The call-to-action mistakes costing you customers

The most expensive call-to-action mistake is a generic button like "Submit" that doesn't tell people what happens next. Replacing vague copy with specific, first-person wording has lifted clicks by 60% or more in tests, and "Submit" performed worst in the studies cited here (CDMG, FoundryCRO). You pay for the same traffic either way. With a weak button, fewer of those visitors act.

Below are the mistakes we see most often, roughly in order of what they cost, with the fixes the evidence supports.

What's the most expensive CTA mistake?

Weak button copy. The words on a button move conversion far more than its colour, shape or size.

The usual order of impact is copy, then placement, then contrast against the page, and only then the specific colour. Copy changes alone swing conversion by 10–30%, and colour changes don't come close (Kissmetrics).

What works:

• Start with a verb that names the gain. "Get", "Start", "Claim" and "See" beat "Submit", "Send" and "Order". Swapping a generic verb for an outcome verb lifted clicks by up to 95% in one test (MarketingExperiments).
• Say what the visitor gets. Changing "Get membership" to "Find your gym & get membership" made the offer more relevant and lifted conversion 213% (ContentVerve via Icopify).
• Write in the first person. "Start my free trial" beat "Start your free trial" by up to 90% on landing pages (ContentVerve via Icopify). There's an exception at the final payment step, where "your" can win: first-person copy reduced conversions by about 27% in one test (Unbounce).

Replacing "Submit" with copy that names the next step lifted clicks by 60% to 87.5% across several tests (CDMG).

The percentages vary by site, so they aren't a forecast for yours. The pattern across these tests is consistent: specific, first-person copy that names the benefit beats generic wording.

Are too many buttons hurting you?

Probably, if your page offers several competing actions.

Across 18,639 landing pages, conversion was 13.5% with one call to action, 11.9% with two and 10.5% with three or more (Unbounce). When Whirlpool cut an email from four CTAs to one, clickthrough rose 42% (MarketingSherpa).

Repeating one goal is fine. "Start free trial" in the hero, mid-page and footer is a single message, and repeating it on long pages adds 8–25% (Conversion Rate Experts). Competing goals are the problem: "Free trial" next to "Book a demo", "Download the guide" and "Contact sales" splits attention four ways.

Our rule is one primary action per page, repeated at natural points, with at most one lighter secondary option for visitors who aren't ready yet. We cap the page at around five buttons in total, with a clear visual hierarchy.

Where should the main button go?

It depends on how much convincing the offer needs. For simple, high-intent offers, put the primary button above the fold so people see it straight away. For complex offers or cold visitors, show value and proof before asking them to act. One CXL test found a 304% lift from moving the CTA below the fold on a complex-offer page (FoundryCRO).

Keep the button near trust signals in either case. CTAs positioned next to testimonials converted 18% faster than standalone buttons (Arahoster via the research).

On long pages and on mobile, a sticky bottom bar is worth adding. Across several A/B tests it lifted conversion by 3.7% to 10% (Zipify, Clean Commit). Show it only after the main button scrolls out of view, make the tap target at least 48px tall, and keep it inside the phone's safe area so it never sits under the home indicator.

Does button design matter at all?

Yes, but less than copy and placement, so fix those first. A few design choices have good evidence behind them:

• Use filled buttons for the primary action. Outline buttons underperform filled ones by about 2.8% on average across six tests (GoodUI). In one triple test, ghost buttons drew 20% fewer clicks and took a second longer to find (CXL). Keep ghost buttons for secondary actions.
• Make the button contrast with its surroundings. There's no universally best button colour. The button needs to be the most distinct element on the page. Top-quartile pages average a 7.2:1 contrast ratio against their background, versus 4.1:1 for median pages (roast.page). If your accent colour also appears in icons and dividers, the button disappears.
• Round the corners. Rounded buttons produced 17–55% higher click rates than sharp-edged ones in published experiments (Biswas et al., Journal of Consumer Research, 2024). The high end may be inflated, but the direction holds.

Do popups help or hurt?

Both, depending on how you use them.

Exit-intent popups convert about 3.9% of leaving visitors on average, across more than a billion displays (TryFlint), and up to 17% on abandoned carts (Claspo). Matching the offer to the page improves results further: URL-matched popups converted 152% better than generic ones (Pixelform).

They backfire when done badly. Roughly 62% of users link "too many popups" with brand annoyance, while 72% welcome well-timed, relevant offers (Popupsmart). Google penalises popups that block content as soon as a mobile visitor arrives from search (Omnisend). Don't fire a popup on page load. Delay it, cap how often it shows, and stop showing it to people who've already converted.

FAQ

What is the single worst CTA word? "Submit." It was the lowest-performing call-to-action word in the studies cited here. Replace it with what happens next, such as "Get my free audit" or "See pricing" (CDMG).

Does button colour matter? Less than people think. The order of impact is copy, then placement, then contrast, then colour. Copy changes can swing conversion 10–30%, and colour changes don't come close (Kissmetrics). Make sure the button contrasts strongly with everything around it.

How many calls to action should a page have? One primary goal, repeated. Conversion drops from 13.5% to 11.9% to 10.5% as competing CTAs go from one to two to three or more (Unbounce). Add one lighter secondary option at most.

Should I use a button or a text link? Buttons win on landing pages, pricing pages and forms. Text links suit blog content and download pages, where buttons can feel like ads or trigger spam fear (GoodUI, Copy Hackers).

Every page we build has one primary action, and its button says what happens next. To have us review the buttons on your site, book a call.