You are halfway through a long comparison page. The offer finally makes sense, but the action was at the top, several screens ago. A button that stays within reach could save a trip back. Now put the same button over the last field of a mobile form. Visibility has become the problem.
A call to action (CTA) is the button or link asking a visitor to take a next step. A persistent CTA stays available as the person moves through the page, often in a sticky or floating bar. Its value is convenience after the visitor has enough context. Its cost is a piece of the screen and a continuing request for attention. The decision is whether this action earns that space for this visitor, on this page, at this moment.
When the action has earned a place
Persistence helps most when the page is long enough to lose sight of a relevant action and the visitor may be ready at several points. A product comparison can keep “Compare plans” available while someone reads the differences. A detailed course page can retain “See course dates” after the reader has checked the syllabus. The label should tell people what will happen next, not merely ask them to “Act now”.
Before adding a bar, try the page without it. Is the existing action hard to return to? Does a table or long explanation naturally create a decision point halfway down? Would a clear inline link at that point solve the problem with less screen space? A persistent control should solve a real access problem, not compensate for a page that has forgotten to make its next step clear.
The purpose matters too. “See accessibility details” or “Review delivery options” may help a person decide; a permanent “Book a demo” on an early educational article may ask for a larger commitment than the reader came to make. Keeping a request visible does not make it timely.
A worked page decision
Imagine a software site with a long plan-comparison page, a signup form and a help article. The marketing team proposes the same floating “Start trial” button on all three. This is an illustrative scenario, not a Barrava customer result.
On the comparison page, visitors need to read plan limits and may decide before reaching the bottom. “Choose a plan” in a compact, persistent place could be useful if it takes them to the relevant plan choice. Test the mobile view with the comparison table open, at 200% zoom, and with the consent notice present. If the button covers a plan difference, the layout needs changing before launch.
On the signup form, the visitor is already acting. A second “Start trial” button adds no direction and may cover a field, error message or the phone's on-screen keyboard. Leave the form's own submit action in charge. On the help article, someone is trying to solve a problem. “Start trial” is the wrong companion to that task; a relevant help or contact route may be better, or no persistent CTA at all.
That leads to three different decisions from one campaign idea: a candidate on the comparison page, no duplicate on the form, and a different or absent action on the help page. If your current tool cannot apply those page-level distinctions, solve the targeting through a supported site implementation or choose a genuinely useful site-wide action. Do not assume a bar can target individual URLs because a future campaign plan mentions it.
Where persistent controls take more than they give
Start by looking at the bottom and edges of the page on a real phone. A narrow desktop bar can become a large mobile slice after labels wrap. Add the site's consent control, chat launcher, accessibility control, cookie-settings entry, feedback tab and phone keyboard. Then tap, scroll, zoom and use the Tab key. A screenshot taken before these states appear can miss the overlap that matters.
WCAG 2.2's focus-not-obscured rule says a keyboard-focused component must not be entirely hidden behind content the site created. Its explanation calls out sticky footers and banners as possible causes. The target-size rule sets a 24 by 24 CSS pixel minimum or sufficient spacing, with defined exceptions; a crowded row of tiny controls needs closer inspection. These are concrete reasons to test the whole page, not a claim that every sticky bar fails accessibility.
Promotional interruptions also deserve restraint. Google Search Central advises against dialogs and interstitials that obstruct page content and describes small banners as a less intrusive pattern. That guidance does not give a blanket search ranking guarantee to any bar. A nominally small banner can still be too much when several other layers occupy the same corner.
Look beyond physical overlap. A persistent request can also interrupt understanding. If it appears before the benefit is explained, repeats the same action as a visible page button, uses vague urgency, or draws attention away from a required choice, it is taking focus without adding help.
Decide with a four-part page check
Use this check on the actual page and device, rather than asking whether “sticky CTAs work” in general:
| Question | Evidence to inspect | A reason to change course |
|---|---|---|
| Relevance: Does the action match the visitor's job now? | Page intent, search/ad promise, support questions | A demo request on a help task |
| Timing: Has the page earned the ask? | Where value and proof become clear | The bar asks before the offer is understood |
| Space: Can every task still be completed? | Narrow mobile, zoom, keyboard focus, forms, consent | A covered field, heading, price or control |
| Result: Does it help a worthwhile outcome? | Completed actions and later quality, not clicks alone | More clicks with fewer useful completions |
One counterexample is worth keeping: a short page with a strong, visible inline action may not need a persistent one at all. Another is a service journey where keeping a help action visible is valuable even if it produces no sale. Business and visitor value need not be identical to be worth measuring.
Give the action an owner and a review date. Campaigns change; a bar that was relevant for a launch can become wallpaper six weeks later. At review, compare not only interactions but valid submissions, task completion or support outcomes. For the difference between a CTA click and a meaningful result, see our guide to measuring CTA interactions.
Barrava offers a practical layout choice here. Its current website toolbar can hold ordered modules, including up to two CTAs on Starter and other eligible plans. Teams can preview a docked bar, which reserves page space, or a floating bar, which overlays the page; they can also control module visibility by device. On a tight mobile layout, fewer active actions or a docked layout may be a better starting point than shrinking everything to fit. Those settings let you inspect the trade-off. They do not provide a universal conversion lift, and planned URL-specific campaign targeting is not a current promise.
The next time someone asks for a persistent CTA, ask them to show the page at the exact moment it would help—and at the moment it might get in the way. If the first case is strong and the second is solved, the control has earned its place.
