Calculated Column vs Measure

Controls

Filter Context: Region

Filter Context: Channel

What to observe

The highlighted row is just for learning the row-by-row nature of a calculated column.
The Performance Band column never changes when filters change.
The measure cards change immediately because measures respond to filter context.

Visible rows 0

Active row -

Calculated Column

Stored once per row

Measure

Recomputed for current filter context

Common mistake

Trying to use columns for dynamic report logic

Start by filtering the page. The row-level calculated column will stay fixed, but the measures will recalculate for the visible rows.

Salesperson Snapshot

Gold row = one stored row in the model

Rep Region Channel Sales Target Calculated Column

Dynamic Measure Results

Same data, filter-responsive answers

Stored classification

Count of visible rows that already carry the stored calculated-column label Above Target.

Dynamic target achievement

A measure that recalculates using only the currently visible rows in the report.

What would go wrong?

Static

If you expected a stored column to behave like a report-time KPI, it would not recalculate with slicers.

When to use which

Practical guidance

Use a calculated column when

You need a row-level value stored in the model.
You want to slice, group, sort, or relate data using that result.
Examples: banding, flags, concatenated keys, date parts, static classifications.

Sales[Performance Band] =
IF(
    Sales[SalesAmount] >= Sales[TargetAmount],
    "Above Target",
    "Below Target"
)

Use a measure when

You need the result to change with filters, slicers, or visual context.
You want aggregated report logic rather than stored row logic.
Examples: total sales, achievement %, YTD, rankings, dynamic comparisons.

Target Achievement % :=
DIVIDE(
    SUM(Sales[SalesAmount]),
    SUM(Sales[TargetAmount])
)

Real-life difference

A calculated column is prepared during model refresh, then saved. A measure is computed only when the report asks for it.

Calculated column: increases model size, but useful for slicing and row-based attributes.
Measure: does not store row results, but gives dynamic answers for reports.

Common mistakes

Creating a calculated column for something that should react to slicers.
Trying to put a measure on an axis or use it directly as a relationship key.
Using calculated columns for every report KPI and bloating the model.

Rule of thumb: if the answer should be stored per row and reused as a category, use a calculated column. If the answer should change based on filters, visuals, or report selections, use a measure.