Shopify Speed Benchmark 2026: Performance Data from 1,533 Stores Across 9 Industries

Most Shopify speed studies test one page type, report one number, and call it a day. We wanted more than that.

For this benchmark, we screened 2,332 Shopify stores across 9 industries and retained 1,533 stores with CrUX field data, mobile lab, and desktop lab data for the performance comparisons below. Each eligible store was tested on 3 page types (homepage, product detail page, and collection page), with 3 runs per page, on both mobile and desktop. That's roughly 28,000 successful PageSpeed Insights tests and 20.7 GB of raw Lighthouse data.

Beyond lab tests, we collected real-world Chrome User Experience Report (CrUX) field data, detected the theme each store runs, and tracked the third-party services loading on each site.

The result is, to our knowledge, one of the most detailed public Shopify speed benchmarks available. Here's what the data shows.

How We Tested

Transparency matters for benchmarks because different studies measure different things. Here is exactly what we did, so you can judge the data on its own terms.

Store Selection: We started with 11,464 Shopify domains sourced from ecommerce directories and classified each one into 9 industries using an AI classification pipeline with a manually verified rubric. Each store was checked for reachability, confirmed as running on Shopify, and validated against a product sitemap. We excluded stores that were password-protected, pre-launch, headless (custom storefronts not using a standard Shopify theme), B2B-only, or service businesses without a shippable product catalog. We also excluded stores with fewer than 10 products. Duplicate parent/child domains were deduplicated by myshopify subdomain. The final dataset contains 2,332 stores. Of these, 1,533 had enough real-world Chrome traffic to appear in the Chrome User Experience Report. These 1,533 received the full benchmark treatment: 3 runs across all 3 page types on both mobile and desktop. The remaining stores were screened on their homepage only and are not included in the performance comparisons below.

Page URL Selection: For each store, we tested three page types. The homepage URL came directly from the store's domain. The product detail page (PDP) URL was selected by taking the first product listed in the store's Shopify product sitemap (sitemap_products_1.xml). The collection page URL was selected from the store's sitemap or, when available, the /collections/all page. This means PDP and collection results reflect one representative page per store, not an average across all products or collections.

Testing Window: All PageSpeed Insights tests were run between March 19 and March 24, 2026 (6 days). Industries were tested in parallel batches. Some completed within a single day (Automotive: 9 hours, Food & Beverage: 9 hours), while others ran across multiple days due to API rate limits (Beauty: 85 hours, Consumer Goods: 84 hours). Each page type was tested 3 times on both mobile and desktop, giving 18 individual test runs per store. We report the median of the 3 runs for each page/strategy combination to reduce noise from outlier measurements.

Lab Environment: Tests used Google's PageSpeed Insights API v5, which runs Lighthouse under the hood. Mobile tests simulate a Moto G Power (2022) on a throttled 4G connection with CPU slowdown. Desktop tests run without network or CPU throttling. The Lighthouse version used was the latest available through the PSI API at the time of testing.

Field Data (CrUX): For stores with enough real-world Chrome traffic, we also collected Chrome User Experience Report data. CrUX reflects actual visitor experience over the prior 28 days, measured from opted-in Chrome browsers. 1,533 of our 2,332 stores (66%) have CrUX data. This subset is likely biased toward larger or more established stores, since CrUX requires a minimum traffic threshold. The lab and field data were not collected in the same time window, so comparisons between the two are directional rather than exact. Only stores with data across all three dimensions (CrUX field data, successful mobile lab runs, and successful desktop lab runs) are included in the performance comparisons below. Having CrUX data means the store appears in the Chrome User Experience Report, but individual metrics (LCP, CLS, INP) require sufficient samples of each type. Most stores have LCP and CLS data, but 11% lack INP because their visitors did not generate enough click or tap interactions for Google to report a p75 value.

Third-Party Detection: When we refer to "third-party services" in this study, we mean entities detected by Lighthouse's built-in entity recognition system, which identifies known services by their network request domains. This is not the same as the number of apps installed in a store's Shopify admin. A single Shopify app might load zero external scripts, while a manually pasted tracking pixel counts as a third-party entity. Across the 1,533 stores, Lighthouse detected over 2,200 unique third-party entities. Of those, 733 were categorized into one of 24 functional groups (advertising, analytics, marketing, reviews, payments, and so on). Well-known services like Google Analytics or Facebook Pixel were categorized automatically by Lighthouse. The remaining 438 less common entities were verified individually through web research. These categories feed the entity breakdown and blocking time tables later in this article.

Theme Detection: We detected each store's active Shopify theme by reading the Shopify.theme JavaScript object from the store's HTML and cross-referencing with the store's meta.json endpoint. This returns the theme's theme_store_id, which we mapped to canonical theme names. Stores with a theme_store_id use a theme from the Shopify Theme Store. Stores without one use a fully custom theme.

What Is a Good Shopify Speed Score in 2026?

Before getting into the full breakdown, here is a quick orientation. These numbers come from the median performance scores across the 1,533 fully tested stores in our dataset.

Mobile (Lighthouse, throttled 4G):

Desktop (Lighthouse, unthrottled):

These ranges shift by industry. A score of 50 on mobile is above average in Beauty but below average in Automotive. The industry tables below show where each segment falls.

Performance by Industry: Who's Fastest?

The gap between the fastest and slowest industries is larger than most people expect. In our homepage dataset, Beauty & Personal Care scored 15 points lower than Automotive Parts on mobile.

Mobile Homepage Performance (median of 3 runs per store):

Desktop Homepage Performance:

A few things stand out here.

In this dataset, Beauty & Personal Care had the lowest median performance score on both mobile and desktop, across all three page types. The median mobile performance score sits at 40, with a TBT of 922ms. That TBT number is telling. Beauty stores in our dataset loaded a median of 31 third-party services per page, roughly double the median of 15 in Automotive. More on that later.

The mobile-desktop gap is consistent and large. Mobile LCP runs roughly 5-7x higher than desktop LCP across every industry. Mobile homepage LCP ranges from about 10s to 13s. Desktop stays between 1.7s and 2.1s. This gap comes from Lighthouse's mobile throttling (simulated slow 4G), but it also suggests many stores degrade sharply on constrained devices and networks.

CLS looks low in lab tests. Median CLS stays below 0.01 for most industries on mobile. Field data (CrUX) captures additional layout shifts from pop-ups and lazy-loaded content that lab tests miss.

Core Web Vitals: What Real Users Experience

Lab data tells you how a page performs under controlled conditions. CrUX tells you what actual visitors experience. The two paint different pictures.

The chart above shows an asymmetry worth understanding. For CLS, lab and field results largely agree: 88% of stores pass in Lighthouse and roughly 85% pass in CrUX. But for LCP, the gap is extreme: only 1% of stores pass the 2.5-second threshold in lab, while 86% pass in CrUX. That's because Lighthouse simulates a throttled 4G connection with CPU slowdown, which makes almost any image-heavy page exceed 2.5 seconds. Real visitors browse on faster connections with cached assets, and in our benchmark, CrUX data was collected at the origin level (the whole domain) rather than a single cold page load.

Neither measurement is wrong. They answer different questions. Lab data is useful for comparing stores against each other under identical conditions. CrUX data reflects what actual visitors experience.

For the 1,533 stores with CrUX data, here are the p75 field metrics:

Across all 1,533 stores, 68% pass all three Core Web Vitals. This is based on the CrUX overall_category field, which reflects Google's own assessment of whether a site meets all three thresholds. Individual metrics pass at higher rates among stores with sufficient data for each: LCP at 86% (1,512 stores), CLS at 85% (1,531 stores), and INP at 98% (1,362 stores). The lower INP sample count reflects stores where visitors did not generate enough interactions for Google to report a value.

The overall number hides real variation. Pet Care reaches 74%, while Beauty & Personal Care sits at 57%. The industry breakdown in the table above shows where each segment falls.

Product Pages Are Heavier Than You Think

No other public Shopify benchmark tests multiple page types. We tested homepage, product detail page (PDP), and collection page for each store. The performance gap between them is consistent across all 9 industries.

PDPs scored lowest on both mobile (44) and desktop (64). Collection pages came out fastest (49 mobile, 65 desktop). The difference is not dramatic in score terms, but the resource profiles behind those numbers tell a clearer story.

Using Consumer Goods as an example (225 stores, the largest industry slice):

In the Consumer Goods slice shown here, PDPs loaded about 10% more JavaScript than homepages. The gap ranged from 4-17% across industries. The extra JS comes from product-specific features: variant selectors, review widgets, "add to cart" logic, upsell recommendations, and BNPL payment calculators. PDPs also generated 9-25% more HTTP requests depending on the industry.

In our dataset, collection pages were generally the lightest page type. For a deeper analysis of why PDPs underperform and how to fix them, see our product page speed analysis. They're essentially filtered product grids with thumbnails, and they consistently score best across industries.

Unused JavaScript appears consistently across all three page types. The median PDP in Consumer Goods carries 971 KB of unused JavaScript. On homepage it's 929 KB. Across all industries, unused JS on PDPs ranged from 689 KB to 1,063 KB. This is often a high-impact cleanup opportunity for stores trying to improve speed.

77% of stores have an image as their LCP element. This means image optimization (proper sizing, modern formats like AVIF/WebP, preloading the hero image) is likely one of the highest-impact LCP fixes for many stores.

Theme Performance: Dawn Leads, but Context Matters

We detected the Shopify theme running on each store using the Shopify.theme object and meta.json endpoint. Here are the themes with 15 or more stores in our dataset.

The table below adds LCP and TBT detail that the chart above doesn't show.

Theme comparisons should be read directionally, especially for themes with smaller sample sizes.

Among themes with at least 15 stores in our dataset, Dawn had the lowest median mobile homepage LCP at 7,367ms. That likely reflects a lighter default setup, though observed results also depend heavily on third-party services, media weight, and customizations. But Dawn's overall performance score (55) sits below Warehouse (57) and Debut (57) because those themes have even lower TBT.

Among themes with at least 15 stores, Focal had the lowest median mobile homepage performance score at 44 and a 15.2s median mobile LCP.

Here's the thing that makes theme benchmarks tricky: the same theme performs very differently depending on the industry. Dawn scores 63 in Pet Care but only 39 in Beauty & Personal Care. The theme didn't change. What changed is the number of apps, the image weight, and the amount of custom code merchants added.

Theme choice matters, but what you load on top of it matters more.

Third-Party Services: The Strongest Performance Relationship

This is the finding we keep coming back to. The correlation between third-party service count and performance score is strong and negative: r = -0.66.

In our bucketed comparison, stores with 0-5 third-party services had a median mobile score of 74, versus 44 for stores with 21-30 services.

The heaviest third-party services by main thread blocking time (p25-p75 shows the range where the middle 50% of stores fall):

Google Tag Manager deserves special attention. It's on 90% of stores in our dataset and blocks the main thread for a median of 486ms. That's roughly half a second of JavaScript execution on every page load. GTM itself is lightweight, but the tags it fires (Facebook Pixel, Google Analytics, conversion tracking, remarketing pixels) compound quickly.

Third-party service density varies by industry. Beauty & Personal Care stores load a median of 31 third-party services. Automotive Parts stores load 15. This difference likely contributes meaningfully to the performance gap between industries. For a deeper analysis of which services block the browser the most and how to audit your store, see our third-party apps speed analysis. To apply this benchmark data to your own store, the 8-issue speed diagnostic for Shopify walks through detection criteria across all the patterns surfaced in this dataset.

The breakdown of what's loading across all stores:

In our entity classification, advertising and analytics made up roughly a third of detected third-party resources. Many of these tools are deployed sitewide, which means they often add work on page loads where they are not directly needed.

Why This Matters for Revenue

Speed numbers in isolation are just numbers. What makes them worth caring about is the connection between page performance and money.

A joint study by Deloitte and Google tested 37 brand websites across roughly 30 million user sessions. For retail sites, a 0.1-second improvement in mobile load time correlated with an 8.4% increase in conversion rate and a 9.2% increase in average order value (web.dev). For luxury brands, the same 0.1s improvement drove a 40.1% increase in product-to-basket progression.

Vodafone ran an A/B test where a 31% improvement in LCP led to 8% more sales and a 15% increase in lead-to-visit rate (web.dev). Lazada saw a 16.9% mobile conversion bump after tripling their LCP performance (web.dev).

Google recommends achieving good Core Web Vitals, noting that they align with the kind of page experience its core ranking systems seek to reward (Google Search Central). The SEO effect is rarely the clearest standalone reason to prioritize speed, but the UX and conversion upside can still be meaningful.

That context makes the numbers in this benchmark more tangible. Industries with heavier service density tend to score lower on performance, which suggests there is meaningful room for improvement. This benchmark does not estimate the revenue gain from reducing service count alone, but the external evidence points in a consistent direction. In our dataset, the slowest industries also tend to be the ones with the heaviest third-party service density, which strengthens the business case for third-party audits.

What the Data Points To

A few patterns keep surfacing across the dataset.

Third-party service count showed the strongest observed negative relationship with performance score in this dataset. Theme choice, image optimization, and code quality all matter. But the r = -0.66 correlation between service count and performance is the strongest signal in the data. A store on a slower theme with few third-party services will often outperform a store on a faster theme buried under 30 services. For stores with heavy service density, a third-party audit is often the best first diagnostic step.

Product pages appear to be the heaviest and often the least optimized page type in the dataset. Most speed optimization advice focuses on the homepage. Our data shows PDPs consistently carry more JavaScript than homepages (4-17% more depending on industry) and 689-1,063 KB of unused code at the median. For stores where product pages drive the majority of organic and paid traffic, this is where the performance gap hurts most.

Lab and field results tell different stories, especially for LCP. Nearly all stores fail the 2.5s LCP threshold in lab but pass it in the field, because Lighthouse simulates a throttled connection that penalizes image-heavy pages. For CLS, lab and field show similar pass rates (88% in lab, 85% in field). The takeaway: lab tests are useful for comparing stores under identical conditions, but CrUX gives a more realistic picture of what visitors actually experience.

The industry you're in shapes your baseline. A Beauty store running a lean setup faces a different starting point than an Automotive store, because customer expectations around visuals, social proof, and personalization drive heavier app stacks. Knowing your industry median gives you a realistic target instead of chasing a generic "good" score.

For practical optimization steps based on these findings, see our guide on how to read a Lighthouse report.

FAQ

What is a good Shopify speed score?

Across our dataset of 1,533 fully tested stores, the median mobile Lighthouse performance score is around 47. In several industries, scores above 55 place a store in the faster-performing segment. On desktop, the median is around 65. These are lab scores measured under throttled conditions. CrUX often shows much better LCP than Lighthouse because real visitors use faster connections than what the lab simulates. For CLS, lab and field results are closer, with both showing high pass rates (88% in lab, 85% in field).

Why is my mobile score so much lower than desktop?

Lighthouse mobile tests simulate a mid-range phone on a throttled 4G connection. Desktop tests run without network or CPU throttling. This creates a consistent gap: mobile LCP runs roughly 5-7x higher than desktop LCP across all 9 industries in our data. The gap is normal, but it does highlight how much mobile optimization matters.

How many apps is too many for Shopify?

Our data shows a strong negative correlation (r = -0.66) between the number of third-party services on a page and performance. Stores with 0-5 third-party services score around 74 on mobile. At 16-20 services, the median drops to 51. Above 30, it falls to 36. The number that's "too many" depends on which services you're running and how they load, but the data is clear that every addition has a cost.

Does my Shopify theme affect speed?

Yes, but less than most people think. Dawn (the most popular free theme) scores a median 55 on mobile. Focal (the slowest in our dataset with 15+ stores) scores 44. That's an 11-point gap. But the same theme can score 63 in one industry and 39 in another depending on what apps and customizations are layered on top. The theme sets the baseline. The app stack determines where you actually land.

Does the number of products in my catalog affect speed?

Not meaningfully. Across the 1,533 stores in our dataset, the correlation between product count and mobile performance score is near zero (r = -0.03 for homepage, r = -0.08 for collection pages). Stores with 50 products scored about the same as stores with 5,000. That's because Shopify paginates collection pages (typically 24-48 products per page), so catalog size doesn't change how much content renders on any given page load. What does affect speed is the number of third-party services, theme weight, and media on the page itself.

What are Core Web Vitals and why do they matter?

Core Web Vitals are three metrics Google uses to measure real-world page experience: LCP (loading speed), CLS (visual stability), and INP (interactivity). In our dataset, 68% of Shopify stores pass all three. Google strongly recommends achieving good Core Web Vitals and says they align with the kind of page experience its core ranking systems seek to reward (Google Search Central), but their bigger impact is on conversion rates. A Deloitte and Google study found that a 0.1-second speed improvement led to an 8.4% conversion increase for retail sites (web.dev).

What's Next

For more on Shopify speed optimization, see our complete site speed guide and site speed best practices.

This overview covers the cross-industry findings. We're publishing individual deep-dive articles for each of the 9 industries, covering industry-specific app stacks, theme usage patterns, and detailed resource breakdowns. The first is our beauty industry deep-dive, which breaks down why Beauty & Personal Care ranks last and where the weight comes from. Our Fashion & Apparel deep-dive covers why Fashion carries the heaviest image payload and how return-prevention tools like Kiwi Sizing and Route shape the performance profile.

If you want to see where your store stands, try our store speed test. It runs a PageSpeed Insights test on your store and shows your Core Web Vitals alongside a platform-specific fix checklist. For a multi-dimensional audit covering speed plus 27 other SEO, UX, trust, and CRO checks, see our Shopify audit checklist, or run the automated Shopify store audit on your own store. For a walkthrough of what the speed tool reports and how to interpret the results, see how the speed test works.

Download the Benchmark Data

• Summary dataset (CSV) - one row per store/page/strategy, 9,173 rows
• Run-level dataset (CSV) - individual test runs, 27,467 rows
• Entity dataset (CSV) - third-party entities per store/page/strategy, 225,396 rows
• Data dictionary (CSV)
• README / Methodology

All public files use anonymized store IDs (SHA-256 hash). Version 1.1, licensed under CC BY 4.0.

v1.1 update (April 2026): All performance metrics, blocking times, and correlation values recalculated using median of 3 Lighthouse runs per store (previously single run). Differences are minor and do not affect conclusions. Added entity-level CSV with per-service blocking data for full reproducibility. See the CHANGELOG in the dataset for details.