The short version
- A modal popup buys the most attention and costs the most goodwill. Reserve it for something the visitor would be worse off missing.
- An announcement bar is the cheapest interruption and the one most often ignored. That is the right trade for standing information.
- A slide-in sits between the two: noticeable, dismissible, and it does not block anything.
- Google's intrusive interstitial guidance applies to MOBILE and to content shown immediately on arrival from search. A bar is essentially never affected; an immediate full-screen modal can be.
- Exit intent does not exist on touch devices. There is no pointer leaving the window, so any exit-intent popup silently becomes a timed one on more than half your traffic.
- Frequency capping matters more than the format. The same popup shown once is persuasive and shown on every page view is an ad blocker installation.
Every site eventually needs to say something that is not on the page the visitor asked for: a sale ends Friday, deliveries are delayed, there is a newsletter. The three standard formats each buy attention at a different price, and the cost is paid by a different part of the experience.
The choice is usually made by whatever the tool offers rather than by what the message deserves, which is why so many sites open a full-screen modal to mention free delivery. Matching the format to the message is most of the work.
What each format actually costs
A modal popup covers the page and usually blocks interaction until it is dismissed. It gets read, which is why it converts, and it also interrupts a task the visitor chose over the one you are proposing. That is a real cost, and it compounds on repeat views.
An announcement bar takes a strip at the top or bottom and never blocks anything. It is the format people scroll past, which is exactly right for information that should be available rather than acted upon — opening hours over a holiday, a delivery cutoff, a shipping delay.
A slide-in appears in a corner without covering the page. It is noticed at roughly the rate a bar is ignored, and it is dismissible without a decision. For most newsletter and discount offers it is the honest middle: visible enough to work, quiet enough not to be resented.
| The message | Format | Why |
|---|---|---|
| Site is down for maintenance tonight | Bar | Everyone should see it; nobody needs to act |
| Free delivery over a threshold | Bar | Standing information, relevant at checkout not on arrival |
| Newsletter or discount signup | Slide-in | Optional, and a modal is disproportionate to the ask |
| Sale ends in two hours | Popup | Time-critical and the visitor loses by missing it |
| Age or region confirmation | Popup | It genuinely has to block the page |
| Cookie or consent choice | Bar or popup | Depends on your jurisdiction, not on conversion |
The trigger matters more than the format
A popup shown the instant a page loads is a different product from the same popup shown after the visitor has read half of it. The first interrupts someone who has not yet decided whether your page is worth their time; the second interrupts someone who has already decided it is.
Scroll depth is the most reliable proxy for interest that does not require guessing. Someone who has read past halfway is engaged, and an offer at that point reads as a next step rather than a toll gate.
A time delay is the blunt version and works well enough. What is worth avoiding is the immediate trigger, which combines the highest annoyance with the lowest intent, and which is also the one Google's mobile guidance is aimed at.
- Scroll trigger: fires after a share of the page has been read. The best default for an offer.
- Delay: simple, predictable, and fine at ten seconds or more. At two seconds it is an immediate popup wearing a hat.
- Exit intent: fires when the pointer heads for the tab bar. Effective on desktop, and see the next section for what it does on a phone.
- Immediate: correct only when the popup genuinely has to block — an age gate, a region choice — and wrong for everything else.
Exit intent does not exist on a phone
Exit intent works by watching the mouse pointer leave the top of the viewport. A touchscreen has no pointer and no hover state, so there is nothing to detect. The visitor closing your tab, switching apps or hitting back generates no signal a page can act on in time.
Every exit-intent implementation therefore falls back to something else on touch — usually a timer or a scroll threshold — and the important thing is knowing which, because your mobile visitors are getting that behaviour and not the one you configured. Given that most traffic is now mobile, an exit-intent popup is mostly a timed popup with a misleading name.
If exit intent is the right idea for your desktop traffic, configure the fallback deliberately rather than accepting whatever the vendor picked.
The ranking risk, stated accurately
Google's intrusive interstitial guidance is narrower than the folklore around it. It concerns content that is inaccessible on MOBILE because a visitor arriving from search is shown an interstitial — immediately, or while scrolling through the page — that obscures the main content.
Three things follow. A standard announcement bar occupying a reasonable strip is not affected. A popup triggered well after arrival, in response to scrolling or engagement, is a much weaker case than one shown on load. And banners required by law — cookie or age verification — are explicitly acknowledged as legitimate uses, provided they are no larger than they need to be.
The practical version: do not cover a mobile page on arrival from search. Everything else in this guide is a conversion decision rather than a ranking one.
- Affected: a full-screen modal on load, on mobile, for a visitor from search.
- Not affected in practice: a top or bottom bar of reasonable height.
- Explicitly acknowledged: legally required notices, sized reasonably.
- Separate concern: a bar that pushes the page down after a delay causes layout shift, which is a Core Web Vitals problem rather than an interstitial one.
Frequency capping is the setting that decides everything
The difference between a popup people tolerate and a popup people install a blocker to escape is almost never the design. It is whether closing it means anything.
A popup that reappears on every page view teaches the visitor that dismissing it does not work, which is the point at which they stop reading it and start resenting it. Once per visit is the minimum defensible setting. Once every few days is better for anything that is not time-critical.
This is also why frequency capping should never be a paid feature. It is the setting that makes the widget acceptable to the people it is shown to, and charging for it means the free version of the product is the version that damages the site it is on.
- Once per visit: dismissal lasts until the browser is closed. A reasonable floor.
- Once every few days: the right default for a recurring offer.
- Every page view: correct for a genuine service notice, and almost nothing else.
- Whatever you choose, verify that closing it actually persists across pages — that is the part that breaks in a private window, and it should fail towards showing the widget rather than crashing.
How Zyff's widgets handle these
Zyff has all three formats — a popup, an announcement bar and, in the chat launcher, a corner panel that behaves like a slide-in. The popup offers four triggers: a delay of up to two minutes, a scroll threshold between 5% and 95% of the page, exit intent, or immediate. Exit intent falls back to the delay on touch devices, and the editor says so rather than leaving you to discover it.
Frequency capping is free on every one of them, in all three modes: once per visit, once every one to 365 days, or every page view. That is deliberate — charging for the setting that stops a popup nagging someone would monetise the version that damages the site it is on.
The announcement bar can push the page down instead of covering it, which avoids sitting on top of a navigation the site already has pinned there. That option reserves its space rather than inserting it late, so it does not trade one problem for a layout shift.
Try it on Popup Widget
Free, no signup, no watermark. Runs on Zyff’s servers, so you get the same result on a phone as on a desktop.
At a glance
| Popup triggers | Delay, scroll depth, exit intent, immediate |
|---|---|
| Delay range | 0-120 seconds |
| Scroll range | 5-95% of page height |
| Exit intent on touch | Falls back to the delay - there is no pointer to watch |
| Frequency options | Once per visit, once every 1-365 days, every page view |
| Announcement bar height | 32-96px, optionally pushing the page down |
| Frequency capping | Free on every format |
Frequently asked questions
Popup or announcement bar - which converts better?
A popup converts better per impression and costs more goodwill per impression. The right question is what the message deserves: a bar for standing information nobody needs to act on, a popup for something time-critical the visitor loses by missing, and a slide-in for the optional offer in between.
Will a popup hurt my Google rankings?
Only in a specific case. Google's intrusive interstitial guidance is about mobile pages where a visitor arriving from search cannot reach the content because something is covering it. A popup triggered after scrolling, or a bar of reasonable height, is a much weaker case. An immediate full-screen modal on mobile is the one to avoid.
Does exit intent work on mobile?
No. It watches the mouse pointer leave the viewport, and a touchscreen has no pointer. Every implementation falls back to something else there — usually a timer — so on the majority of traffic an exit-intent popup is a timed popup. Configure the fallback deliberately rather than inheriting it.
When should a popup appear?
After the visitor has shown some interest. A scroll threshold is the most reliable proxy — someone past halfway has decided your page is worth reading. A delay of ten seconds or more works nearly as well. Immediate is right only when the popup genuinely has to block the page, like an age gate.
How often should the same popup be shown?
Once per visit at minimum, and once every few days for anything not time-critical. A popup that reappears after being closed teaches visitors that dismissing it does not work, and that is the point where they stop reading it entirely.
Is a sticky bar bad for Core Web Vitals?
It can be, if it inserts itself after the page has laid out and pushes everything down — that is Cumulative Layout Shift. A bar that reserves its space on the first paint, or one that overlays rather than displaces, avoids it. The delay is what causes the problem, not the bar.
Do cookie banners count as intrusive interstitials?
Google explicitly acknowledges legally required notices, including cookie and age verification, as legitimate uses, provided they use a reasonable amount of screen space. A full-screen consent wall on mobile is pushing that considerably harder than a bar at the bottom.
Should I use more than one at a time?
Rarely, and never simultaneously. A bar carrying standing information alongside a popup for a time-limited offer is defensible; two things competing for the same attention at the same moment means neither is read. If both are on the page, stagger them and make sure closing one does not reveal the other.