The pillar settled that colour on a dashboard should mean something. Give each colour one job, grey out everything that isn’t the point, check it survives colour vision deficiency — that’s the argument, and I’m going to assume it here.
This is the practice: which colours, to what standard, and exactly where in Data Studio they go. I’ll walk through every panel involved, because the shape of a Data Studio palette is more constraining than people expect — it isn’t a list of nice colours, it’s an assignment to named slots, and most of the slots aren’t chart colours at all.
Four ways to colour a chart in Data Studio, and what each one tells the reader
Before any of the mechanics, the decision that actually matters. A colour scheme is a claim about the data, and four schemes on the same six numbers produce four different readings.
Every series its own colour. The default, and the right answer for an all-channels breakdown where the reader picks their own row. It says: six categories, equally important, help yourself. In Data Studio this is just the palette doing its job — you set the first four or five slots and leave it alone.
One accent, the rest grey. The scheme for the chart you’re actually going to write a paragraph about. Set one palette slot to your accent, set the others to a light neutral, and the eye goes exactly where the commentary goes. In a monthly marketing report this is the single highest-value colour move available, and it takes about ninety seconds.
Semantic green and red. Reserve these for direction — change against last period, target hit or missed. The moment you use green and red for categories, you’ve told the reader that email is good and paid is bad, which is a judgement you didn’t mean to make and can’t take back. Data Studio has a dedicated setting for this pair, covered below, and using it keeps the semantic colours out of your category palette.
A single hue, light to dark. The scheme for ranked things: a top-twenty landing pages table shaded by sessions, a heatmap of hour against day. It says these are ordered, and the dark end is bigger. Because it varies in lightness rather than hue, it’s also the most robust scheme on this page for readers with colour vision deficiency.
The rest of this guide is how to build each of those in the product.
Where a Data Studio palette actually lives
Open Theme and layout → THEME → Customise and you’re editing a structured object: a set of named style properties plus one ordered array of series colours. The array holds twenty slots, not the three or four that most branding guides imply, and Data Studio fills them in order.
The slots people forget are the ones that make a report look half-finished: accent styles (table headers, filter controls, buttons, links), the “Others” colour that appears on every chart with a row limit, the positive and negative change colours on scorecards, and the page navigation colours.
Setting the twenty chart palette slots
Scroll the Edit theme panel to Data styles and the palette is the swatch grid — four rows of five. Click any swatch to open a picker that takes a hex code directly, so you can paste from a brand guideline without converting anything.
Work with the report visible, as in that shot, rather than on a blank page. The charts repaint live, so you can see whether slot two is distinguishable from slot three at actual bar width — which is a different question from whether two hex codes look different side by side in a brand deck.
Directly above the swatches sits the setting that decides whether any of this holds.
Colour by: Series order or Dimension values
Series order means slot one colours whatever happens to be first. Dimension values pins a colour to a value. The difference is invisible until somebody uses a filter.
So under Series order, Organic is blue this month and green next month, when Paid overtakes it. Under Dimension values, Organic is the same colour on every chart, in every month, in every report you apply the theme to. For anything a client can filter — which is most client reporting — that’s the one you want.
The same trap runs in the other direction. If you’re greying out “series 2, 3 and 4” to emphasise series 1, you’re greying positions, not things. The moment the data reorders, you’ve greyed the wrong line and highlighted something you didn’t mean to.
How to pin a colour to a dimension value
Under Colour by, click Manage dimension value colours. That opens a report-wide map — not a chart-level setting — listing every value you’ve pinned and offering Add a value at the bottom.
Two things that shot answers which the documentation doesn’t emphasise. The map is report-wide — pin Organic once and every chart in the report obeys it. And it caps at 1,500 values, which is irrelevant for channels and very relevant if you were planning to pin a colour per landing page.
Where this earns its place in a marketing report: a channel mix that appears on the summary page, the acquisition page and the appendix. Pin the six channels once, and a client who filters to two of them on page three still sees the colours they learned on page one.
What setting the theme buys you afterwards
Here’s the payoff that makes the half-hour worth it, and I’ve not seen it mentioned anywhere. Once a theme is set, every colour picker in the editor carries a THEME row — your palette, offered as swatches, everywhere a colour can be chosen.
That’s the mechanism that keeps a report consistent under maintenance. The person who adds a chart in three months doesn’t need your brand guide open — the right colours are the nearest ones in the picker. Without a theme, they’ll pick from the generic grid, and that’s where the eleventh colour in a six-colour report comes from.
Setting the positive and negative change colours
Scroll further down Edit theme and there’s a group called Positive and negative change style, holding exactly two colours.
These drive every up-arrow and down-arrow on every scorecard. Setting them deliberately does two jobs: it stops your brand red from also being your bad-number red, which is a genuine problem for any client whose logo is red, and it keeps green and red out of the category palette, where they’d invent the judgement I warned about at the top.
If the client’s brand is red, this is the setting to reach for first. Push the negative colour to a distinctly different red — or to an amber — and the scorecards stop arguing with the header.
The four contrast ratios a Data Studio palette has to hit
Three of these are widely quoted in design writing. The one that actually governs your chart colours is almost never mentioned.
WCAG 2.2 — a W3C Recommendation since December 2024 — sets 4.5:1 for normal text and 3:1 for large text at AA, and 7:1 at AAA. Those are the familiar ones.
The one that matters here is 1.4.11 Non-text Contrast, which requires 3:1 for “graphical objects” — the parts of a graphic required to understand the content. That is your bars, your lines, your pie segments and your status icons. A pastel palette that looks tasteful against white can fail this comfortably, and it’s the criterion that the entire dashboard-colour genre skips.
And a unit mismatch nobody flags. Data Studio sets type in pixels. WCAG defines large text in points — 18pt, or 14pt bold. At the standard 96dpi that’s 24px, or about 20px bold. So the relief for “large text” starts much higher than people assume, and everything at Data Studio’s 12px default owes the full 4.5:1. The type guide works through what that means for chart titles and axis labels.
Who can’t read your palette, and by how much
The pillar makes the case for checking. Here are the numbers, properly sourced, because this is a field that repeats figures without them.
The US National Eye Institute puts colour vision deficiency at about one in twelve men, most commonly red-green. It gives no figure for women; Birch’s review of worldwide prevalence puts that at roughly 0.4%.
The part that matters more than either number, and which almost nothing on this topic says: prevalence varies by population. The same review gives about 8% for men of European Caucasian descent but 4–6.5% for men of Chinese and Japanese ethnicity. If your client’s leadership sits in three continents, “one in twelve” is a European figure being applied globally.
Palettes that already solve colour vision deficiency
Don’t design a colour-vision-safe palette from scratch. Use one that’s been tested.
The Okabe–Ito set, from Masataka Okabe and Kei Ito’s Color Universal Design work, is eight colours chosen to stay distinguishable under the common forms of colour vision deficiency. It’s the default choice in a lot of scientific publishing for exactly that reason, and it maps cleanly onto Data Studio’s first eight palette slots — paste the eight hex codes in order and you’re done.
For sequential and diverging scales — heatmaps, conditional formatting — ColorBrewer‘s colourblind-safe sets are the equivalent, and they carry the advantage of varying in lightness as well as hue. Lightness survives every form of colour vision deficiency, which is why a palette that varies on that axis is robust in a way that a palette varying only in hue never is. It’s also why the fourth scheme at the top of this page is the safest of the four.
Making a colour carry a rule: conditional formatting
Conditional formatting is where colour stops being decoration and starts being information — a cell that turns amber when a metric crosses a threshold is doing work no static palette can.
Select a table, open the Style tab, and click Conditional formatting → Add. The builder is a full-screen panel in three parts: which elements to format, the rule itself, and an Appearance row that decides what the colour lands on.
Entire row versus single cell is the choice that decides whether the table reads as a heatmap or as a flagging system. For a landing-pages table where one row needs attention, entire row and a light fill. For a table where three different columns each have their own threshold, single cell — otherwise the rules fight over the same row.
Worth knowing before you build: a chart takes up to 20 rules, each combining conditions; when two rules touch the same property, the last true one wins; and colour scales take between two and five points. Two of the preset colour scales are optimised for colour vision deficiency — they’re at the end of the dropdown, and they’re the ones to reach for.
One trap: conditional formatting compares the underlying value, not the displayed one. A cell showing “2.0%” that’s really 0.0195 won’t match a rule for “greater than 2%”.
Extract theme from image, and why it isn’t brand matching
Data Studio can build a palette from an image — point it at a logo and it returns candidate themes. It’s a real time-saver and it is not brand matching.
It samples the colours present in the image, which isn’t the same as the colours your brand specifies. A logo that’s mostly white on a coloured field gives you a palette built from the field. And nothing about the process considers contrast, so the result can be perfectly on-brand and fail 1.4.11 on every chart. The branded theme guide walks through the extraction dialog and what to fix afterwards.
Use it to skip typing four hex codes. Then check the result against the numbers above.
The order to set a Data Studio palette in
Putting the panels above into a sequence that avoids setting the same colour twice:
- Start from the closest preset rather than Default, so the properties you don’t set are already sensible
- Set Colour by — Dimension values for client reports
- Fill the first four palette slots deliberately; let the remaining sixteen be reasonable
- Set the “Others” colour, or every row-limited chart gets a grey wedge
- Set positive and negative change, so your brand red isn’t also your bad-number red
- Set accent styles — the table headers and filter controls that make it look properly branded
- Set page navigation colours last
- Pin the dimension values that appear on more than one page
Two behaviours to know once it’s built. Component-level styling beats the theme and survives you applying a new one, so a rebranded report often has stragglers — select the chart, Style tab, Reset to report theme. And that reset won’t clear custom colours set in the dimension value colour map, which are held separately.
What Google gives you, and what you still have to check
Three things worth crediting: the default theme is a reasonable starting point rather than an arbitrary one, two of the conditional-formatting colour scales are explicitly colour-vision-optimised, and image components support alt text.
What doesn’t exist: no alt text for charts, no contrast checker, and no accessibility documentation for reports. So the checking is yours to do — a browser extension that simulates colour vision deficiency, and any contrast checker, applied to a finished report before it goes out. Ten minutes, once per brand.
Frequently asked questions
How many colours are in a Data Studio chart palette?
Twenty, assigned in series order. Most branding guides talk about three or four because that’s how many you realistically need to choose deliberately — but the slots exist, and anything you don’t set stays at whatever the preset had.
How do I stop a channel changing colour between charts in Data Studio?
Set Colour by to Dimension values rather than Series order, then use Manage dimension value colours to pin each channel to a colour. The map is report-wide and holds up to 1,500 values, so a client filtering two channels out on page three still sees the colours they learned on page one.
What contrast ratio do chart colours need?
3:1, under WCAG 2.2’s 1.4.11 Non-text Contrast, which covers graphical objects — bars, lines, segments and icons. It’s a different requirement from the 4.5:1 for text, and it’s the one dashboard colour guides almost always miss.