Do popups hurt UX? When they do, and when they do not

Popups affect UX in both directions. A delayed, easy-to-close bar can save a cart or explain shipping. A first-load full-screen wall on mobile search is the pattern Google still calls an intrusive interstitial. Core Web Vitals now include INP at 200ms. Consent dialogs are a separate object. Do not stack them.

When do popups improve the visit?

When they answer a question the page did not. Shipping total, a size guide, guest checkout, a related article at 75% scroll. The visitor was already doing the task. The overlay shortens it.

Baymard still puts average cart abandonment at 70.22% (50 studies, 22 September 2025). 40% of the non-browsing reasons are extra costs. A cart overlay that states the shipping number is UX, not "interruption marketing."

When the visitor opens it. Click-to-open demo forms, help panels, and share sheets are popups. They do not fight the page because the page asked.

When the timing proves interest. Omnisend saw email forms at 2.4% conversion after a 6 to 10 second delay vs 1.9% on immediate display (2025, 1.24 billion displays). Late is ruder in theory and kinder in the data.

When do they make the site worse?

First paint on a phone, covering the thing the person tapped in search. Google's intrusive interstitial page (updated 10 December 2025) tells you not to obscure the entire page with a promotional dialog, and to prefer a small banner. Cookie and age gates are mandatory exceptions. A newsletter wall is not.

Hard to close. No visible X, no Escape, a 5-second lock, a fake close that opens a second step. That is not a conversion tactic. That is a trap.

Stacked dialogs. Cookie banner plus sale modal plus chat plus a wheel. The person cannot reach the article. They also cannot give valid consent to any one of those layers.

Layout shift and slow taps. Core Web Vitals in 2026 are LCP (2.5s), INP (200ms, replaced FID in March 2024) and CLS (0.1). A bar injected into document flow after load is CLS. A 200KB popup bundle on every URL is INP.

Irrelevant offers. Exit-intent on a help article offering 15% off a product the reader was not looking at. Target by URL and by "already subscribed."

Do popups change search rankings?

Not as a popup-shaped penalty with that name. Google documents two separate issues:

1. Intrusive interstitials, aimed at promotional overlays that block content after a search click, especially on mobile.
2. Page experience, of which Core Web Vitals are the measured part.

A well-timed slide-in that does not cover the article, does not shift layout, and does not block the first tap is unlikely to be the ranking story. A full-screen takeover on every blog URL from Google can be.

Hiding the overlay from Googlebot while showing it to users is the kind of cloaking conversation you do not want. Fix the overlay.

How do you keep the good version?

• One campaign per page. Two if one is a tiny bar.
• Easy close, including tap-outside and Escape.
• Delay or scroll. No first-load modal on content URLs.
• Mobile sheet or bar, not a desktop modal scaled down.
• Cookie CMP first, and finished, before any marketing overlay.
• Measure bounce, INP and revenue with the campaign on and off.

Sleeknote is best if the overlay has to look like the store and you will use teaser states instead of a hard modal. OptinMonster is best if you will test timing. Hello Bar is best if a bar was the whole UX requirement. Popupsmart is fine for a first delayed campaign if you will set frequency caps.

Older versions of this page quoted brands we cannot verify from primary sources (YouCof, Astcario, a bicycle campaign). Those stories are out. If you need a method, A/B the overlay against nothing for a week and keep the variant that does not raise landing-page bounce.

More setup detail: how to use popups and popup mistakes.