A customer moves from London to Berlin. Should last year's orders now count as German? Your answer is the SCD type.
A customer moves house. Do you correct every old invoice to the new address, or leave each invoice showing where they lived when it was sent?
Choosing wrong silently restates published history — last quarter's numbers change and nobody knows why.
Dimension attributes change, and the warehouse has to decide what happens to history. Type 1 overwrites: the dimension always shows current truth, and every past fact is restated to match. Type 2 versions: the old row is closed with an end date, a new row opens, and each fact keeps pointing at the version that was current when it happened — so history stays as it was reported. Type 3 keeps a 'previous value' column for the one attribute where you want both. The reason this is a perennial interview question is that it exposes whether someone understands that a warehouse stores history, not state, and that the choice is a business decision rather than a technical one.
SCD Type 1 overwrites the attribute, so all history is restated to current values. Type 2 closes the old row with an end date and inserts a new version with a fresh surrogate key, so facts stay joined to the value that was true at the time. Type 3 adds a previous-value column. Type 2 is the default for anything you report on, because restating history silently changes last quarter's numbers.
Slowly Changing Dimensions For Data Engineers — Seattle Data Guy, 8:15