Shopify INP / Interaction to Next Paint poor — heavy theme JS, third-party apps blocking add-to-cart.

Several products can look like “my store feels slow / laggy.” They do not share a settings panel. This page is only when Interaction to Next Paint (INP) is poor — lab or field INP red / needs improvement — while first paint and layout may still look fine. Poor LCP / slow paint is shopify-pagespeed-lcp-poor. High Cumulative Layout Shift is shopify-cls-layout-shift. A cart drawer / AJAX cart that never updates after add is cart-drawer-not-updating. A sticky / mobile buy bar that never shows or never adds is sticky-add-to-cart-not-working. An App embed that never appears is app-embed-not-loading. This page is only: INP / Interaction to Next Paint poor.

Six ordinary reasons INP stays poor: theme JavaScript that runs long tasks after every click, App embeds and marketing scripts stacking listeners on add-to-cart, variant / filter handlers that re-render the whole section, chat / reviews / upsell widgets polling on the main thread, third-party analytics firing synchronously on interaction, or you are debugging LCP, CLS, cart drawer staleness, sticky ATC, or a missing app embed instead. Walk them in that order. You already pay for Online Store. You do not need another seat to read a PageSpeed INP report.

Open PageSpeed Insights on a live product URL (not only the homepage). Force mobile. Read INP score, Total Blocking Time, long tasks, and Diagnostics that mention main-thread work or third-party code. Field data (CrUX) and lab data can disagree — treat field as the score that matters for Search, lab as the debug surface.

DevTools → Performance: record a mobile session, then click add-to-cart, a variant swatch, and a collection filter. Mark long tasks between pointer-up and next paint. Honest proof: if LCP is the only red metric and clicks feel instant, leave for shopify-pagespeed-lcp-poor. If the page jumps after load without laggy clicks, leave for shopify-cls-layout-shift. If the cart never updates at all, leave for cart-drawer-not-updating. Do not mix those proofs with a page that paints fine then stalls on interaction.

Theme bundles that parse large product JSON, run animations, or re-query the DOM on every interaction create long tasks. INP measures the delay from input to the next paint — a 300ms+ main-thread block after add-to-cart is the product.

Fix: Themes → Edit code / Customize on Current. Defer non-critical theme scripts, split cart / product JS from homepage chrome, and avoid synchronous work in click handlers. Prefer Sections API updates over full-page re-renders. Publish Current. Hard-refresh and re-run mobile PSI INP. Slow first paint without interaction lag is shopify-pagespeed-lcp-poor— different from a theme that paints fast then stalls on tap.

Reviews, upsells, wishlists, chat, loyalty, and analytics embeds often attach to the same add-to-cart button. Each listener runs before the next paint. Stack five of them and INP collapses — especially on mobile PDPs.

Fix: Themes → Customize → App embeds on Current. Turn off embeds you do not need on the PDP. Prefer deferred / after-interaction loading when the vendor allows. Capture Performance with one embed at a time to see who owns the long task. An embed that never appears at all is app-embed-not-loading— different from embeds that load and then block clicks. Retest mobile PSI INP after each embed change.

Variant pickers, collection filters, and cart / Sections API handlers that re-render large DOM trees, fetch serial network calls, or run analytics sync on the critical path delay the next paint even when the cart eventually updates.

Fix: keep the first paint after click cheap — update the pressed control and cart count first, defer secondary fetches. Avoid awaiting three apps before opening the drawer. A cart that never updates (wrong section id, unbound drawer) is cart-drawer-not-updating. A sticky / mobile buy bar that never shows or never adds is sticky-add-to-cart-not-working. This page is when the click works but feels laggy before paint.

PageSpeed / LCP poor on mobile is shopify-pagespeed-lcp-poor. CLS / layout shift high is shopify-cls-layout-shift. Cart drawer / AJAX cart not updating is cart-drawer-not-updating. Sticky add to cart not working is sticky-add-to-cart-not-working. App embed not loading is app-embed-not-loading.

The only honest test for this page: mobile PageSpeed / INP is poor while the page still renders — heavy theme JS, third-party apps stacking on add-to-cart, or handlers that do too much before the next paint. Capture the long tasks, repair those three, publish Current, retest mobile PSI. Anything else is a different product.

Stay DIY if you can spend one afternoon on the list above: prove mobile INP, defer heavy theme JS, rein in App embeds on the PDP, keep add-to-cart / variant handlers lean, and stop confusing LCP / CLS / cart-drawer / sticky-ATC / app-embed with interaction lag. That is the whole playbook. No agency required. Short answers also live on the merchant FAQ. Related pages: shopify-pagespeed-lcp-poor, shopify-cls-layout-shift, cart-drawer-not-updating, sticky-add-to-cart-not-working, app-embed-not-loading.

Pay for help when the work is not the INP row — it is connecting the stack you already have so lead capture is not dying in DMs, storefront and checkout-ready exceptions land in a channel you already check, and the live store is one workflow instead of a spreadsheet plus a Slack bot you forgot. That is the Latch AI Ops install.

What $997 USD one-time buys: lead capture on the live store, ops alerts in a channel you already check, and a checkout-ready workflow. 72 hours after collaborator access, not after payment. One live Shopify store. Access plus a delivery thread. Store URL collected at checkout. Not a SaaS seat, not a theme rebuild retainer, not a conversion guarantee, not a PageSpeed agency job.

Other shops’ public ranges, not testimonials: a checkout / ops install typically runs $500–$2,500. A full ops consulting block typically runs $2,000–$5,000. $997 is the cheap end of that work, sold as a 72-hour install of three systems — not as an earnings number or conversion-lift claim.

What it does not buy: more orders, recovered revenue, a forever INP / Core Web Vitals guarantee, or a case study. Latch AI Ops has not published customer proof because there are no customers yet. You are buying the install.

Checkout is on Whop. Operator is Nico. United States offer. You can revoke collaborator access after handoff. We do not take owner passwords.