On October 28, Shopify’s Fall ’26 Edition gave apps a new way into the merchant’s own reports. With App Events, participating apps can send their activity into Shopify Analytics, where merchants can chart it or query it with ShopifyQL next to orders and sessions. Shopify’s example is a loyalty app: an apparel brand compares points earned and redeemed against an outerwear campaign.
The same release adds automatic annotations for events such as app installations. An install now shows up as a marker on your sales chart.
This is a good change. Apps have lived in separate dashboards for years, and every vendor’s dashboard shows that vendor winning. Putting app activity beside revenue makes it easier to ask whether the loyalty program or the upsell widget is paying its way.
But it only fills in one side of the ledger. App Events reports what an app says it did for you. The release notes don’t describe anything that reports what each app costs your storefront in load time and interaction delay.
The only app speed standard is a lab test
Shopify does publish a speed bar for apps. To earn Built for Shopify status, requirement 2.2.1 says an app “must not reduce the storefront Lighthouse performance score by more than ten points.” That’s a sensible gate for an app review queue. It’s also a lab score, not a real-user score, and it judges each app on its own rather than how it performs in combination with others on the same site.
Shoppers don’t get apps one at a time. They get the review widget, the loyalty launcher, the chat bubble, the popup and the analytics tags together, on a mid-range phone. Each app can clear a ten-point bar while the stack as a whole fails. Yottaa’s data puts third-party apps at 44% of eCommerce load time.
There’s a second gap. Lighthouse can’t measure Interaction to Next Paint, because nobody taps anything during a synthetic run; Google’s web.dev guidance says to use Total Blocking Time as a proxy. Interaction delay is where apps bite hardest on product and cart pages, and it’s expensive. Shopify’s April 27, 2026 analysis found that 32 milliseconds of interaction lag costs 1.5% of conversion.
Merchants are already saying this out loud. In a Shopify Community thread from October 19, a speed consultant who ran 5,991 stores through mobile PageSpeed reported a median score of 65 for stores with zero to two homepage apps and a score of 45 for stores with 10 or more. He also found about one in eight stores still loading code from apps that looked uninstalled. Those are his numbers, from lab tests, but the replies matched what we hear: leftover snippets in theme files and overlapping widgets, each loading its own scripts.
An annotation is a timestamp, not a diagnosis
Install annotations will create a new kind of meeting. Conversion dips on Tuesday. Someone pulls up the chart, and there’s the marker: new upsell app, installed Monday. The app looks guilty.
It might be. It might also be a promotion that ended, a shift in traffic mix or a tracking change. Merchants already struggle with this. A long-running Shopify Community thread, still drawing replies in October, started with a merchant blaming two new apps for a jump in abandoned checkouts. The useful answers were about measurement: reconcile checkout IDs against paid orders, segment by device and change one script at a time.
To know whether an app moved conversion through speed, you need the speed data on the same timeline. That means INP and LCP from real sessions, by page template, before and after the install, with the app’s own scripts identified.
Build the cost column before Black Friday
Black Friday is November 27. A one-page app scorecard turns this month’s keep-or-cut decisions into a data call instead of a vote:
- List every app that loads on product, cart and checkout pages. Include code left behind by apps you’ve removed. Shopify says blocks built with theme app extensions are removed automatically on uninstall; code added any other way can stay long after it was supposed to be removed.
- Fill the value column. Use App Events where a vendor supports it and the vendor’s own dashboard where it doesn’t. Ask each vendor whether they plan to send App Events.
- Fill the cost column from field data: each app’s script timing on real sessions, plus INP and LCP on the pages where it runs. Check mobile, and check Safari as well as Chrome.
- Decide per app: keep it, load it after the first interaction, limit it to the pages that need it, or remove it.
- If something must be installed before peak, watch real-user INP and LCP for a few days before you trust the sales chart.
This is the work Yottaa does. Yottaa RUM captures every session across browsers, and Application Sequencing changes when a third-party script loads without a code change. Whatever tooling you use, the principle holds: an app earns its place with a value number and a cost number, measured on the same shoppers.
Where this goes from here
Shopify has made app activity a first-class dataset. The obvious next step is putting app cost beside it, whether Shopify builds that or merchants bring their own. Until then, the sales chart will tell you when an app arrived. Your field data has to tell you what it costs your site in terms of performance.
Which apps on your store would survive a value-versus-cost review?
Subscribe to our blog
Performant Experimentation – A Guide to A/B Testing
Every A/B test you run is a purchase. You’re buying information—an answer to a hypothesis—and you’re paying for it in degraded performance and lost conversion during the test window.
Download eBook
