Playbooks
How much does page speed actually affect ecommerce conversion rates?
The data is consistent and old enough to be well-tested: a 1-second delay costs the average ecommerce store roughly 7% of conversions (Deloitte, 37 retail sites), and separate Akamai/SOASTA analysis of about 10 billion visits found a 1-second delay can cut conversions by up to 22%. On the gains side, the Google/Deloitte "Milliseconds Make Millions" study — 37 brands, over 30 million mobile sessions, 30 days of hour-by-hour monitoring — found a 0.1-second mobile speed improvement lifted retail conversion 8.4% and average order value 9.2%. Pages loading in 1–2 seconds see the highest average conversion rates (around 3.05%), and sites loading in 1 second convert roughly 2.5x higher than sites loading in 5 seconds. These figures are still being cited as current benchmarks in 2025–2026 recap research, not treated as outdated.
Last updated Aug 21 · 16 min read
The core numbers everyone cites
Three figures recur across virtually every serious page-speed-and-conversion resource, and it's worth being precise about where each comes from. Deloitte's study of 37 retail sites found a 1-second delay costs the average ecommerce store about 7% of conversions (btng.studio/articles/page-speed-ecommerce-conversions-guide/; note.com/masakazu_urabe/n/nef7804fa20f9). Separately, Akamai/SOASTA analysis covering roughly 10 billion visits found a 1-second delay can cut conversions by up to 22% — a notably larger figure, likely reflecting a different sample, methodology, and time period, which is why both numbers get cited together rather than treated as interchangeable. Multiple sources also report that sites loading in 1 second convert roughly 2.5x higher than sites loading in 5 seconds (nitropack.io/blog/how-page-speed-affects-conversion/; fleexy.dev/blog/how-page-speed-affects-conversion-rates-study/, as cited alongside the primary sources reviewed here).
The consistent message across all three: the relationship between load time and conversion isn't linear and gentle — it's steep, especially in the first few seconds, which is exactly the window most ecommerce checkout and product-page interactions happen in.
The Google/Deloitte "Milliseconds Make Millions" study, in detail
This is the single most detailed and most frequently cited study in this space, and it's worth walking through its actual methodology rather than just quoting the headline number. The study was commissioned by Google and conducted by 55 and Deloitte, studying 37 leading European and American brand sites and collecting data from over 30 million user sessions, with mobile load times monitored hour-by-hour across a 30-day window (web.dev/case-studies/milliseconds-make-millions). That's a substantially larger and more rigorously monitored dataset than most web-performance claims get to cite.
The headline retail findings: a 0.1-second improvement in mobile site speed lifted retail conversion by 8.4% and average order value by 9.2%. The study also broke down funnel-stage impact specifically: the same 0.1-second improvement produced a 3.2% increase in progression from the product listing page to the product detail page, and a 9.1% increase in progression from the product detail page to the add-to-basket page — meaning the conversion benefit compounds at exactly the funnel stages where ecommerce sites lose the most volume. Beyond retail, the same study found travel conversions increased by 10.1% (with average order value up 1.9%), luxury brand page views per session rose 8.6%, and lead-generation page bounce rate improved by 8.3% — evidence the speed-conversion relationship isn't retail-specific but holds across several verticals with different purchase behaviors.
Mobile bounce rate and the 3-second cliff
Google/SOASTA's machine-learning analysis of mobile ecommerce sessions found bounce probability rises 32% as load time goes from 1 second to 3 seconds, and climbs to 90% by 5 seconds (eggknite.com/reports/web-performance-2026; note.com/masakazu_urabe/n/nef7804fa20f9). Google's own DoubleClick research separately found that 53% of mobile visits abandon a page taking longer than 3 seconds to load — a figure independently corroborated by more recent 2026 Shopify-focused performance guidance citing the same "53% abandon past 3 seconds" statistic as still current.
The practical read: 3 seconds functions as something close to a hard ceiling for mobile ecommerce, not a soft guideline. Past that point, the majority of visitors are gone before they ever see the product they clicked through for.
Amazon and Walmart's own reported numbers
Two of the largest ecommerce operators in the world have publicly attributed real revenue to page-speed changes, and both figures are widely repeated because they come directly from the companies rather than third-party research firms. Amazon has reported that every 100 milliseconds of added latency cost it about 1% in sales; Walmart reported that each 1-second improvement in load time raised conversions by about 2% (cybertising.com/blog/need-for-speed-milliseconds-make-millions; portent.com/blog/analytics/research-site-speed-hurting-everyones-revenue.htm). At the scale either company operates, even fractional percentage shifts translate into very large absolute revenue swings — which is a large part of why both companies have invested heavily and publicly in performance engineering over the years.
What "good enough" load time actually looks like in 2026
Current guidance converges on under 3 seconds as the acceptable ceiling and 1–2 seconds as the actual target, with pages loading in that 1–2 second window seeing the highest average conversion rates, reported around 3.05% (queue-it.com/blog/ecommerce-website-speed-statistics/; hostinger.com/tutorials/gtmetrix-for-testing-websites-speed/; mirasvit.com/blog/page-speed-conversion-rate.html). For Core Web Vitals specifically — the metrics Google uses as a ranking signal — the current thresholds are Largest Contentful Paint (LCP) under 2.5 seconds, Interaction to Next Paint (INP) under 200 milliseconds, and Cumulative Layout Shift (CLS) under 0.1, all measured on real mobile visits rather than lab conditions.
For a Shopify or WooCommerce store specifically, current optimization guidance identifies a short list of highest-impact, lowest-effort fixes: compressing and properly sizing images, removing unused apps/plugins that load unnecessary scripts, and explicitly setting image dimensions to prevent layout shift — three changes reported to move most stores a meaningful distance toward passing Core Web Vitals thresholds without a full technical rebuild.
Practical examples
A mid-size fashion ecommerce store running Shopify notices its product listing pages load in about 4.5 seconds on mobile, well past the 3-second abandonment threshold cited in Google's DoubleClick research. After compressing product images, removing three unused marketing apps that were injecting extra scripts, and setting explicit width/height attributes on all product images to eliminate layout shift, load time drops to roughly 2.3 seconds — landing inside the 1–2 second window associated with the highest reported conversion rates in the reviewed benchmark data.
A B2B ecommerce operation running a major seasonal sale notices checkout page load times creeping up under the traffic spike, from a normal 1.8 seconds to over 4 seconds during peak hours. Following the practical guidance to re-test site speed daily during high-traffic periods (rather than relying on a pre-sale test alone), the team catches and fixes a CDN caching misconfiguration mid-sale rather than losing days of peak-traffic conversions to a preventable slowdown.
Data and evidence
– Deloitte (37 retail sites): 1-second delay costs the average ecommerce store about 7% of conversions.
– Akamai/SOASTA (~10 billion visits): 1-second delay can cut conversions by up to 22%.
– Google/Deloitte "Milliseconds Make Millions" (37 brands, 30M+ sessions, 30-day mobile monitoring): 0.1s mobile speed improvement → 8.4% retail conversion lift, 9.2% average order value lift, 3.2% lift in listing-to-detail-page progression, 9.1% lift in detail-page-to-add-to-basket progression.
– Same study, other verticals: travel conversions +10.1% (AOV +1.9%), luxury page views/session +8.6%, lead-gen bounce rate improved 8.3%.
– Google/SOASTA machine-learning analysis: mobile bounce probability rises 32% going from 1s to 3s load time, 90% by 5 seconds.
– Google DoubleClick research: 53% of mobile visits abandon a page taking over 3 seconds to load.
– Amazon: every 100ms of added latency reportedly costs about 1% in sales.
– Walmart: each 1-second load-time improvement reportedly raised conversions by about 2%.
– Pages loading in 1–2 seconds see the highest average conversion rates, reported around 3.05%.
– Sites loading in 1 second convert roughly 2.5x higher than sites loading in 5 seconds, per multiple industry sources.
– Current Core Web Vitals thresholds: LCP under 2.5s, INP under 200ms, CLS under 0.1, measured at the 75th percentile of real mobile visits.
– These same Deloitte/Google/Akamai/SOASTA figures continue to be re-cited in 2025–2026 recap research as still-current benchmarks rather than superseded ones (eggknite.com/reports/web-performance-2026; cazyweb.com/research/website-performance-and-conversion-evidence/).
Comparisons
GTmetrix vs. Google PageSpeed Insights. GTmetrix combines Google PageSpeed Insights and YSlow scoring into a single report, offering a fuller picture; PageSpeed Insights is Google's own free tool and the direct source of official Core Web Vitals field data — many practitioners use both together rather than choosing one exclusively (debugbear.com/software/gtmetrix-speed-test; amasty.com/blog/gtmetrix-overview-and-user-guide/, as referenced in the reviewed guides).
Mobile vs. desktop impact. The most detailed available study (Milliseconds Make Millions) specifically measured mobile sessions, and separately, mobile bounce-rate research (the 32%-to-90% bounce curve) is mobile-specific — mobile ecommerce traffic is generally the more speed-sensitive segment to optimize for first, given both the larger dataset behind the mobile findings and mobile's typically higher share of ecommerce traffic overall.
1-second load time vs. 5-second load time. A roughly 2.5x conversion-rate gap between these two benchmarks, cited across multiple industry sources, is one of the starkest single comparisons available and a useful anchor for justifying performance investment to a non-technical stakeholder.
Real-world use cases
– A small local ecommerce brand preparing for a seasonal promotional push runs a free speed test a week before launch, finds its checkout flow loads in 3.8 seconds on mobile, and prioritizes image compression and app cleanup specifically to get under the 3-second abandonment threshold before traffic increases.
– A CRO specialist pitching performance work to a client uses the Amazon (1% sales loss per 100ms) and Walmart (2% conversion gain per 1-second improvement) figures as concrete, company-attributed evidence when a client is skeptical that speed work translates to revenue, since these numbers come directly from the companies rather than a third-party study.
– A web development agency builds page-speed monitoring into an ongoing retainer specifically because a store's load time can silently degrade after every new app, tracking pixel, or content update, not just at initial launch — reflecting the daily-recheck-during-sales-events guidance from practical speed-testing sources.
– A Shopify store owner without technical staff works through the three highest-impact, lowest-effort fixes (image compression, removing unused apps, setting image dimensions) as a first pass before considering a full technical audit or developer engagement.
Common mistakes
– Testing site speed once at launch and never again, missing gradual degradation caused by newly added apps, tracking scripts, or content over time.
– Treating "passes Lighthouse" or a high lab score as proof the site is fast for real users, when field data (real mobile visits) can diverge from lab conditions.
– Optimizing desktop performance while leaving mobile — the segment most of the cited research specifically measured — comparatively neglected.
– Assuming these page-speed-conversion studies are outdated because some are several years old, when 2025–2026 recap research continues to re-cite the same figures as current, durable benchmarks.
– Chasing marginal millisecond improvements on an already-fast site (diminishing returns) while ignoring a checkout flow still loading past 4–5 seconds — the bigger lever is almost always fixing what's clearly slow first.
– Not re-testing during high-traffic events (sales, promotions), when load spikes can silently push a normally acceptable load time past the abandonment threshold exactly when conversion stakes are highest.
Data and evidence — additional context on tools
Google PageSpeed Insights, GTmetrix, WebPageTest, and Pingdom are the most commonly recommended free tools for ecommerce site owners checking speed without a dedicated performance engineering team (per the practical testing guides reviewed). None require a paid subscription to get a baseline reading, making an initial audit accessible to even the smallest store before deciding whether further investment is warranted.
Best practices
– Test on mobile specifically, since mobile is both the more speed-sensitive segment in the available research and typically the larger share of ecommerce traffic.
– Prioritize getting under the 3-second mobile load threshold before chasing marginal gains on an already-reasonably-fast site.
– Compress and properly size images, remove unused apps/scripts, and set explicit image dimensions as the first three fixes, since these are consistently identified as the highest-impact, lowest-effort changes.
– Track Core Web Vitals (LCP, INP, CLS) against their current thresholds using real-user field data (Google Search Console's report), not just lab scores.
– Re-test site speed frequently during high-traffic periods like sales events, when load spikes are most likely and most costly.
– Use the funnel-stage breakdown from the Milliseconds Make Millions study as a mental model: speed improvements compound most heavily at listing-to-detail and detail-to-cart transitions, so prioritize those specific pages if resources are limited.
Key takeaways
– A 1-second delay costs roughly 7% of conversions per Deloitte's study, and up to 22% per Akamai/SOASTA analysis of billions of visits — these are large, not marginal, effects.
– The Google/Deloitte Milliseconds Make Millions study found a 0.1-second mobile speed improvement lifted retail conversion 8.4% and average order value 9.2%, with the biggest funnel-stage gains at listing-to-detail and detail-to-cart transitions.
– Mobile bounce probability rises steeply between 1 and 3 seconds of load time, and 53% of mobile visits abandon pages past the 3-second mark.
– Amazon (1% sales loss per 100ms) and Walmart (2% conversion gain per 1-second improvement) are the two most citable company-reported figures for making a revenue-based case internally.
– These figures remain the current industry benchmark as of 2025–2026 recap research, not outdated statistics superseded by newer data.
Relevant tools.scult.in resources
Before making any changes, run your store through the Website Speed Test tool to get a current baseline reading on mobile and desktop, and re-check it after implementing any of the fixes covered above to confirm the improvement actually landed.
If your store's load time is well past the 3-second mobile threshold and the fix requires more than image compression and app cleanup — CDN configuration, code splitting, third-party script auditing — that deeper technical work is exactly what SCULT.IN's web development team handles for ecommerce clients trying to close the gap between "passes a lab test" and "actually converts better in production."
If this is a gap worth closing properly rather than patching once, that is exactly the kind of work our team handles.
For a related, free starting point, try the AI Visibility Checker.
Frequently asked questions
How much does a 1-second delay cost an ecommerce store in conversions?
Deloitte's study of 37 retail sites found about a 7% conversion cost; separate Akamai/SOASTA analysis of ~10 billion visits found up to 22%.
What conversion lift comes from a 0.1-second mobile speed improvement?
The Google/Deloitte study found an 8.4% retail conversion lift and 9.2% average order value lift.
What's considered a good load time for an ecommerce site in 2026?
Under 3 seconds is the acceptable ceiling; 1–2 seconds is the target, with that range showing the highest average conversion rates (around 3.05%).
How does load time affect mobile bounce rate?
Bounce probability rises 32% going from 1 to 3 seconds and reaches 90% by 5 seconds, per Google/SOASTA analysis.
What percentage of mobile visitors abandon a slow-loading page?
53% abandon a page taking longer than 3 seconds to load, per Google's DoubleClick research.
How much revenue did Amazon attribute to page speed?
Amazon reported every 100ms of added latency cost about 1% in sales.
How much did Walmart's conversions improve from faster load times?
Walmart reported roughly a 2% conversion increase per 1-second load-time improvement.
What's the conversion-rate gap between a 1-second and a 5-second site?
Multiple industry sources cite roughly a 2.5x higher conversion rate for 1-second sites versus 5-second sites.
What is the Deloitte "Milliseconds Make Millions" study?
A Google-commissioned study by 55 and Deloitte covering 37 brand sites and over 30 million mobile sessions monitored hourly over 30 days, measuring the conversion impact of mobile speed improvements.
Are these page-speed studies still relevant given how old some are?
Yes — 2025–2026 recap research continues to re-cite the same Deloitte, Google, and Akamai figures as current, durable benchmarks rather than treating them as superseded.
What is a good LCP (Largest Contentful Paint) score for ecommerce?
Under 2.5 seconds, measured at the 75th percentile of real mobile visits.
What is a good INP (Interaction to Next Paint) score?
Under 200 milliseconds.
What is a good CLS (Cumulative Layout Shift) score?
Under 0.1.
Does site speed affect Google Search rankings for an ecommerce site?
Core Web Vitals are used by Google as a ranking signal, though the reviewed sources focus primarily on the conversion-rate relationship rather than a specific ranking-impact figure.
What free tools can I use to test my ecommerce site's speed?
Google PageSpeed Insights, GTmetrix, WebPageTest, and Pingdom are the most commonly recommended free options.
Is GTmetrix or PageSpeed Insights better for testing?
GTmetrix combines PageSpeed Insights and YSlow scoring in one report; PageSpeed Insights is Google's own tool and the source of official Core Web Vitals field data — many practitioners use both.
How often should I test my store's speed?
Regularly, and especially daily during high-traffic sales periods when load spikes are most likely.
What page-speed fixes have the highest impact for the least effort on Shopify?
Compressing/sizing images, removing unused apps, and setting explicit image dimensions to prevent layout shift.
Does page speed affect average order value, not just conversion rate?
Yes — the Milliseconds Make Millions study found a 0.1-second improvement lifted average order value by 9.2% for retail.
Does page speed matter differently at different funnel stages?
Yes — the same study found a 3.2% lift in listing-to-detail-page progression and a 9.1% lift in detail-page-to-add-to-basket progression from the same 0.1-second improvement.
How do I test my ecommerce site's speed for free?
Run your site's URL through Google PageSpeed Insights or GTmetrix, both of which are free and provide a numeric score plus specific recommendations.
How do I improve Shopify site speed for conversions?
Start with image compression and sizing, remove unused apps, set explicit image dimensions, and prioritize load-critical content before progressively enhancing with JavaScript.
How do I check site speed before a major sale event?
Test both average and simulated peak-load conditions, and re-test daily in the run-up to and during the event rather than relying on a single pre-event check.
How do I know which page on my site to optimize first?
Prioritize the pages tied to the funnel stages the Milliseconds Make Millions study found most speed-sensitive: product listing and product detail pages leading into add-to-cart.
How do I measure real-user speed rather than just a lab test score?
Use Google Search Console's Core Web Vitals report, which reflects real visitor field data rather than a synthetic lab test.
How do I prevent layout shift (CLS) on product pages?
Set explicit width and height attributes on all images and embedded content so the browser reserves space before the asset loads.
Why is my slow-loading website losing sales?
Given the cited Deloitte and Akamai figures, even modest additional load-time delays can measurably reduce conversion, particularly once a page crosses the roughly 3-second mobile threshold.
Why is my mobile checkout bounce rate so high?
If checkout load time is pushing past 3 seconds, that aligns with the documented steep rise in mobile bounce probability in that range.
Why did my conversion rate drop after a site redesign?
A common cause is added scripts, larger unoptimized images, or new apps injected by the redesign pushing load time past the 1–2 second range associated with the highest conversion rates — worth re-testing speed immediately after any redesign.
Why does my site score well on a lab speed test but still feel slow to real visitors?
Lab and field data can diverge; check your Core Web Vitals report in Google Search Console for real-user data rather than relying solely on a synthetic lab score.
Is the Deloitte/Google page speed study still valid for today's JavaScript-heavy ecommerce sites?
2025–2026 recap research continues to treat its figures as current benchmarks, though the underlying study itself was conducted on the site architectures common at the time it ran; the directional relationship (faster = more conversions) is treated as durable regardless.
Mobile vs. desktop — which matters more for page speed optimization?
Mobile is the more extensively studied and typically more speed-sensitive segment in the available research, making it the priority if resources are limited.
GTmetrix vs. PageSpeed Insights vs. WebPageTest — which should I use?
All three are commonly recommended free options; GTmetrix combines PageSpeed Insights and YSlow scoring, while WebPageTest offers more granular waterfall-level diagnostics — many practitioners use more than one.
1-second vs. 5-second load time — how different is the actual conversion outcome?
Roughly a 2.5x difference in conversion rate, according to multiple industry sources reviewed.
Does the Amazon/Walmart data apply to a small independent ecommerce store too?
The directional relationship (faster load time, higher conversion) is treated as broadly applicable, though the specific percentage figures reported by Amazon and Walmart reflect their own scale and traffic patterns rather than a universal constant.
My site passes Lighthouse but conversions are still low — what else could be wrong?
Check real-user field data via Search Console, since lab scores don't always reflect actual visitor experience, and also review non-speed factors (pricing, trust signals, checkout friction) that speed alone doesn't address.
My mobile bounce rate is high but desktop looks fine — what should I check first?
Load time specifically on mobile devices and networks, since the steep bounce-rate curve documented in Google/SOASTA's research is a mobile-specific finding.
My site was fast at launch but has gotten slower over time — why?
Newly added apps, tracking scripts, or content (especially unoptimized images) are the most common causes of gradual speed degradation after launch.
My checkout page slows down specifically during sales — what's the likely cause?
Traffic-driven load on the server or CDN, or a caching misconfiguration that isn't apparent under normal traffic — worth testing under simulated peak load before a major sale, not just under average conditions.
Is it worth hiring a developer or agency to fix ecommerce site speed, or can I DIY it?
The three highest-impact, lowest-effort fixes (image compression, removing unused apps, setting image dimensions) are commonly DIY-able even without deep technical skill; a full Core Web Vitals audit and deeper technical fixes (code splitting, CDN configuration) more often benefit from developer or agency involvement.
Is a professional site speed audit worth the cost for a small ecommerce store?
Given the documented conversion impact (up to 22% loss per 1-second delay in some analyses), even a modest-cost audit can pay for itself quickly if it identifies fixable, high-impact slowdowns — a reasonable investment for a store with sufficient traffic and revenue to buy back the lost conversion.
Should I prioritize page speed work over other CRO tactics (e.g., copy, offers, design)?
Page speed is generally a foundational prerequisite — a fast page with weak copy still underperforms, but a slow page can suppress conversion regardless of how good everything else is, so it's reasonable to fix speed issues before or alongside other CRO work rather than after.
How much should I expect to pay for ongoing site speed monitoring?
Free tools (PageSpeed Insights, GTmetrix, Search Console) cover baseline monitoring at no cost; paid monitoring services or agency retainers add cost but also add continuous tracking and expert interpretation a store owner may not have time for.
Is it worth investing in a CDN specifically for page speed?
A CDN is commonly recommended as part of a broader speed strategy, particularly for stores with geographically distributed traffic, though it addresses only part of the load-time equation (network latency) rather than issues like unoptimized images or excessive third-party scripts.
What's the ROI case for page speed work to a skeptical stakeholder?
Lead with company-reported, revenue-attributed figures (Amazon's 1% sales loss per 100ms, Walmart's 2% conversion gain per 1-second improvement) since these come directly from the companies rather than a third-party study, making them harder to dismiss as academic.
Should I prioritize speed fixes on high-traffic pages or low-traffic pages first?
High-traffic, high-funnel-value pages (home, category listing, product detail, cart, checkout) first, since that's where the documented conversion impact compounds most, per the funnel-stage breakdown in the Milliseconds Make Millions study.
Is it worth switching ecommerce platforms just for better page speed?
Platform migration is a significant undertaking best justified by multiple factors, not speed alone; most stores can achieve meaningful speed gains within their current platform through the fixes outlined here before considering a full migration.
How do I justify a page speed optimization budget to leadership?
Combine the company-reported figures (Amazon, Walmart) with your own store's current load time and traffic volume to estimate a rough revenue-at-stake figure specific to your business, rather than relying on generic industry statistics alone.
Does page speed optimization ever hit diminishing returns?
Yes — once a site is solidly within the 1–2 second target range, further millisecond-level optimization typically yields much smaller marginal conversion gains than fixing a site still loading in 4–5+ seconds.
What's the single highest-priority speed fix for a small ecommerce store that's never done this before?
Get mobile load time under 3 seconds first (the documented abandonment cliff), primarily through image compression, removing unused apps/scripts, and fixing layout shift — before pursuing any more advanced or marginal optimization.
Sources
- https://www.btng.studio/articles/page-speed-ecommerce-conversions-guide/
- https://note.com/masakazu_urabe/n/nef7804fa20f9
- https://queue-it.com/blog/ecommerce-website-speed-statistics/
- https://www.cloudflare.com/learning/performance/more/website-performance-conversion-rates/
- https://www.eggknite.com/reports/web-performance-2026
- https://portent.com/blog/analytics/research-site-speed-hurting-everyones-revenue.htm
- https://www.debugbear.com/software/gtmetrix-speed-test
- https://www.hostinger.com/tutorials/gtmetrix-for-testing-websites-speed/
- https://cazyweb.com/research/website-performance-and-conversion-evidence/
- https://web.dev/case-studies/milliseconds-make-millions
Need this built into your business?
The free tools and prompts on this site handle the small, solved problems. If what you need is bigger — web development, built and maintained for you — that's Scult's day job.
