FP-118 · Funeral Prints · 15 Sep 2026 · Trello card

Click to Design: the 8 requests for Infigo

The 8 requests

tap → page unloads → what came back

Recording links open the player 10 s before the tap. Times UTC. "start_design" in the player's activity list is the tap.

"Nothing" means whatever came back carried no tracking, so it was not a site page and not the designer. A blank or error page fits. A visitor closing the tab looks the same in the event log, which is why the request lookup settles it.

Three more cases sit in Hotjar with order product ids Infigo can look up directly: 973511, 973794, 976286. JC's 44 remaining Hotjar recordings are grouped with links on the Trello card. Order products 980509 to 980511 are our own test from 6 Sep, ignore those.

Why the home page case is proof, in order

21 Aug, recording 01a02660 MP4, times UTC. No designer opened, and nobody closed one.

1. 22:11:23 to :26. Three taps on Click to Design. The page did not change, so the visitor tapped again twice.
2. 22:11:26 to :32. Waiting. No designer page loaded, no designer event, no designer console line. Opening and closing a designer leaves all three.
3. 22:11:32. The home page appears, and its own tracking reports where it came from: the StartEditor address of the click. So Infigo's start step received the click and answered with the home page.
4. 22:11:32 to 22:18:10. The visitor searches for the product by hand.
5. 22:18:24. The designer opens normally on the second try.

Five more visitors gave up during the wait

Tap, page unloads, then the previous page of the visit appears. A Back press, not a server answer. These show the silent wait, not a wrong answer. An earlier version of this page listed them as "server sent a category page", and that reading was wrong.

How often

• Every tap waits with no feedback on screen. Median 4.6 s from tap to designer, 1 in 4 visitors over 8 s, 1 in 10 over 11 s (4,143 sessions, 19 Aug to 5 Sep).
• Recorded taps since 19 Aug: 2,720 in 1,308 visits. 2,299 opened the designer within a minute, 99 opened it without a page view being logged, 8 are on the lookup list, 5 are Back presses.
• Re-clicks: 9% of visitors tapped again within 15 s. Since 2 Sep the button dims on tap and re-clicks fell to 1 to 3%.
• PostHog recorded 15% of visits until 14 Sep and 100% since. Expect about one new lookup case every 2 days from now on.

Ruled out

• The browser. The button is a plain form submit. The three scripts listening to it (field validator, Infigo diagnostic, Klaviyo tracker) never cancel it. No console error at the tap. Every recorded tap left the browser.
• The replay picture skipping or freezing. All times come from the event log clock, not the player.
• The designer's own errors ("$ is not defined" and others). They fire inside an already open designer.

A second, separate bug: the designer stalls inside its own loading

8 cases in 2,708 loads. Here the click worked and the designer page arrived within seconds, then one of its own loading steps hung. Read from the designer's console log. Median designer load is 3.5 s and 9 in 10 finish under 8 s. The 32 to 131 s stalls are real; the 4 to 41 min ones are phones put to sleep mid-load. Not part of the Infigo request lookup.

Also removed from the click problem: three phone visits on 14 and 15 Sep with the designer 7 to 10 min after the tap. The phone was asleep, and once the request went out the server answered in under 9 s.

How to read a recording

1. Trust the activity list, not the picture. Events are sent with their own time stamps. The picture is sent in chunks, and the last chunk before a page unloads often never arrives.
2. Switch off "Skip inactivity" (settings cog under the video). A frozen page counts as idle, so the player fast-forwards the wait.
3. After a Back press the browser restores the page from memory without a fresh picture, so 01a03bd8 and 01a06522 look like "already in the designer".
4. When the next page opens under a new tab id the player switches tabs, which is why 01a029c5 and 01a02c9e jump.

The event log cannot tell a blank server answer from a closed tab. Infigo's server log settles that, which is the reason for the lookup list.

Next

1. Infigo: status, redirect target and response time for the 8 requests, starting with StartEditor/2793 on 21 Aug.
2. Us: watch the 100% recording for new cases, and match the 6 "nothing" visits in Hotjar.

Sources: PostHog events, console logs and replays (project 365336), Hotjar recording markers (site 2580074), the product page code as served on 6 and 14 Sep 2026.