How much does 0.1 seconds matter to a shopper holding a phone? A study commissioned by Google and carried out by Deloitte and the agency 55 measured it across 37 brand sites and more than 30 million user sessions, using data collected at the end of 2019. When mobile load times improved by 0.1 seconds, retail conversion rates rose 8.4% and retail shoppers spent 9.2% more. Travel sites saw conversions rise 10.1% from the same change.
Scope of the Deloitte Study
The researchers tracked mobile load times hour by hour for 30 days and compared them with what visitors did next. The effect appeared at every stage of the purchase path. On lead generation sites, the share of users who reached the form submission page rose 21.6%, and page views rose 7%. On luxury retail sites, the move from a product page to the basket rose 40.1%.
The study measured small changes on sites that were already reasonably quick, so even modest gains in speed showed up in sales. A buyer who has already chosen a product is still sensitive to delay at each page that follows.
The Core Web Vitals Benchmarks
Google publishes 3 measures, called Core Web Vitals, that describe how a page feels to a visitor. Largest Contentful Paint records when the main content appears, and the target is 2.5 seconds or less. Interaction to Next Paint measures how quickly the page responds to a tap or click, with a target of 200 milliseconds. The third measure scores how much the layout jumps while loading, and a score of 0.1 or less is good. Each target applies to the 75th percentile of real visits.
More than 50% of sites miss at least 1 target on mobile. The HTTP Archive’s 2025 Web Almanac, based on July 2025 data, found that 48% of websites passed all 3 measures on mobile and 56% passed on desktop. Both figures improved on 2024, when the shares were 44% and 55%.
Largest Contentful Paint on Mobile
Of the 3 measures, Largest Contentful Paint is the one mobile sites fail most often. The Web Almanac found that 62% of sites had a good score on mobile, compared with 77% for responsiveness and 81% for layout stability. For a store, the largest element is usually the main product photo, and it is often the heaviest file on the page.
Shrinking that image has a direct effect. A photo exported at full camera resolution can be several times larger than the screen needs, and serving it in a modern compressed format at the displayed size cuts the wait without visible loss. Loading the main image first, ahead of reviews and recommendation widgets further down the page, helps as well. Lazy loading, which delays images lower on the page until the visitor scrolls toward them, frees bandwidth for the main photo.
Server Processing Time
Before the browser can draw anything, the server has to build the page. On a store that means querying the product database, checking stock and applying prices, and on an underpowered server that work alone can use up much of the 2.5-second target.
A content delivery network sends images from a location near the visitor, and a lighter theme reduces the work each page requires. Moving the store to fast VPS hosting gives it its own processing power and memory for pages that must be built fresh, such as the cart.
Page Weight in 2025
The same Almanac’s page weight chapter measured what a typical home page carries. On mobile in 2025, the median home page included 911 KB of images and 632 KB of JavaScript. Images take time to download, and JavaScript also has to be processed by the phone after it arrives, which is why script-heavy pages can feel unresponsive after they look finished.
A small business home page rarely needs that much. A main photo and a contact form can be delivered in a fraction of the median weight, and each kilobyte removed is one the visitor’s phone never has to fetch.
Responsiveness at the Moment of Purchase
Interaction to Next Paint matters most at the moments a buyer acts. A visitor who taps “Add to cart” and sees nothing happen for 0.5 seconds may tap again, which can add 2 items to the cart or trigger an error. Heavy scripts for chat widgets and tracking tags compete with that tap for the phone’s processor.
Layout movement creates a related problem. When a banner or an image loads late and pushes the page down, the button a visitor was about to press moves, and the tap lands on something else. At the checkout stage that mistake can send the visitor to the wrong page or cancel a step in the payment form.
The Desktop and Mobile Gap
Phones have slower processors and less reliable connections than office computers, which is consistent with the 8-point gap between desktop and mobile pass rates in the Web Almanac. A site tested on an office computer with a fast connection can feel quick to its owner and slow to a customer on a train. Owners should test their own pages on a mid-range phone over a mobile connection, since that is closer to the conditions many buyers face.
Field data from real visitors gives a more accurate picture than a single lab test. Free tools that report Core Web Vitals from real-user data show how the site performs across devices and connection types, and they flag which pages fail.
A Priority List for Small Sites
The order of fixes matters for a business with a small budget. Compressing and resizing images comes first, because cutting image file sizes is cheap and can produce a large improvement in Largest Contentful Paint. Removing scripts that no longer serve a purpose comes next, since each one adds weight and competes for the processor. Fonts are a further item to check, because each custom typeface is another file the page must download before text appears in its final form. Server and network changes follow once the pages themselves are lean.
Each change should be measured. Recording the Core Web Vitals scores and the conversion rate before and after a fix shows if the improvement reached customers, and which pages still need work.
Gains From the First Fixes
The gains in the Google-commissioned speed study came from shaving 0.1 seconds off sites that were already reasonably fast. Many small business sites have product pages that take several seconds to show their main image, and for them the first fixes are the cheapest and bring the largest change. On those sites, the final 0.1 seconds belongs at the bottom of the list.
