Speed guides tell you how to make a Shopify store faster. Conversion guides tell you how to make it sell more. This one sits between them: it is the plan for turning performance work into revenue you can see in Shopify Analytics, with a business owner’s calendar rather than a developer’s backlog. For the technical how-to, read how to speed up a Shopify store and Core Web Vitals for Shopify; for the tactics that do not involve speed at all, read Shopify conversion rate optimization. If you would rather hand the whole thing over, that is our Shopify speed optimization service.
Why performance is a sales problem, not a technical one
The numbers on this topic are well established and worth repeating because they are the whole argument. Roughly 47% of shoppers expect a page to load in two seconds or less; 53% of mobile visits are abandoned when a page takes longer than three seconds; as load time grows from one second to five, the probability of a bounce rises by about 90% (Google); and every extra 100 milliseconds of load time has been shown to cut conversions by around 7% (Akamai). About 82% of consumers say slow loading affects their decision to buy. Put those together and “the site is a bit slow” becomes “we lose a measurable share of every ad dollar before the product page renders”.
That framing changes who owns the problem. A slow store is not a developer’s backlog item; it is a line in the marketing P&L. The plan below is written so that the person who owns revenue can run it, with a developer executing the fixes.
Put a revenue number on your load time
Do this once and the rest of the plan funds itself. Take three numbers from Shopify Analytics for the last 30 days: sessions, conversion rate and average order value. Take your mobile largest contentful paint (LCP) from PageSpeed Insights field data for the product page template. Then run one simple model: if performance work moves you from your current LCP to the “good” threshold, assume a conservative conversion improvement in line with the industry figures above, and multiply by sessions and order value. You do not need the model to be exact; you need it to tell you whether the work is worth $2,000 or $20,000. A store doing $100,000 a day, to use the figure cited on our headless guide, can generate about $7,000 more daily from a one-second improvement; scale that to your size and you have the budget line.
The metrics that actually move sales
Not every performance metric matters to shoppers. Three do, in this order:
- Largest Contentful Paint on mobile product and collection pages. This is the moment the product image appears. It is the metric closest to “can I see what I came for”, and it is where most Shopify stores fail the Core Web Vitals threshold.
- Interaction to Next Paint. How quickly the store reacts when someone taps a variant, opens the cart drawer or filters a collection. Slow interaction feels broken even when the page loaded fast; it is usually caused by app scripts.
- Cumulative Layout Shift. Content jumping as images, banners and reviews load, which causes mis-taps and abandoned carts on mobile.
Total page weight, request count and desktop scores are diagnostics, not goals. A store can score 60 on desktop Lighthouse and convert well; a store with a slow mobile LCP will not.
The 30-day speed-to-sales plan
Days 1–5: baseline and the money model
- Record mobile LCP, INP and CLS for the homepage, a top collection and a top product page, from field data (real users) not only lab tests.
- Record conversion rate, mobile conversion rate, bounce rate on product pages and revenue per session for the same 30 days.
- Build the revenue model above and agree the budget with whoever signs it.
- List every app installed and what it does on the storefront. Most stores find apps nobody remembers installing.
Days 6–12: the app cull and script hygiene
The fastest win on almost every Shopify store is removing what should not be there. Uninstall apps with no owner and no measurable benefit; for the rest, check whether they inject scripts on every page or only where needed. Move tracking to Shopify’s web pixels where possible and load remaining third-party scripts asynchronously. This step alone often moves INP from red to green and costs nothing but decisions.
Days 13–19: images and the above-the-fold path
Product and collection images are the LCP element on nearly every Shopify page. Serve them at the right size for each viewport, in WebP or AVIF, with lazy loading below the fold and eager loading for the hero and first product images. Cut hero videos or defer them. The details are in our Shopify image optimization guide. Measure LCP again on the same three pages.
Days 20–26: theme code
Minify and defer JavaScript, remove unused CSS, cut render-blocking requests, reserve space for late-loading elements to stop layout shift, and fix any section that loads all its data on every page. This is developer work and the largest single step; scope it from the diagnostics of days 1–19 so the developer fixes what field data says is slow, not what a lab tool lists.
Days 27–30: measure the business result
Compare the same four business metrics with the baseline, on the same traffic mix if you can (pause big campaigns for the comparison window or compare like-for-like weeks). Attribute honestly: performance work rarely explains a 30% jump; it usually explains a few points of conversion, which at scale is real money. Write the result down next to the model from day 5 so the next round has evidence.
Where the money hides: template by template
Not every page deserves the same attention, because not every page carries the same revenue. Look at Shopify Analytics for sessions by landing page and the conversion funnel, then rank templates by the revenue that flows through them. For most stores the order is: product pages first (they are the landing page for paid and organic traffic and the LCP element is a product image), collection pages second (filters and grids are where INP problems live), cart and cart drawer third (a slow drawer is a direct abandonment cause), homepage last. Homepages get the most attention from owners and the least from shoppers who arrive from an ad or a search result. Spend the developer’s time in revenue order, and measure each template separately so a homepage win does not hide a product-page loss.
Why performance work fails to show up in sales
- Measuring the lab, not the field. A Lighthouse score is a simulation; Search Console and CrUX field data are your real customers on real phones. Optimize for the field number.
- Fixing desktop. Most Shopify traffic is mobile. A desktop score of 95 with a mobile LCP of four seconds is a failed project.
- Installing a speed app. Apps that “optimize speed” add a script to do it. Some help with images; none fix theme code or third-party scripts, and several make INP worse.
- Comparing different traffic. A sale week against a quiet week tells you about the sale. Compare like for like or the result is noise.
- Stopping at the score. Speed is a precondition, not a guarantee. If the page is fast and still does not convert, the problem is the offer, the product page or the checkout, and that is the next plan.
What to do after day 30
Performance decays. New apps arrive, banners get heavier, someone uploads a 4 MB hero. Put two habits in place: a monthly look at field Core Web Vitals in Search Console, and a rule that no app is installed without a named owner and a check on storefront scripts. Then move on to the conversion work that speed unlocks: product page structure, checkout completion (our checkout optimization guide), and the CRO testing programme. Speed is the floor; the ceiling is everything you build on it.
When to bring in help
Days 1–12 are within reach of any store owner with admin access. Days 13–26 need a developer who has done Shopify performance work before, because theme code is where amateurs make stores slower. If you do not have that person, our speed optimization service starts with the same baseline and the same money model, and our CRO service takes over from day 30.
