One Primary CTA per Page: The Rule People Keep Breaking

Everyone in web design knows the rule: a page should have one primary call-to-action. It’s in every guide. Nobody argues against it. And yet when I audit real websites, pages that follow it are the minority. This is interesting. When a rule is universally endorsed and routinely broken, the breakage usually isn’t ignorance. Something is generating the violations, and it’s worth figuring out what.

I got to see the pattern at scale building GazeSite, which grades websites on, among other things, whether a skeptical visitor would know what action to take. The failing pages aren’t chaotic. They’re usually well made. There’s a hero with “Start Free Trial.” Next to it, “Watch Demo.” A bar above the header announcing a webinar. A section inviting you to read the blog. A newsletter box. A chat widget blinking in the corner. Each element is reasonable. Together they form a committee of asks, each pulling in a different direction.

And a committee is exactly what produced them. A page with six competing CTAs is almost always the org chart rendered in HTML. Marketing needed the webinar promoted. Sales wanted “Book a Demo.” Content wanted newsletter signups because that’s their metric. Nobody made a decision so much as everyone made a request, and the page absorbed them all. That’s the generator. The rule gets broken not by designers who disagree with it, but by organizations where no single person is accountable for what the page is for. Which suggests the fix isn’t a design fix at all. The question “what is the one thing we want a first-time visitor to do” is a business decision, and pages with too many CTAs belong to businesses that haven’t made it. The clutter is honest, in a way. The page is confused because the company is.

Why does the confusion cost anything, though? The optimistic case for many CTAs sounds plausible: different visitors want different things, so offer many doors and let each person pick. The problem is that this model imagines a visitor who arrives with a formed intention and merely needs the matching door. Most visitors aren’t like that. Most arrive with a vague interest and a few seconds of patience, and their intention gets formed by the page. Show that visitor one clear action and the page is proposing an intention: this is what people do here. Show them six and you’re handing them a decision they never wanted, and the standard human response to an unwanted decision is deferral. I’ll look at this later. Later never comes.

There’s a subtler cost too. Multiple asks don’t just split clicks; they degrade the average value of the clicks you get. The visitor who might have started a trial gets diverted into the newsletter box because it asked for less. You “converted” them, in the sense that a metric went up, but you traded your most valuable possible outcome for your least, and the trade is invisible because the newsletter signup gets counted as a success.

The rule, stated carefully, is not one CTA per page. It’s one primary CTA per page. Secondary actions are fine, and often necessary — but they should be visibly subordinate. One button that looks like a button, filled, high-contrast, and everything else demoted to text links. The visual hierarchy does the arguing for you. A visitor can absorb “there’s one main thing here, and some side doors” instantly, without reading anything. What they can’t absorb is five buttons of equal weight, because equal weight is the page refusing to state a preference.

The hard version of this discipline is that a “page” is really a screenful. A long landing page can repeat its primary CTA many times — that’s fine, repetition of one ask is focus, not clutter. What it can’t do is change the ask as you scroll: trial at the top, demo in the middle, webinar at the bottom. Each screenful should answer “what does this site want from me” the same way.

If you want to apply this today, don’t start by redesigning. Start by listing every clickable ask on your homepage — every button, banner, popup, and widget — and then answer one question in writing: which single action, if a visitor took only it, would you choose? That sentence is the decision your page has been failing to make. Once it exists, the demotions are easy, almost mechanical. Everyone knew the rule all along. What was missing was someone willing to make the call the rule depends on.