Mobile INP (>200 ms) issues appeared after ending a one‑month SEO agency engagement; Google Search Console flagged the problem and PageSpeed Insights scores have been declining. Shopify support noted the agency added code during the engagement that was removed when the service stopped.
A respondent warns the previous agency may have artificially boosted reported performance, so the site might have been slow all along and only now reflects true metrics after their code was removed. The concern is about “fake” improvements versus genuine optimization.
Suggested validation steps:
Attempt to restore the removed code to verify its effect (possible if it was theme code; not feasible if injected via Shopify ScriptTags API—Shopify mechanism to add JavaScript to a storefront).
Review archived site versions on archive.org to inspect prior code and behavior.
Provide context (store URL, theme name) to enable deeper troubleshooting.
Technical context: INP (Interaction to Next Paint) measures delay from user input to the next paint; >200 ms on mobile is poor. Tools used: Google Search Console and PageSpeed Insights.
Status: Unresolved. Key questions remain whether past improvements were manipulative and precisely where the code changes occurred.
Summarized with AI on December 19.
AI used: gpt-5.
I recently hire SEO agency for a month and decided to stop the service due to my budget. On the day I stopped the service that’s when I started seeing INP issue longer than 200ms mobile on my google console. I contacted shopify and they said the agency had added Java code when we started working with them. and it looks like the code was removed on the date i stopped the service. I then went to PageSpeed Insights to dive deep on this issue and I am currently feeling helpless trying to fix this issue. As the days go by the score keeps on dropping with more red flags. Please help!
I contacted shopify and they said the agency had added Java code when we started working with them. and it looks like the code was removed on the date i stopped the service
Did the really fix things, or did they use a fraudulent trick to makes reporting tools spit out fake numbers.
Did you only ever read their reports, did you personally only ever use one tool like pagespeed and not something more in depth.
Depending on where the code was removed I could potentially restore it and validate if what they were doing was legitmate; though not if it was a script injected by the scripttags api. But restoring it would be for validation only you wouldn’t be able to continue using code without license.
You could also look through archive.org to see if there’s an archive of the site with the code to validate you didn’t get scammed.
If you have an actual budget and need actual performance customizations really done to stop losing revenue then contact me by my email for services, or if you have access private message by clicking here (sloooower).
Contact info in forum signature.
ALWAYS please provide context, examples: store url, theme name, post url(s) , or any further detail in ALL correspondence.
It’s not the current issue or current scores, did the previous “service” make the OLD scores look better with tricks so now it only seems like there’s an issue because there was an issue the whole time because the numbers your basing things on were fake.
