Convention is a feature
Two decades of the web have taught visitors that chat lives in the bottom-right corner. A launcher in a creative location is not distinctive — it is invisible, because nobody looks for it there. Spend your creativity on the words and the timing; keep the location boring.
Words beat icons
A bare bubble icon tests worse than a labeled button, because a label answers the visitor's actual question: is this a human, a bot, or a mailto? "Chat with us" sets the expectation; "Questions? We're online" adds state; "Leave a message" honestly reframes the offline hours. Match the label to reality — a button that says "we're online" over an unstaffed queue burns exactly the trust chat exists to build.
Mobile: the hard mode
- Stay off the thumb path. Bottom-right on mobile collides with floating action buttons, cookie bars, and the visitor's own scroll thumb. Keep the launcher small (a circle, not a bar) and respect the safe-area insets on notched phones.
- Never cover the primary action. A launcher squatting on the Add-to-Cart or Submit button costs more revenue than chat recovers. Audit your key templates at 375px width — the launcher must not overlap the one thing each page exists for.
- Expanded = full attention. When the panel opens on a phone, half-measures fail: a desktop-width panel clipped by the viewport is unusable. Full-screen (or viewport-clamped) expanded chat is correct on mobile even though it feels heavy on desktop.
Share the corner politely
Your launcher is rarely alone down there: cookie consents, back-to-top buttons, and accessibility overlay tools all claim the same corner. The widget should detect neighbors and give ground rather than stacking on top or — worse — fighting them for pixels. (MyLiveChat's widget yields to accessibility-tool repositioning for exactly this reason.) If your site runs several floaters, assign the corner deliberately instead of letting five scripts negotiate by force.
Design: on-brand, but legible first
- Brand color on the launcher is right — as long as label contrast stays readable (4.5:1). Legibility outranks palette.
- One attention animation on first load is plenty. A launcher that bounces forever trains visitors to ignore it — or to close it on sight.
- The unread badge is your strongest attention tool: quiet until there is genuinely something new, unmistakable when there is.
The five-minute audit
Open your three most important pages on a phone: can you read the label, tap the launcher with a thumb, complete the page's primary action with chat closed, and use the full chat when open? Then load your slowest page: does the launcher's arrival shove content around? Fix what failed, in that order. In MyLiveChat, position, offsets, size, and label are per-site settings under Deployment — test variations without touching code.
A launcher in the bottom-right corner is sitting on top of whatever your site already put there, and on
a surprising number of sites that is something important: a cookie banner, a sticky add-to-basket button, a
back-to-top control, a promotional bar, or the last row of a long form.
Audit the overlap on real pages rather than the homepage. Checkout, the pricing table and any page with a
sticky footer are where collisions are both most likely and most expensive, because the thing being covered
is usually the action you want the visitor to take.
Cookie banners deserve a specific check. They typically appear on the first visit, occupy the bottom of
the viewport, and are exactly the moment a new visitor is deciding whether your site is competent. If your
widget and your banner fight, the visitor's first impression is of a site that does not work.
Placement is a per-page decision
Treating the widget as one global setting is the most common missed opportunity. The pages where a
question blocks a purchase — pricing, checkout, delivery, product detail — justify a more
prominent, more inviting presence than a blog archive or a careers page.
The reverse is also worth doing. Pages where chat cannot help, or where an interruption is actively
unwelcome, are candidates for a quieter widget or none at all: legal pages, an account area during a
sensitive task, a form someone is midway through completing.
Where you vary it, keep the launcher recognisably the same object in the same corner. Varying prominence
is fine; moving the control around the screen between pages makes it feel unreliable and costs you the
convention you were relying on.
What to measure
Compare the open rate of the widget across your top pages rather than sitewide. A page with heavy traffic
and a near-zero open rate usually has a placement or visibility problem rather than an audience with no
questions.
Check the mobile numbers separately, always. Mobile is where the widget competes hardest for space and
where a badly placed launcher does the most damage, and a healthy blended figure routinely conceals a poor
mobile one.
Watch for visitors opening the widget and closing it immediately without typing. That pattern often means
they opened it by accident, which is a placement finding — usually a launcher sitting where a thumb
naturally rests or where another control used to be.