The add-to-cart button is the single most important click on your product page, every sale passes through it. That makes it a tempting first A/B test, and also the subject of more myths than any other element on your store. This article covers what’s genuinely worth testing on your CTA (call to action, the button you want people to click), what’s mostly folklore, and how to run the tests properly.

Why the add-to-cart button deserves a proper test

Most product page elements influence the decision to buy. The CTA is different: it’s the mechanism of buying. If people who want the product can’t find the button, don’t trust it, or hesitate over what it will do, you lose sales you’d already earned.

That said, keep expectations realistic. A button change is rarely going to double your add-to-cart rate. Button tests tend to produce small-to-moderate effects, which means they need decent traffic to detect. If your store is on the smaller side, check the feasibility numbers before you commit weeks to a button colour experiment, you may be better off testing bigger swings first.

The right metric for these tests is add-to-cart rate: the percentage of product page visitors who click the button. It’s the closest thing to a direct measurement of what you changed. There’s a longer discussion of why add-to-cart rate beats conversion rate as a product page metric, but the short version is: the further downstream your metric, the more noise drowns out your signal.

Button copy: “Add to cart” vs the alternatives

Copy is the most interesting thing to test on a CTA, because it changes what the click means to the visitor.

The standard: “Add to cart”

“Add to cart” (or “Add to basket” in the UK) is the default for a reason. It’s unambiguous, it matches what every other store says, and it makes a low-commitment promise: you’re adding something to a cart, not buying it yet. Familiarity is a feature here, not a weakness.

Alternatives worth testing

  • “Add to bag”, common in fashion and beauty. It’s a tone signal more than a functional change. Worth testing if your brand leans premium.
  • Ownership language, “Get yours”, “Claim yours”, “Make it mine”. These raise the emotional temperature. They can work for impulse-friendly, personality-led brands and fall flat (or feel pushy) for considered purchases.
  • Benefit-led copy, “Start sleeping better”, “Try it for 30 days”. This pairs the click with the outcome or the guarantee. It tends to work best when the guarantee is genuinely strong and the product is unfamiliar.
  • Urgency copy, “Get it before it’s gone”. Only test this if it’s true. Fake urgency erodes trust and, in some jurisdictions, invites regulatory attention.

Copy to be cautious with

Avoid anything that makes the click sound like a bigger commitment than it is. “Buy now” as the primary button label can suppress clicks from people who wanted to keep browsing, they’re not ready to “buy”, but they were ready to “add”. If you want a buy-now path, that’s what secondary CTAs are for (more below).

A good hypothesis format: “Because our audience buys on impulse, changing ‘Add to cart’ to ‘Get yours’ will increase add-to-cart rate.” If you can’t articulate the because, you’re guessing, which is fine occasionally, but guesses should be your lowest-priority tests.

Size, colour and contrast: myths vs reality

You’ve probably read that a red button beat a green one somewhere and lifted conversions by some suspiciously round percentage. Here’s the honest version.

The myth

There is no universally best button colour. Results from other people’s stores don’t transfer, because the effect of a colour depends entirely on the page around it. A red button on a white minimalist page is loud; a red button on a red-accented theme disappears.

The reality: contrast and visibility

What reliably matters is whether the button is visually findable:

  • Contrast with surroundings. The button should be the most visually distinct clickable element on the page. If your theme uses the same accent colour for the button, the sale badge and the review stars, the button isn’t distinct.
  • Size and tap target. Small buttons get missed, especially on mobile. A full-width or near-full-width button on mobile is a sensible default to test against a narrow one.
  • Looking clickable. Ghost buttons (outline only, no fill) photograph beautifully in design mockups and routinely underperform solid buttons because they read as less interactive. That’s a testable hypothesis, not a law, but it’s a better-motivated test than red vs green.

So: don’t test colour for colour’s sake. Test low contrast vs high contrast, or ghost vs solid, or standard vs large. Those tests have a mechanism behind them.

Placement: above the fold, and what “the fold” means now

The classic advice is to keep the add-to-cart button “above the fold”, visible without scrolling. On desktop that’s usually easy. On mobile, where most Shopify traffic lives, a large image gallery plus title plus price plus variant picker often pushes the button off the first screen.

Placement tests worth running:

  • Tightening the layout above the button, shorter title area, smaller gallery, variant picker moved, so the CTA appears on the first screen on a typical phone.
  • Button before vs after the variant selector. Convention says the button follows the choices (size, colour), and breaking that convention usually causes confusion. But if your product has no meaningful variants, test removing or collapsing the selector entirely.
  • A second button further down the page for long, content-heavy product pages, so a reader who’s just been convinced by your description doesn’t have to scroll back up.

On Shopify, layout changes like these are template changes. The clean way to test them is to build the alternate layout as a second product template and split traffic between the two, the approach described in how to A/B test Shopify product pages. This is exactly the kind of test Atchoo! runs: you create the alternate template in your existing theme (e.g. product.cta-b), and Atchoo assigns each visitor deterministically to one version and tracks add-to-cart rate for each, with Bayesian analysis to call the winner.

Sticky mobile CTAs: the highest-probability button test

If you only run one CTA test this quarter, make it this one: a sticky add-to-cart bar on mobile, a slim bar pinned to the bottom of the screen with the price and an add-to-cart button, visible no matter how far the visitor scrolls.

Why it’s a strong hypothesis:

  • Mobile product pages are long. Galleries, accordions, reviews, the visitor can easily be four screens away from the button at the moment they decide to buy.
  • The bottom of the screen is where thumbs naturally rest, which is the whole argument of mobile product page optimization.

Things to get right in the variant: keep the bar slim (it’s stealing screen space from your content), make sure it doesn’t cover your footer or cookie banner, and if the product has required variants, decide what tapping the bar does before a variant is chosen, scroll to the selector, or open a picker. A sticky bar that adds the wrong variant to the cart will hurt more than it helps.

This is still a test, not a guaranteed win. Some themes ship with sticky bars that are so intrusive they get resented. Measure it.

Quantity selectors: helpful or friction?

The humble quantity stepper next to the button is worth questioning:

  • For most single-purchase products (a mattress, a jacket), the quantity selector is clutter. Testing its removal simplifies the buy area and removes one decision.
  • For consumables and multi-buy products (coffee, socks, supplements), the opposite test applies: make quantity more prominent, or test preset bundles (“1 / 3 / 6 bags”) instead of a bare stepper.

The mechanism in both cases is the same: every control near the CTA is a question you’re asking the visitor. Only ask questions that earn their keep.

Secondary CTAs: buy-now buttons and their trade-offs

Shopify supports dynamic checkout buttons, “Buy it now”, or accelerated wallets like Shop Pay, Apple Pay and Google Pay, that skip the cart and jump toward checkout.

The trade-offs are real in both directions:

For: genuinely faster path for ready buyers; wallet buttons remove address and card entry, which is a big deal on mobile.

Against: a second button competes with the first, splitting attention and adding visual noise; buyers who go straight to checkout skip your cart and any cross-sells or free-shipping-threshold nudges that live there; and two side-by-side buttons force a “which one?” micro-decision.

Sensible tests: primary CTA alone vs primary-plus-buy-now; and visual hierarchy tweaks, the add-to-cart button styled boldly with the buy-now button quieter beneath it, versus equal weight.

One measurement caveat: if you’re using add-to-cart clicks as your metric, decide upfront how buy-now clicks count. A variant that shifts clicks from add-to-cart to buy-now hasn’t necessarily lost purchase intent, it’s rerouted it. Define your success metric before launch, not after you’ve peeked at results.

Running button tests properly

A quick checklist:

  1. One change per test. If the variant has new copy and a new colour and a sticky bar, you’ll never know which element did the work.
  2. Write the hypothesis down with the mechanism (“because…”).
  3. Check feasibility first. Small effects need large samples. Use a sample size calculator with your current traffic and add-to-cart rate before you start, and see sample size and duration for how long to run.
  4. Run whole weeks. Weekend shoppers behave differently from Tuesday-lunchtime shoppers.
  5. Segment by device when reading results. A sticky bar test can win big on mobile and do nothing on desktop; the blended number hides that.

For the full methodology, priors, peeking, when to call a winner, see the complete guide to Shopify A/B testing.

Ideas ranked by expected value

If you’re prioritising, a reasonable order for most stores:

  1. Sticky mobile add-to-cart bar
  2. Getting the CTA onto the first mobile screen
  3. Solid vs ghost / low-contrast vs high-contrast button styling
  4. Primary CTA alone vs with buy-now
  5. Copy variants (“Add to cart” vs alternatives)
  6. Quantity selector removal or bundling
  7. Pure colour swaps (last, if ever)

Because Atchoo tests run as template splits inside your existing theme, each of these is a template variant away, no theme duplication, no code changes to your live template. The Pro plan includes full A/B testing with a 14-day free trial if you want to put your button beliefs on trial.

Frequently asked questions

Do button colour changes ever win?

Sometimes, usually when the original button had poor contrast with the page around it. The win comes from visibility, not from the psychology of a particular hue. If your button already stands out clearly, a colour swap is a low-probability test.

How long should a CTA button test run?

At minimum one full week, and usually two to four, depending on traffic. The honest answer comes from a sample size calculation using your visitor numbers and baseline add-to-cart rate, see our sample size and duration guide.

Should “Buy now” replace “Add to cart” as my main button?

For most stores, no, it raises the perceived commitment of the click. The better-motivated test is offering buy-now (or a wallet button) as a clearly secondary option beneath the primary add-to-cart button.

Can I test a sticky add-to-cart bar without editing my theme code?

Many Shopify themes include a sticky bar as a template or section setting you can toggle. If yours does, create a second product template with the bar enabled and split traffic between the two templates, no code required.