Clustered vs Stacked Column Chart in Power BI: Bar vs Column, Legend, 100% Stacked and Small Multiples

By Ravindra Bagale Last updated: 19 October 2026

The column chart is the visual you will use most, and it comes in several versions that look alike but answer different questions. Bar or column? Clustered or stacked? Stacked or 100% stacked? Pick the wrong one and the chart still draws, but it tells the wrong story. In this chapter you build each version from the same Blinkit sales table, learn the one question that decides between them, and finish with Small multiples and the Tooltips well.

What you'll learn in this chapter

Bar vs column: long labels vs time order
Clustered: groups side by side, using Legend
Field order: what goes on X and what goes in Legend
Stacked: parts of one whole
100% stacked: shares, e.g. weekday vs weekend
Small multiples, switching visual type, and the Tooltips well

Practice file: blinkit-sales-2023-2025.csv (1,702 order lines).

Made-up data: fictional Blinkit-style sales, for practice only. Amounts are in ₹ and dates are DD-MM-YYYY.

1. The problem

The category names are long ("Home, Kitchen & Pet Supplies"), and the manager wants to compare online and cash-on-delivery revenue, see each state's category mix, and know which categories sell more on weekends. One chart type cannot do all of that well.

2. The idea in simple words

Chart	Use it when
Bar (horizontal)	Labels are long, or you show a ranked list
Column (vertical)	The X-axis is time (months, years, left to right) or labels are short
Clustered	Comparing sub-groups side by side (Online vs COD)
Stacked	The pieces add up to one whole (each category's share of a state's revenue)
100% stacked	Only the shares matter (weekday vs weekend %)

Put the comparison in Legend, and keep it to 2–4 values; a legend with many items is hard to read.

Classroom line

When do we use a bar chart? When the names are long and the full word must show.

3. The data

Category	Revenue (₹)	Online	COD	Weekday	Weekend
Atta, Rice, Oil & Dals	1,01,540	74,905	26,635	70,247	31,293
Dairy, Bread & Eggs	55,512	38,895	16,617	40,521	14,991
Home, Kitchen & Pet Supplies	52,843	38,068	14,775	52,843	—
Snacks & Munchies	44,655	32,425	12,230	27,250	17,405
Fruits & Vegetables	39,107	27,024	12,083	29,077	10,030
Cold Drinks & Juices	37,186	25,509	11,677	26,708	10,478
Total	3,30,843	2,36,826	94,017	2,46,646	84,197

Home, Kitchen & Pet Supplies has no weekend orders in this data.

4. Step by step

Bar vs column

Add a Stacked bar chart: Y-axis category, X-axis revenue. Every category name shows in full.
Switch it to a Stacked column chart (click the icon while the visual is selected). The fields carry over; the long names wrap or get cut.

Figure 1. A bar chart keeps long category names readable (illustration).

Clustered

New Clustered column chart: X-axis category, Y-axis revenue, Legend payment. Each category gets two columns, Online and COD.
Swap them: X-axis payment, Legend category. Now there are two clusters with six thin columns each: harder to read. Field order matters.

Stacked and 100% stacked

Switch the first chart to Stacked column: Online and COD sit one on top of the other, and each column's height is the category total.
New chart: X-axis state, Y-axis revenue, Legend category: each state's column is split into its categories.
Switch to 100% stacked column, with X-axis category and Legend week_type: every column is 100% tall, showing the weekday and weekend shares.

Figure 2. Clustered puts Online and COD side by side; stacked puts them on top of each other as parts of the category total (illustration).

Small multiples and tooltips

In the state chart, move category from Legend to Small multiples: one small chart per category, each showing revenue by state.
Drag sales into the Tooltips well: hover over a column to see units sold, without a second series on the chart.

Classroom line

What does stacked column chart mean? One on top of the other.

Classroom line

Clustered means side by side; that's the difference.

5. Your result should look like this

Figure 3. 100% stacked: weekday vs weekend share for each category (illustration).

Clustered: Atta, Rice, Oil & Dals has the tallest pair, Online 74,905 and COD 26,635.
State chart with Legend: Maharashtra's column (₹1,22,142) is largest; its biggest segment is Atta, Rice, Oil & Dals (₹39,276).
100% stacked: Snacks & Munchies has the biggest weekend share (39%); Home, Kitchen & Pet Supplies is 100% weekday.

Classroom line

People go out on weekends; maybe nobody cooks at home.

6. What happened

Legend splits each column by a second field, by colour.
Clustered draws the pieces next to each other; stacked piles them up, so the total is the column height.
100% stacked divides each piece by its column total, so all columns are the same height.
Changing visual type keeps the fields in their wells.

Classroom line

What is a legend? It splits the data by colour; that's what we call the legend.

7. Explore further

Weekend snacks. Snacks & Munchies earns 39% of its revenue on weekends, the highest share: weekend habits are different, with more snacking and less cooking. Home, Kitchen & Pet Supplies orders arrive only on weekdays.
Small multiples. They can look crowded with many groups; use them for 3–6 groups, and keep the Y-axis shared so the small charts compare fairly.
Don't stack unlike things. Revenue (₹) and sales (units) are not parts of one whole; show them clustered, or put sales in the tooltip.
More. The main course lessons Bar and Column Charts: Stacked, Clustered, 100% and Small Multiples and Visual Calculations go further.

Classroom line

Small multiples means groups get created here, one small chart each.

Classroom line

Small multiples don't look that good, so we mostly don't use them.

8. Common mistakes

Wrong result	Reason	Fix
Category names cut off	Column chart with long labels	Use a bar chart
Two clusters of thin columns	Category in Legend, payment on X	Category on X, payment in Legend
Revenue and units piled up	Stacked used for unlike values	Clustered, or units in Tooltips
Legend with 20 colours	Field with many values in Legend	Keep Legend to 2–4 values
100% chart says nothing	Values are not parts of one whole	Use a Legend like week_type

Ravindra Bagale's Tip

Before choosing a chart, finish this sentence: "I want to compare …". If it is groups side by side, cluster. If it is parts of one total, stack. If it is only the percentages, use 100%.

Classroom line

Interviews will ask the difference between a bar chart and a column chart.

9. Practise on your own

Practice task

Using blinkit-sales-2023-2025.csv:

Build a clustered column: category on X, revenue on Y, payment in Legend. What is COD revenue for Dairy, Bread & Eggs?
Build a 100% stacked column with category on X and week_type in Legend. What is Fruits & Vegetables' weekday share?
Build a stacked column with state on X and category in Legend. What is Karnataka's Atta, Rice, Oil & Dals segment?
Online vs COD overall: which chart and what totals?

Show solution

₹16,617.
74% (29,077 of 39,107).
₹23,464.
A simple column (or a pie) by payment: Online ₹2,36,826, COD ₹94,017.

10. Think like a professional

Business view. The chart type is part of the message; pick it from the question.
Time left to right. Use columns (or lines) for time; bars for categories and rankings.
Few legend items. Two to four colours read at a glance.
Use the tooltip. Secondary numbers belong in the tooltip, not as another series.

Interview questions

Bar vs column chart? Bars are horizontal and suit long labels and rankings; columns are vertical and suit time order.
Clustered vs stacked? Clustered places sub-groups side by side; stacked piles them as parts of one total.
When is 100% stacked useful? When only the shares matter, e.g. weekday vs weekend.
What do Small multiples do? Split one visual into a grid of small charts, one per value of a field.