Splitlab home

Book a Demo

Splitlab home

Book a Demo

Splitlab home
The splitlab.io first screen in a desktop browser with six Book Demo buttons in front of it, in near-black, yellow, purple, blue, burnt orange and green, with the yellow one outlined as the winner

November 18, 2026

By

Bishal Katuwal

·

16 min

read

Best Colors for Conversion Rates: What the Evidence Says

November 18, 2026

By

Bishal Katuwal

·

16 min

read

Best Colors for Conversion Rates: What the Evidence Says

In the ever-evolving business environment, achieving a competitive edge demands innovative solutions beyond conventional methods. Organizations require smart systems that can optimize workflows, enhance customer interactions, and reveal new avenues for growth. Enter Biotix, the cutting-edge AI-powered CRM management tool.

Red doesn’t convert best, and neither does green. The right color depends on your audience: their age, eyesight and your brand. Here’s the famous red vs green test read properly, and how I pick and test button colors.

Bishal Katuwal

Bishal Katuwal

Splitlab Founder

The splitlab.io first screen in a desktop browser with six Book Demo buttons in front of it, in near-black, yellow, purple, blue, burnt orange and green, with the yellow one outlined as the winner
The splitlab.io first screen in a desktop browser with six Book Demo buttons in front of it, in near-black, yellow, purple, blue, burnt orange and green, with the yellow one outlined as the winner

There is no single best color for conversion rates. Red doesn’t win everywhere, and neither does green or orange. What converts is contrast, and the right color depends on who your audience is. An older audience needs stronger contrast than a young one, color-blind visitors see some pairs as the same color, and your brand sets what people expect. So pick a few colors that fit your audience and brand, check the contrast, then test them on your own traffic and judge them on revenue.

I get asked about button colors a lot, usually by someone who read that “red buttons convert 21% better” and wants to know if they should switch. This post goes through where that number comes from, what it actually showed, and what I do instead.

Key takeaways

  • No color wins on every page. The famous red vs green test measured clicks on one homepage, and its author said not to generalize it.

  • Choose for your audience. Age, eyesight and color blindness change what people can see. Older visitors need more contrast.

  • Contrast beats hue. A button converts when it stands out from the page around it, not because of the color it is.

  • Test three to six colors, not two. Red vs green is too narrow. Run them in a multivariate test with your headline and let the data pick.

  • One accent, one job. Use your accent color only for the main action. If everything is yellow, nothing is.

  • Make the text readable. WCAG 2.1 asks for 4.5:1 contrast for normal text and 3:1 for large text.

  • Color is a small lever. On small traffic, test the offer, headline and proof first.

Is there a best color for conversion rates?

No. Color affects conversions only through what it does on a specific page: whether the button is easy to find, whether its label is easy to read, and whether it fits the brand people expect when they land.

That’s why the same color can win on one page and lose on another. An orange button on a blue page stands out. The same orange button on a page full of orange headings, badges and banners disappears. The color didn’t change. The contrast did.

So the useful question isn’t “which color converts best?” It’s “does my main button stand out on this page, and can my audience read it?”

The red vs green button test, read properly

Most “best button color” articles trace back to one test. In August 2011, Josh Porter wrote it up on the HubSpot blog as “The Button Color A/B Test: Red Beats Green” (the original link now redirects, so that’s an archived copy). Here is what it actually says:

  • What was tested: the home page of Performable, a marketing software company. One version had a green button, the other a red one. Nothing else changed.

  • The traffic: “over 2,000 visits” over “a few days”. That’s roughly 1,000 per version.

  • The result: “21% more people clicked on the red button than on the green button.”

  • What was measured: clicks on the button. Not sign-ups, not customers, not revenue. One of the first comments under the post asked whether the extra clicks turned into more form completions. The post doesn’t say.

And the part that usually gets left out is the author’s own takeaway. He wrote that “we cannot generalize these results to all situations”, that they hold only “in this page design, on this site, with the audience that viewed it”, and that red may have won because it “happened to contrast nicely with the green in Performable’s color scheme”. Green was already the site’s brand color, so a green button blended in. His advice: “do not go out and blindly switch your green buttons to red without testing first.”

So the honest summary is this. On one green-branded homepage, a button that contrasted with the brand got more clicks over a few days. That’s a lesson about contrast, not about red.

Why “red converts best” is a myth

The test gets repeated as a rule because a rule is easier to share than a caveat. But there are three reasons not to copy it:

  • It measured the wrong thing for most of us. A click isn’t a sale. If you sell a $27 ebook or a $997 course, a color that pulls more curious clicks but the same number of buyers hasn’t helped you. I explain why I judge pages on revenue per visitor vs conversion rate.

  • The reason it won doesn’t travel. Red stood out against green. On a page that’s already red, a red button would do the opposite.

  • One small test isn’t a law. About 1,000 visitors per version, over a few days, on one page, is a starting point for your own test, not a result to apply everywhere.

What actually makes a button convert: contrast and hierarchy

People don’t read a landing page top to bottom. They scan it, and their eye goes to whatever looks different. Psychologists call this the isolation effect, or the von Restorff effect, after Hedwig von Restorff’s 1933 study: when several similar things are on show, the one that differs is the one people remember.

For a landing page, that means the main button should be the most different-looking thing on the screen. Here’s our own homepage with one change. On the left is the live page: a dark button on a light cream page. On the right, the same button in a pale beige that’s close to the background.


The same splitlab.io first screen in two desktop browsers. Left, the live site: a dark Book a Demo button on the light cream page, 13.3:1 contrast against the page. Right, tagged EXAMPLE: the same button in pale beige, only 1.15:1 against the page, so it blends in

Same headline, same guarantee, same page. On the right, your eye lands on the headline and then has nowhere obvious to go. The button is still there, it just doesn’t ask to be clicked.

This matters most on the first screen of a phone, where most paid traffic lands. If the button doesn’t stand out there, many people never act on the page at all.

Use one accent color, and only for the action

The simplest way to make a button stand out is to give it a color nothing else on the page uses. I call this the accent color. It goes on the main button, and on the same button wherever it repeats down the page. Nowhere else.

Why: the accent color becomes a signal. After one screen, visitors learn that yellow (or orange, or whatever you picked) means “this is the thing to click”. If you also use it for headings, badges, icons and banners, the signal is gone.

Here’s an example with a free webinar sign-up page, the kind I run for course launches:


Example free webinar sign-up page shown two ways, tagged EXAMPLE. Left, accent on the button: yellow is used only for the Save my seat button. Right, accent everywhere: the badge, headline highlight, bullets and date banner are yellow too, so the button is one yellow thing among many

Both pages use the same yellow on the same “Save my seat” button. On the left, it’s the only yellow thing, so the eye goes straight to it. On the right, the badge, the highlighted headline, the bullets and the date banner are yellow too, and the button is just one more yellow shape.

A few rules I follow:

  • Start from your brand color. Sometimes the best button is simply your brand color, because people recognize it from your ads. If your page is already full of it, though, a brand-colored button blends in. Then pick a color that contrasts with your brand, which is exactly what happened on the Performable page.

  • Keep secondary actions quiet. “Book a demo”, “watch the video” or “see the curriculum” can be a text link or an outlined button. They shouldn’t compete with the main action.

  • Don’t use the accent for warnings or errors. If your button is red, a red error message under a form field looks like part of the button, and a red button can look like a warning.

Make it readable: contrast you can measure

Standing out is half of it. The other half is whether people can read the button. This one you don’t have to guess, because there’s a standard.

The W3C’s WCAG 2.1 success criterion 1.4.3 (Level AA) asks for a contrast ratio of at least 4.5:1 for normal text and 3:1 for large text, which it defines as at least 18 point, or 14 point bold. A related criterion, 1.4.11 Non-text Contrast, asks for 3:1 between interface components and the colors next to them, where a visible shape is needed to identify the control.

You can check any pair of colors for free in WebAIM’s contrast checker. Paste the button color and the text color, and it tells you whether the pair passes. Here are some common button choices, worked out with the WCAG formula:


Contrast check of five buttons against WCAG 2.1 (4.5:1 for normal text, 3:1 for large text): splitlab.io dark button with white text 14.3:1 pass; dark text on yellow 11.1:1 pass; white on red 3.8:1 large text only; white on green 2.1:1 fail; white on yellow 1.6:1 fail

The surprise for most people is white text on bright colors. White on a bright yellow comes out at about 1.6:1, and white on a bright green at about 2.1:1. Both fail. Those buttons look fine on a designer’s large monitor and turn into a blur on a phone in sunlight. Dark text on the same yellow passes easily.

Our own brand yellow is a good example of why both checks matter. It works as a highlight, but next to our cream background the contrast is only about 1.5:1. So the main button on splitlab.io is near-black with white text, not yellow.

Choose colors for your audience

This is the part most color articles skip. Who is looking at the page changes which colors work.

  • Age. Eyes change with age. Older visitors usually need stronger contrast and larger text to read a button comfortably, and soft pastel buttons that look elegant to a 25-year-old designer can be hard to see for a 65-year-old buyer. If you sell to seniors, or to their adult children, go for bold, high-contrast buttons.

  • Color blindness. About 1 in 12 men can’t tell some colors apart. More on that below.

  • Your niche. A fitness audience, a finance audience and a crafts audience are used to very different palettes. Look at the ads and pages your audience already trusts.

  • Your brand. People clicked through from your ads, emails or videos. A button in your brand color feels familiar, as long as it still stands out on the page.

So instead of asking “red or green?”, write down who your buyer is, then pick three to six colors that fit them. Not just red and green: yellow, orange, purple, teal, dark navy, black. Those are the colors you test.


Example palettes, tagged EXAMPLE, for three audiences. Younger buyers (fitness, creators, gaming): purple, teal, rose, yellow, near-black. Middle-aged buyers (business, coaching, finance): blue, burnt orange, green, gold, near-black. Older buyers: bold colors with strong contrast and big labels, dark red, navy, dark green, bright yellow, black. Pick 3 to 6 that pass the contrast check, then test them with your headline in one multivariate test

Color meaning depends on context

You’ll find charts that say blue means trust, green means growth and orange means urgency. Some of that is real association, but the meaning of a color depends heavily on where people see it:

  • On a form, red often means an error.

  • In a store, red often means a sale.

  • On a trading or crypto page, red means the price went down, and green means it went up.

  • In your own funnel, a color means whatever your brand has trained people to expect.

So I don’t pick a color for what it “means”. I pick it for whether it stands out on this page and fits the brand people clicked through from. If your Instagram posts and Meta ads are all deep purple and the page they land on is orange, the bigger problem is that it doesn’t feel like the same brand.

Don’t rely on color alone: color-blind visitors

Color vision deficiency is common. The US National Eye Institute says about 1 in 12 men have it, and that the most common type makes it hard to tell red and green apart. Colour Blind Awareness puts it at about 1 in 12 men (8%) and 1 in 200 women.

Think about what that means for a red vs green button test. For a meaningful share of your male visitors, the two versions may look much more alike than they do to you. What they do see clearly is the difference between light and dark. That’s one more reason contrast matters more than hue.

Two practical rules:

  • Never use color as the only signal. WCAG 1.4.1 says color shouldn’t be “the only visual means” of “indicating an action”. A button should look like a button because of its shape and label, not only its color. A form error should have a message, not just a red border.

  • Check in grayscale. Turn your phone’s display to grayscale (most phones have it under accessibility settings) and look at the page. If the button still stands out, it works for color-blind visitors too.

How to choose a button color in five steps

This is the process I use on any funnel page, from a $27 ebook to a webinar registration page.

  1. Write down who your buyer is. Age, niche, the brand they know you from. Why: the best color for a 25-year-old gamer isn’t the best color for a 60-year-old retiree.

  2. Pick three to six colors that fit that audience. Include your brand color and a few that contrast with the page: background, headings, images, logos. Why: a button can only stand out from what’s around it, and you don’t know in advance which of your candidates will win.

  3. Check the text contrast of each. Put the button color and label color into WebAIM’s checker and aim for at least 4.5:1. Why: a button people can’t read on a phone in sunlight costs you clicks, and this check takes ten seconds. Drop any color that fails.

  4. Do the squint test. Step back from the screen and squint, or blur a screenshot, and see what’s left. Why: blurring removes the words, so you see only what the eye is drawn to first. If the button isn’t one of the first things you notice, it doesn’t stand out enough.

  5. Test them together, and judge on revenue per visitor. Run a multivariate test: your button colors against two or three headlines at the same time, with a multi-armed bandit split so the weak combinations get less traffic early. Why: a color that wins with one headline can lose with another, and testing them together finds the best combination in one go. A color that gets more clicks can still bring in fewer buyers, so revenue per visitor decides.

Here’s the squint test on our homepage. Both phones are blurred on purpose. On the live site, the dark button is still the clearest shape on the screen. On the pale version, it’s gone, and the headline is all you see.


The squint test on the splitlab.io phone first screen, blurred on purpose. On the live site the dark Book a Demo button is still the clearest shape. On an EXAMPLE pale version the button disappears and only the headline stands out

If you want a step-by-step on running the test itself, see my guide on how to A/B test landing pages.

Why button color tests rarely matter on small traffic

Here’s the uncomfortable part. Most button color changes, if they help at all, help a little. And small effects need a lot of traffic to detect.

A rough example, using the standard two-proportion formula (95% confidence, 80% power), the same math as in my split testing guide. Say your $27 ebook sales page converts at 5%, which is about where mine sit:

  • A small change, like a button color that lifts sales from 5% to 5.25%, needs roughly 122,000 visitors per version before you can trust the result.

  • A big change, like a new offer or headline that lifts sales from 5% to 6.5%, needs roughly 3,800 visitors per version.

Those are my own calculations, and the exact numbers depend on your traffic. The point holds anyway: if your page gets a few thousand visitors a month, a button color test on its own will most likely end in noise. That’s why I fold colors into a multivariate test with the headline instead of running them alone.

That’s why color is near the bottom of my list:


Bishal Katuwal's testing order, biggest lever first: 1 the offer (price, bonuses, guarantee), 2 the headline (the promise from the ad), 3 proof by the button, 4 form and checkout (fields, steps, order bump), 5 button color, last and only if it blends in
  • The offer. Price, bonuses, payment plans, the guarantee. Why: this is what people are deciding on. A better offer can move sales far more than any design change.

  • The headline. It should repeat the promise from the ad or post that brought people in. Why: people decide in seconds whether they’re in the right place.

  • Proof near the button. Student results, screenshots, reviews. Why: the moment of most doubt is just before the click.

  • The form and checkout. Fewer fields, fewer steps, a relevant order bump. Why: this is where people who already decided to buy drop out.

  • Button color. Last, and mainly if the squint test shows the button blends in. In that case, fixing it isn’t a test. It’s a repair.

For more on choosing what to test first, read my landing page split testing guide and my landing page optimization best practices.

Examples from info-product funnels

Here’s how this plays out on the kinds of pages I work on. These are typical situations, not results from a specific test:

  • A $27 ebook sales page. Say the brand is navy and white, and the buy button is navy too, so it looks like another heading. The fix: a yellow button with dark text, used for every “Get the book” button on the page and nothing else. Why: the button now differs from everything around it, and dark text on yellow passes the contrast check easily.

  • A free webinar registration page. Say the designer used the brand orange for the “Free training” badge, the date banner, the bullet icons and the button. The fix: make the badge, banner and icons gray or dark, and keep orange for “Save my seat” only. Why: one accent, one job.

  • A $997 course checkout. Say the “Complete my order” button is white text on a bright green, which fails the contrast check at around 2:1. The fix: a darker green, or dark text. Why: at the moment someone is about to pay, the button label should be the easiest thing on the page to read.

  • An order bump. Say the bump box is highlighted in the same color as the main button, so it competes with the buy button for attention. The fix: a dashed border and a light tint for the bump, the accent color only for the buy button. Why: two things in the accent color split attention at the most expensive moment in the funnel.

None of these is a rule to copy. Each one comes from the same check: does the main action stand out, and can everyone read it? If you’re auditing a whole page, my landing page CRO checklist includes button contrast as one of its 40 checks, and how to increase landing page conversion rate shows where pages lose people first.

How Splitlab helps

Splitlab runs A/B, multivariate and multi-armed bandit tests on funnel pages, including pages built in ClickFunnels and HighLevel. If you do test a button color, it shows revenue per visitor next to conversion rate for each version, with every order, upsell and order bump credited to the version the visitor saw. So you find out whether the new color made money, not just whether it got clicked. See the features or book a demo.

FAQ

What is the best button color for conversions?

There isn’t one. The best button color is one that contrasts with your page and brand, is used only for the main action, and has readable text (at least 4.5:1 contrast for normal-size text under WCAG 2.1). On a blue page that might be orange. On a yellow page it might be near-black.

Do red buttons really convert better than green?

Not as a rule. The well-known test found 21% more clicks on a red button than a green one, on one homepage whose brand color was green, with over 2,000 visits. It measured clicks, not sales, and its author said the result may have come from red contrasting with the green brand and shouldn’t be generalized.

Does color psychology matter on a landing page?

A little, but context matters more. Red can mean error, sale or a falling price depending on where people see it. I choose a button color for contrast and brand fit first, and leave “what the color means” as a minor tie-breaker.

Should I A/B test my button color?

Yes, but not on its own and not just two colors. Pick three to six colors that fit your audience and brand, and test them in a multivariate test with your headline, so you find the best combination. On its own, a small lift from a color change can take over 100,000 visitors per version to detect. If your button fails the squint test or the contrast check, just fix it first.

Do older audiences need different button colors?

Usually they need stronger contrast and larger text rather than a specific hue. Soft or pastel buttons that look fine to a young designer can be hard to read for older visitors, so check contrast carefully and test bold options.

How do I make buttons accessible for color-blind visitors?

Use strong light-dark contrast, give the button a clear shape and label, and never use color as the only signal. Then check the page in your phone’s grayscale mode. If the button still stands out, it works for color-blind visitors too.

My final thoughts

I understand why the red button story spreads. It’s simple, it’s cheap to try, and it feels like a shortcut. But the test behind it measured clicks on one green homepage, and its own author told people not to copy it.

What I’d do instead: think about who your buyer is, pick a few colors that fit them, check the text contrast and squint at the page on your phone. Then test those colors together with your headline and let revenue per visitor pick the winner.

If your button fails the squint test, don’t A/B test it. Fix it first. When you’re ready to test, book a demo and we’ll set up the multivariate test with you.

The best color is the one your audience can see, and your data proves.

Built for revenue

Test for revenue, not just clicks.

Run A/B, multivariate, and multi-armed bandit experiments — with real revenue tied to every variation.

We’ll increase your conversion rate by 30% or you don’t pay. Terms apply.

✓

Revenue per variation

✓

A/B & MVT

✓

Multi-armed bandit

Test for revenue, not just clicks.

A/B, multivariate and multi-armed bandit experiments, with real revenue tied to every variation.


Start optimizing for revenue today not just conversion rates.

Run experiments, track revenue per variation, and scale what actually makes you money.

We’ll increase your conversion rate by 30% or you don’t pay. Terms apply.


Start optimizing for revenue today not just conversion rates.

Run experiments, track revenue per variation, and scale what actually makes you money.

We’ll increase your conversion rate by 30% or you don’t pay. Terms apply.


Start optimizing for revenue today not just conversion rates.

Run experiments, track revenue per variation, and scale what actually makes you money.

We’ll increase your conversion rate by 30% or you don’t pay. Terms apply.