Reporting and KPIs
How to Design a Great Dashboard
The three design decisions that make a dashboard readable: which KPIs earn a place, which visual fits each one, and how to set visual hierarchy.
Sajagan Thirugnanam
·
Updated
A great dashboard design comes down to three decisions, made after the audience and the layout are already settled. Which KPIs earn a place on the page, which visual fits each one, and how size, position and color guide the reader's eye to what matters first. Get those three right and the dashboard reads clearly before anyone touches a filter.
This post covers those three decisions. For defining the audience and the business question behind a dashboard, see how to create a dashboard. For page size, the layout grid and Power BI's formatting settings, see dashboard layout and formatting.
Choose the KPIs That Earn a Place
A dashboard gets crowded when every available number goes on it instead of only the ones that matter. Before adding a KPI, run it through three tests:
Is it actionable? If nobody would change what they do when the number moves, it does not belong on the dashboard. A metric that only satisfies curiosity is not a KPI.
Is it tied to a real business question? The metric should trace back to the question the audience needs answered, not to whatever field happened to be easy to pull from the source system. Our guide to KPI reports covers defining KPIs against a business objective in more depth.
Does someone own it? Every KPI needs a person responsible for it being correct and for acting on it. A KPI with no owner is the one that goes stale first, because nobody notices when it breaks.
A KPI that fails any of these three tests is a candidate to cut, not to shrink and tuck in a corner. A dashboard with eight KPIs that all pass these tests works better than one with twenty where half are just available data.
Match Each KPI to the Right Visual
Once the KPI list is set, each one needs a visual that shows what the reader actually needs to see: a single value, a change over time, or a comparison across categories. Four visual types cover most dashboard KPIs:
A card, for one headline number the reader needs to see instantly: total revenue this month, the current open ticket count, the current active user count.
A line chart, for a trend: the same metric measured repeatedly over time, where the shape of the change matters as much as the current value.
A bar chart, for a comparison: the same metric split across categories, such as revenue by region or tickets by agent.
A table, only when the reader needs the raw rows behind a number, not a summary of them. A table is not a fallback for "couldn't decide on a chart."
The measure behind a KPI usually does not change depending on which of these it feeds. A simple measure like:
Total Revenue = SUM ( Sales[Revenue] )works as a card on its own, as a line chart once you add a date field to the axis, or as a bar chart once you add a category field. What changes between the four visual types is the field well the measure goes into, not the DAX behind it. Our guide to measures tables covers organizing measures like this one as your model grows.
The exception is a visual built to show progress toward a goal. Power BI's own KPI visual needs three specific fields:
A base measure to evaluate.
A trend axis field, usually a date.
A target measure or value to compare against.
It then shows an icon and color for whether the metric is ahead of or behind that target.
Picking the wrong visual for a KPI makes the dashboard harder to read even when every number on it is correct. A pie chart shows a split at one point in time. It cannot show a trend, because it has no axis for time to run along. Putting twelve months of revenue into a pie chart forces the reader to compare wedge angles across two things at once, share and month, which the eye is bad at. The same data as a line chart shows the trend immediately.
If a chart makes the reader stop and think about how to read it, the chart, not the reader, is the problem. Our guide to data visualization best practices goes further into chart selection.
Use Size, Position and Color to Set the Order
Once the right KPIs sit on the right visuals, the last design decision is which one the reader sees first. Three tools set that order, and each should be used deliberately, not decoratively:
Size. The largest visual on the page draws the eye first. Reserve that space for the KPI the audience cares about most, not the one that happened to produce the most interesting chart or was easiest to build.
Position. Where a KPI sits on the page also sets reading order. For the mechanics of that, the grid, reading direction and Power BI's alignment tools, see dashboard layout and formatting.
Color. Use color as a signal, not a decoration. One accent color, applied only where something needs attention, such as a KPI below target, does more work than a different color on every chart. Keep everything else neutral so the accent actually stands out. This is different from picking a brand color theme, which dashboard layout and formatting covers; that is about consistency, this is about drawing attention to what matters.
A dashboard where every visual is the same size, in the same shade, competing for attention, forces the reader to do the prioritizing the designer should have done. Use size to say "look here first" and color to say "this one needs action," and leave everything else quiet.
FAQs
How many KPIs should a dashboard show?
There's no fixed number. What matters is that every KPI on the page passes the actionable, tied-to-a-question and owned-by-someone tests above. A dashboard with five KPIs that all pass is more useful than one with fifteen where most are just available data.
When should a table replace a chart?
Only when the reader's actual task is to look up specific rows, not to spot a pattern. If the audience needs to know what happened, a chart is usually faster to read. If they need to know which exact records, a table is the right tool, and a chart would just be a worse table.
Why is a pie chart usually the wrong choice for a dashboard?
A pie chart shows a split at a single point in time. It cannot show a trend, and it gets hard to read past five or six slices, because the eye cannot compare angles precisely. A bar chart usually shows the same comparison more clearly, and a line chart is almost always the better choice for anything changing over time.
Sources
Key Performance Indicator (KPI) visuals - Microsoft Learn
A great dashboard design comes down to three decisions, made after the audience and the layout are already settled. Which KPIs earn a place on the page, which visual fits each one, and how size, position and color guide the reader's eye to what matters first. Get those three right and the dashboard reads clearly before anyone touches a filter.
This post covers those three decisions. For defining the audience and the business question behind a dashboard, see how to create a dashboard. For page size, the layout grid and Power BI's formatting settings, see dashboard layout and formatting.
Choose the KPIs That Earn a Place
A dashboard gets crowded when every available number goes on it instead of only the ones that matter. Before adding a KPI, run it through three tests:
Is it actionable? If nobody would change what they do when the number moves, it does not belong on the dashboard. A metric that only satisfies curiosity is not a KPI.
Is it tied to a real business question? The metric should trace back to the question the audience needs answered, not to whatever field happened to be easy to pull from the source system. Our guide to KPI reports covers defining KPIs against a business objective in more depth.
Does someone own it? Every KPI needs a person responsible for it being correct and for acting on it. A KPI with no owner is the one that goes stale first, because nobody notices when it breaks.
A KPI that fails any of these three tests is a candidate to cut, not to shrink and tuck in a corner. A dashboard with eight KPIs that all pass these tests works better than one with twenty where half are just available data.
Match Each KPI to the Right Visual
Once the KPI list is set, each one needs a visual that shows what the reader actually needs to see: a single value, a change over time, or a comparison across categories. Four visual types cover most dashboard KPIs:
A card, for one headline number the reader needs to see instantly: total revenue this month, the current open ticket count, the current active user count.
A line chart, for a trend: the same metric measured repeatedly over time, where the shape of the change matters as much as the current value.
A bar chart, for a comparison: the same metric split across categories, such as revenue by region or tickets by agent.
A table, only when the reader needs the raw rows behind a number, not a summary of them. A table is not a fallback for "couldn't decide on a chart."
The measure behind a KPI usually does not change depending on which of these it feeds. A simple measure like:
Total Revenue = SUM ( Sales[Revenue] )works as a card on its own, as a line chart once you add a date field to the axis, or as a bar chart once you add a category field. What changes between the four visual types is the field well the measure goes into, not the DAX behind it. Our guide to measures tables covers organizing measures like this one as your model grows.
The exception is a visual built to show progress toward a goal. Power BI's own KPI visual needs three specific fields:
A base measure to evaluate.
A trend axis field, usually a date.
A target measure or value to compare against.
It then shows an icon and color for whether the metric is ahead of or behind that target.
Picking the wrong visual for a KPI makes the dashboard harder to read even when every number on it is correct. A pie chart shows a split at one point in time. It cannot show a trend, because it has no axis for time to run along. Putting twelve months of revenue into a pie chart forces the reader to compare wedge angles across two things at once, share and month, which the eye is bad at. The same data as a line chart shows the trend immediately.
If a chart makes the reader stop and think about how to read it, the chart, not the reader, is the problem. Our guide to data visualization best practices goes further into chart selection.
Use Size, Position and Color to Set the Order
Once the right KPIs sit on the right visuals, the last design decision is which one the reader sees first. Three tools set that order, and each should be used deliberately, not decoratively:
Size. The largest visual on the page draws the eye first. Reserve that space for the KPI the audience cares about most, not the one that happened to produce the most interesting chart or was easiest to build.
Position. Where a KPI sits on the page also sets reading order. For the mechanics of that, the grid, reading direction and Power BI's alignment tools, see dashboard layout and formatting.
Color. Use color as a signal, not a decoration. One accent color, applied only where something needs attention, such as a KPI below target, does more work than a different color on every chart. Keep everything else neutral so the accent actually stands out. This is different from picking a brand color theme, which dashboard layout and formatting covers; that is about consistency, this is about drawing attention to what matters.
A dashboard where every visual is the same size, in the same shade, competing for attention, forces the reader to do the prioritizing the designer should have done. Use size to say "look here first" and color to say "this one needs action," and leave everything else quiet.
FAQs
How many KPIs should a dashboard show?
There's no fixed number. What matters is that every KPI on the page passes the actionable, tied-to-a-question and owned-by-someone tests above. A dashboard with five KPIs that all pass is more useful than one with fifteen where most are just available data.
When should a table replace a chart?
Only when the reader's actual task is to look up specific rows, not to spot a pattern. If the audience needs to know what happened, a chart is usually faster to read. If they need to know which exact records, a table is the right tool, and a chart would just be a worse table.
Why is a pie chart usually the wrong choice for a dashboard?
A pie chart shows a split at a single point in time. It cannot show a trend, and it gets hard to read past five or six slices, because the eye cannot compare angles precisely. A bar chart usually shows the same comparison more clearly, and a line chart is almost always the better choice for anything changing over time.
Sources
Key Performance Indicator (KPI) visuals - Microsoft Learn
Show us the report nobody trusts.
1 · A 30-minute call.
2 · We look at your current reports together.
3 · We tell you what we would do.
Show us the report nobody trusts.
1 · A 30-minute call.
2 · We look at your current reports together.
3 · We tell you what we would do.
Show us the report nobody trusts.
1 · A 30-minute call.
2 · We look at your current reports together.
3 · We tell you what we would do.
CaseWhen is a Berlin BI consultancy that builds reporting that leaders can trust, on the Microsoft stack: Power BI, Fabric and Azure.
Berlin, Germany
Free tools
CaseWhen is a Berlin BI consultancy that builds reporting that leaders can trust, on the Microsoft stack: Power BI, Fabric and Azure.
Berlin, Germany
Free tools
CaseWhen is a Berlin BI consultancy that builds reporting that leaders can trust, on the Microsoft stack: Power BI, Fabric and Azure.
Berlin, Germany