Titles: Make the Point

The ladder

A label names the topic. An action title states the finding.

Consulting firms such as McKinsey call these action titles: a full sentence that states the "so what" of the page. Cole Nussbaumer Knaflic makes the same point in Storytelling with Data: make the takeaway explicit instead of leaving the audience to find it.

Here is the same chart with three titles. Switch between them and watch what the reader gets.

Bar chart titled Turnover by Department, all bars grey.

Poor: a label. It names the topic. Every bar is the same grey, so nothing tells the reader where to look.

Bar chart titled Production has the highest turnover, Production bar in blue.

Okay: a takeaway. It states the finding and the blue bar agrees. But "highest" by how much? The reader can't tell if this is a problem.

Bar chart titled Production loses 40% of its staff, three times the rate in Sales. Production in blue, Sales in light blue.

Excellent: an insight. A number (40%) and a comparison (three times Sales). Both bars the title names are coloured, so the chart proves the sentence.

Poor~~Turnover by Department~~Names the topic. Fine for a filing cabinet, not for a reader.

OkayProduction has the highest turnoverStates the finding, but with no scale the reader can't judge it.

ExcellentProduction loses 40% of its staff, three times the rate in SalesFinding, number and comparison. A reader who sees only this line still gets the message.

The formula

Subject, change, number, compared with what.

Who or what + what happened + the number + compared with what, and, if there's room, so what.

Part	In the excellent title	Ask yourself
Who or what	Production	What is the one thing this chart is about?
What happened	loses	Use a verb: rises, falls, leads, trails, loses, costs.
The number	40% of its staff	Give the size, with its unit.
Compared with what	three times the rate in Sales	A target, last year, the average, or a peer.
So what (optional)	"so it accounts for 83 of the 102 people who left"	What does it mean for the reader's decision?

Then check it:

Fifteen words or fewer, in sentence case, left-aligned, ideally on one line.
No field names, jargon or chart types. "Sum of Left by Department" and "Bar chart of turnover" both fail.
The five-second test: someone sees only the title for five seconds. Can they say the takeaway out loud?
Honest about cause. Write "is linked to" unless the data proves "causes".

Match the chart

The chart has to point at what the title says.

The title and the colour should pick out the same thing. Knaflic pairs the two: the words state the takeaway and the colour shows where it is. When they disagree, the reader's eye goes to the colour, and the title loses.

Title says Sales keeps its people best, but the Production bar is the one in blue. — The title is about Sales, but the blue bar is Production.
The eye lands on Production, the opposite story.

Same chart with the Sales bar in blue to match the title. — The one blue bar is the one the title names.
Same data, same sort: only the emphasis changed.

One chart can tell several stories. Whichever one the title tells, colour that part and leave the rest grey.

More examples

The same fix works on any chart.

Power BI writes a title for you from the field names. Every one of these defaults describes the data, not what it shows.

Column chart titled Count of Reason For Term by Reason For Term. — The title repeats a field name twice.
Truncated labels and a counted axis make the reader work.

Bar chart titled Another job or more money explains 30% of exits, with those two bars in blue. — The title adds the two reasons up for the reader: 30%.
Shares instead of counts, so the number in the title matches the bars.

Line chart titled Count of Term Year by Term Year, with gridlines. — "Count of Term Year by Term Year" says nothing.
The spike is visible, but the reader has to decide it matters.

Line chart titled Exits nearly doubled in 2015, to 26 from 14, with those two points labelled. — The title names the change and both numbers.
Only the two points the title mentions are labelled.
The subtitle flags that 2016 is a half year.

Dashboard titles

On a dashboard, a fixed sentence can turn false.

A presentation page shows one fixed view, so its title can state a fixed finding. A dashboard has slicers. The moment a reader filters, the numbers change, and a typed-in claim may no longer be true.

Below, the reader has filtered to employees rated "Fully Meets". The typed title still says Production leads with 40%. On screen, IT/IS leads at 41% and Production is 35%.

Dashboard visual filtered to Fully Meets. Title still says Production loses 40%, but IT/IS is now highest at 41% and Production shows 35%. — The title and the chart now contradict each other. The reader stops trusting both.

Same visual with a dynamic title: IT/IS has the highest turnover: 41% (Fully Meets employees). — The title is calculated, so it names whichever department leads, its rate, and the filter in effect.

Three ways to title a dashboard visual, from safest to strongest:

Type	Example	Use when
Question	Which departments lose the most people?	The visual is for exploring and the answer changes often.
Descriptive	Turnover by department, share who left	Plain words, never field names. The safe default.
Dynamic finding	IT/IS has the highest turnover: 41% (Fully Meets)	You can build it from a measure, so it stays true for every filter.

A dynamic title is a text measure. Power BI recalculates it every time a slicer changes:

Turnover Title =
VAR Top1 = TOPN ( 1, VALUES ( Employees[Department] ), [Turnover Rate] )
VAR Dept = MAXX ( Top1, Employees[Department] )
VAR Rate = MAXX ( Top1, [Turnover Rate] )
VAR Scope =
    IF ( ISFILTERED ( Employees[Performance Score] ),
         " (" & SELECTEDVALUE ( Employees[Performance Score], "selected scores" ) & " employees)" )
RETURN Dept & " has the highest turnover: " & FORMAT ( Rate, "0%" ) & Scope

Page vs. visual titles

The page states the message. Each visual supports it.

A report page has two kinds of title. Keep their jobs separate:

	Page title	Visual title
Job	The one message of the page	What this visual shows, or the question it answers
Where	A text box at the top left, same spot on every page	The visual's own title
Example	Production loses 40% of its staff, three times the rate in Sales	Share who left, by department
Size	20-24 pt, the largest text on the page	12-14 pt

Units and definitions go in a small grey subtitle under the title ("Share of employees who left, 2006-2016"), not in the title itself.

In Power BI

Four settings do all of this.

Write the title

Select the visual, open Format visual > General > Title.
Replace the default text with your sentence. Set Horizontal alignment to Left.
Turn on Subtitle for the unit and time frame, in a smaller grey font.

Colour the bar the title names

Create a measure: Bar Colour = IF ( SELECTEDVALUE ( Employees[Department] ) = "Production", "#2A78D6", "#C8C6C4" )
Open Format visual > Bars > Colors, click fx, choose Field value, and pick Bar Colour.

Make a dashboard title follow the slicers

Create a text measure like Turnover Title above.
In Format visual > General > Title, click fx next to Text, choose Field value, and pick the measure.

Add a page title

Insert > Text box, top left. Type the page's message.
Put it in the same position and size on every page so readers know where to look.

Test your eye

Rewrite these before you open the answer.

"Sum of Pay Rate by Department". What's wrong, and what would you write?

It's a field-name label: "Sum of" is Power BI's default aggregation, and summing hourly pay rates means nothing. Use the average and state the finding, for example: "Production pays $23 an hour on average, the lowest of any department".

A presentation title reads "Turnover is a problem in some departments". Poor, okay or excellent?

Poor to okay. It hints at a finding but names no department, no number and no comparison. Excellent: "Production loses 40% of its staff, three times the rate in Sales".

The title says "Referral hires stay longest" but every bar is a different colour. What's the fix?

Make the chart agree with the title. Grey out every bar, then colour only Employee Referral in the accent. Add the number: "Only 13% of referral hires left, vs. 34% overall".

Why shouldn't a dashboard visual have the title "Production loses 40% of its staff"?

Because slicers change the data. Filter to a performance score or a hiring source and Production may no longer lead, or the 40% changes. Use a question title, a descriptive title, or a dynamic title built from a measure.

Put the point in the title.