You have seen them. A button that says "Get started" which links to the homepage. A "Contact us" link that opens a 404. A "Book a demo" CTA that submits a form to no one.
These are placeholder CTAs: the ghost infrastructure of a site that was built but not fully wired up. They look functional. They do not work.
Where placeholder CTAs come from
Most come from template-based or AI-assisted website builds. When a page builder or AI tool scaffolds a site, it creates the structure of a conversion funnel: hero CTA, pricing CTA, feature section CTA, footer CTA. The buttons exist. The links are set to # or to the homepage because the actual destination does not exist yet.
The site then gets shipped before those destinations get built. The button stays on the page. The destination never appears.
This shows up most often in three places. First: AI page builders like Framer templates and AI site generators, which include CTA buttons in every section by default. The placeholder href is usually # or /signup, and the /signup page frequently never gets created. Second: vibecoded sites where someone generated a full website from a prompt and the output included all the standard sections with all the standard buttons, none of them connected to an actual backend. Third: redesigns where a site gets a visual refresh but the CTA destinations are not updated to match the new URL structure.
Why placeholder CTAs are expensive
A visitor who clicks a CTA and lands on a broken page or loops back to a page they already visited does not try again. They leave.
The damage is larger than the missed conversion. A visitor who found the product interesting was ready to take the next step. The broken step tells them the company is not ready for them.
A broken CTA signals a broken company. Visitors form an opinion of the product based on the quality of the experience, and that opinion forms in seconds.
Common patterns to look for
The self-linking hero CTA. The most prominent button on the page links back to the homepage or to #. This kills conversions on your highest-traffic page.
The pricing-to-contact loop. The pricing page has no way to buy directly. It says "Contact us." The contact page submits a form that goes nowhere or sends an email to an inbox no one checks.
The blog CTA that never changes. Every blog post has a call to action at the bottom set up once and never updated. It promotes a feature that was renamed or a plan that no longer exists.
The mobile menu link that 404s. The desktop nav works. The mobile nav has a slightly different set of links, and one of them was never updated to match the new URL structure.
The "coming soon" page behind a live CTA. The button says "Try for free." The destination says "We're working on something exciting. Sign up to be notified." The product launched six months ago.
How to find them
Walk through your full site as a new visitor. Click every button and every link. Follow the path a customer would follow to buy or sign up. For a standard marketing site this takes about ten minutes. Write down every dead end.
Then look at your source for links pointing to #, /, the current page, or to pages that return 404. These are almost always placeholder or broken links.
Also test on mobile. Some CTAs that work on desktop are obscured or untappable on mobile. The button exists, but the tap target is too small or overlapped by another element.
SiteBlob checks for placeholder CTAs as part of its full site scan. It checks href values, flags buttons that link to the same page, and identifies forms with no valid action endpoint, with the exact HTML location of each issue.
Fixing them
Fix is straightforward once you have the list. For each broken CTA: decide what it should actually do. If you do not know, remove the button. A missing button is less damaging than a broken one.
Build the destination if it does not exist. For a signup CTA, this means a real signup flow. For a demo CTA, this means a booking link or a form that actually sends notifications.
Update the href. Test it. Then run through the full audit again. It is easy to fix one CTA and miss three others.