Timed Pop-ups: When to Trigger for Maximum Conversion
Why a fixed delay is the weakest trigger available, and what to use instead of a number of seconds.
On this page
The short version
- A fixed delay is the weakest trigger available, because time on page is a poor proxy for interest — a reader who left the tab open in the background looks identical to one reading closely.
- Scroll depth, pages per visit and time with the tab actually focused all predict better and are no harder to implement.
- If you must use a timer, measure engaged time rather than elapsed time, and set it long enough that arriving and leaving does not qualify.
The timed pop-up is the default in almost every tool: wait five seconds, then show. It is the easiest trigger to configure and the least informative one available.
The problem is not that five seconds is too short. It is that elapsed time does not distinguish between the states you care about.
What a timer cannot tell apart
After thirty seconds on a page, a visitor might be reading carefully, or have opened your article in a background tab and gone to lunch, or be sitting on a page they landed on by mistake with the window behind another one.
All three look identical to a timer, and only one of them is a reasonable moment to interrupt. That is the whole case against it.
Worse, the second case is common — background tabs are how a great many people read — and it produces a pop-up that is shown, never seen, dismissed automatically on navigation, and counted as an impression that failed to convert. Your conversion rate is being diluted by people who were never there.
Triggers, by how well each predicts interest
| Trigger | Predicts interest | Note |
|---|---|---|
| Second page view in a visit | Strongly | They chose to continue |
| Scroll past a threshold | Strongly | They read down the page |
| Engaged time with the tab focused | Moderately | Excludes background tabs |
| Returning visitor | Moderately | They came back |
| Exit intent | Weakly, but cheap | The visit was ending |
| Elapsed time on page | Barely | The default in most tools |
| Immediate on arrival | Not at all | Nothing has happened yet |
Combining signals
The arrangement that works best is two conditions rather than one: scrolled past a threshold and been engaged for some time. Either alone produces false positives — a fast scroll to the bottom, or a long dwell without reading — and together they are a reasonable proxy for someone who is actually there.
Keep the thresholds modest. The point is to exclude the clearly disengaged, not to require someone to have finished the article before you are willing to speak to them.
If you are stuck with a timer
Set it long. The common five-second delay excludes almost nobody: someone who arrived at the wrong page and is about to leave has usually been there five seconds.
Twenty to thirty seconds of engaged time is a more defensible floor, and it should be measured per page rather than per session, so a reader who has just arrived on a new article is not immediately interrupted because they were on the site earlier.
Auditing a timer you inherited
Before changing anything, find out what the current setting is actually catching. The number to pull is impressions against sessions: if the pop-up is being shown on nearly every session, the timer is not filtering anyone and is functioning as an arrival trigger with a delay.
Then compare the conversion rate of impressions in the first minute against those later. A large gap in favour of the later group is direct evidence that the early impressions are being wasted on people who were never engaged — and it is the argument that will convince whoever set the timer in the first place.
The third check is bounce rate on sessions that saw the pop-up against those that did not. It will be higher, because the pop-up fires on people who were staying; what matters is whether the gap is small or alarming.
Trigger check
- Not firing on arrival
- The timer pauses when the tab is hidden
- Two conditions rather than one, where the tool allows
- At least twenty seconds of engaged time if a timer is the only option
- Measured per page, not per session
- Impressions from background tabs are excluded from the conversion rate
Frequently asked questions
Is there a best delay?
Not as a number, because it depends on how long your pages take to read. Set it as a proportion of the page — after roughly a third — rather than as a fixed count of seconds.
Do longer delays reduce signups?
They reduce impressions and usually raise the conversion rate, and the net often improves because the interruption lands on people who were engaged. Measure signups per session rather than per impression to see it.
What about firing on a specific scroll position tied to the content?
That is the strongest version — an offer that appears immediately after the section it relates to. It is closer to an in-content form than a pop-up, which is rather the point.