative data — categories, names, dates, geography. A measure holds quantitative data you can do arithmetic on — sales, profit, quantity.
Tableau colours them: dimensions appear in blue, measures in green. That colour follows the field onto the shelves, so you can read a view's structure at a glance.
Why measures get aggregated
Drop Sales on Rows and Tableau shows SUM(Sales), not a raw value. That's deliberate. A view shows one mark per combination of dimensions, so every measure has to collapse to a single number per mark.
Change the aggregation from the field's dropdown — Sum, Average, Median, Count, Count Distinct, Min, Max. The function always appears in front of the field name, so you can never be unsure what you're looking at.
When Tableau guesses wrong
Tableau assigns roles on connection, based on data type. Numeric fields become measures. That's usually right, but not always — a Store ID or a postal code is numeric and yet summing it is meaningless.
Right-click the field and choose Convert to Dimension. The reverse works too: a text field of numbers stored as strings can be converted to a measure once you fix the data type.
A separate axis: discrete vs. continuous
Dimension/measure is about what the data is. Discrete/continuous is about how Tableau draws it. Discrete fields (blue) produce headers and distinct panes; continuous fields (green) produce an axis.
Most dimensions are discrete and most measures are continuous, which is why people conflate the two — but you can have a discrete measure or a continuous date dimension. Keeping the two ideas separate saves a lot of confusion later.