Scroll-Triggered Pop-ups: Engagement-Based Targeting
Scroll depth as a proxy for interest, the threshold that works, and the page types where it misfires.
On this page
The short version
- Scroll depth is the best single predictor of interest available to a pop-up, and it is available in every tool.
- Use a percentage of the article, not of the page. A page with a long footer or a comments section makes percentage-of-document meaningless.
- It misfires on short pages, on listings, and on anything a reader scans rather than reads — those need a different trigger or none.
Scrolling is the one thing a reader does that is unambiguous. A cursor can drift, a timer can run in a background tab, but scrolling past halfway is an action taken deliberately by someone present.
That makes scroll depth the strongest simple trigger, and the caveats are all about what you are measuring the depth of.
Percentage of what
Most tools default to a percentage of the whole document. On a page with a long footer, a comments section, a related-posts grid and three ad slots, fifty percent of the document can be the middle of the article — or the end of it, or somewhere in the footer, depending on the page.
The measurement that means something is percentage of the article element. Someone who has read two-thirds of the actual content has demonstrated something; someone who has scrolled two-thirds of a document that is half furniture has not.
If the tool cannot target an element, an absolute pixel depth calibrated per template is a workable substitute — cruder, and still better than a percentage of a document whose length varies with the comment count.
Where a threshold lands on a typical article page
The same 50% setting, on three page shapes. This is why the default produces inconsistent behaviour across a site.
Illustrative of the effect rather than measured figures. Check where your own threshold lands by scrolling your longest and shortest templates with the trigger visible.
Where scroll depth is the wrong signal
Short pages, where reaching any threshold takes one gesture and means nothing. Listing and category pages, where scrolling is browsing rather than reading. Reference pages people jump around in. And anything with infinite scroll, where the document grows as you go and a percentage never arrives.
On those, either use a different trigger — a second page view works well for listings — or accept that the page does not warrant an interruption.
Setting the threshold
Somewhere between a half and two-thirds of the article, for most content. Earlier than that and you are interrupting people who have not yet decided the article is worth their time; later and you are only reaching people who were going to finish anyway.
Add a small requirement that the scrolling was not instant. A reader who hit the end key and jumped to the bottom has not read anything, and treating that as engagement is the most common false positive on this trigger.
Two configurations
The default
- 50% of the document
- Fires on any scroll speed
- Same setting on every template
- Comment threads move the threshold
Worth setting up
- 60% of the article element
- Requires the scroll to have taken some seconds
- Disabled on listings and short pages
- Independent of comments and footers
Instrument it before you tune it
Most tools tell you how many impressions a scroll trigger produced and not where in the page it fired. Those are different questions, and the second is the one that explains an odd conversion rate.
The cheap instrument: record the scroll position and the article height alongside each impression, for a week. What usually comes back is a distribution rather than a point — the same threshold firing at very different places depending on the template, which is the percentage-of-document problem showing up in data rather than in theory.
It also surfaces the pages where the trigger never fires at all, because the article is shorter than the threshold requires. Those pages have a pop-up configured and no pop-up in practice, which nobody notices because the tool reports zero impressions rather than an error.
Scroll trigger check
- Measured against the article, not the whole document
- Threshold between a half and two-thirds
- Instant scrolls to the bottom do not qualify
- Disabled on listings, short pages and infinite scroll
- Checked on your longest and shortest templates
- Still subject to the same frequency cap as any other trigger
Frequently asked questions
Should the threshold differ by page length?
A percentage of the article handles that automatically, which is the main argument for it. A fixed pixel depth needs a per-template value and will drift as templates change.
Can I combine scroll depth with a timer?
Yes, and it is the strongest simple combination: past the threshold and engaged for some seconds. It excludes both the fast scroller and the background tab.
Does firing later reduce total signups?
Fewer impressions, higher conversion, and usually a similar or better total — with a considerably smaller reader cost. Judge it on signups per session.