Possible Bug - High INP from Vanilla Input Boxes with no JS

[Michael Clark]

Dear web-vitals-feedback,

My network of websites provide a number of calculators that allow users to rank their strength and fitness. Our users love our websites and report no issues.

Unfortunately there seems to be a pervasive issue where the Android OS is showing a Presentation Delay in INP metrics when using the on screen keyboard. This is particularly bad on Samsung. Because of this I cannot get my websites passing the INP metric.

I have had two calls with PageSpeed experts and used two RUM solutions (thank you to DebugBear and CoreDash/corewebvitals.io). After improving one element, the first RUM solution showed my site as high performance but PageSpeed was still faiing. I signed up for a second RUM and the second showed my site as failing on inputs due to presentation delay.

Failing - https://strengthlevel.it/calcolatore-di-carico-massimale (2 vanilla inputs, no adverts)

Passing - https://strengthlevel.it/standard-di-forza/panca-piana (inputs, no adverts, but inputs rarely used)

Failing - https://runninglevel.com/ (inputs)

Passing - https://runninglevel.com/running-times/5k-times (no inputs)

Our users love these calculators and really want to find my content on Google so I am sad I am being penalised for operating system behaviour related to the keyboard. I know if I just offered articles I wouldn't have any problems with INP but users do want to be able to check their health and fitness on the homepage.

Would it be possible for someone to look into this further?

Until this bug is resolved is there a possibility to apply mitigation for my network of websites?

Michael Clark

Founder, Strength Level

--

INP is caused by Presentation Delay and is always around 500 ms

INP by Operating System/Browser shows Android/Chrome and Samsung

 

INP Event is keydown/click on inputs

INP element interactions happen on vanilla inputs on pages with no custom JS (#oneRepMaxCalculator)

Distribution on vanilla inputs is 1/3 good, 1/3 bad 1/3 failing

Histogram shows unusual distribution

Finally, I used a dev-phone Pixel 4a on an older Chrome and got an INP score of 552 ms. My personal experience was that the website worked perfectly and I had no delay in using it. Again in this case it is a vanilla input box with no JS.

[Barry Pollard]

Thank you for this feedback.

We have also noted this issue on some Android devices. This has been raised to the Android team and you can follow this bug for updates https://issues.chromium.org/issues/41485748

Unfortunately, we do not have an ETA of when this bug will be resolved.

--
You received this message because you are subscribed to the Google Groups "web-vitals-feedback" group.
To unsubscribe from this group and stop receiving emails from it, send an email to web-vitals-feed...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/web-vitals-feedback/8b469c93-e4e8-4bd8-9b9f-0e314da9bf64n%40googlegroups.com.

[Michael Clark]

Hi Barry,

Thank you for the quick reply. That gives me peace of mind and let's me get back to work -- I have spent weeks optimising the site, so at least I can focus on user-facing improvements while this is being handled at Google. Thankfully this focus has improved performance so it is not all wasted time.

I've added a comment on the Chromium bug tracker #41485748 detailing the INP issue's impact on our site's Core Web Vitals and providing a link to this thread for RUM data insights.

I'm hopeful this added information will help underscore the urgency of the issue for all affected sites. I'm also wondering if this issue is causing the Google homepage to fail in its INP scores (being <input> focussed).

Thank you again for your attention to this matter.

Best wishes,

Michael Clark
Founder, Strength Level