Latency Cost Case Studies: Real Data from Amazon, Google, Walmart, and More

These are the most frequently cited studies on the revenue impact of latency. We have consolidated them here with source attribution, methodology details, and modern context. The core finding has been consistent for 20 years: speed equals revenue.

20 Years of Consistent Evidence

2006 Amazon2006 Google2012 Walmart2020 Deloitte + Google2017 COOK2012 AutoAnything2016 Mobify2009 Bing (Microsoft)2009 Shopzilla

From Amazon in 2006 to Deloitte in 2020, across search engines, ecommerce platforms, and mobile sites, the data consistently shows the same pattern: faster pages generate more revenue. The relationship has held through desktop-only browsing, the mobile revolution, and the current era of Core Web Vitals.

All Case Studies

Amazon

2006

Every 100ms of added latency cost 1% of sales

Methodology

Internal A/B test. Amazon deliberately slowed pages for a subset of users and measured the revenue impact.

Source

Greg Linden presentation, widely cited by Google and industry

Modern Context

At Amazon's current estimated $600B+ annual revenue, 1% equals $6B+. The most frequently cited latency stat in the industry.

Google

2006

Adding 0.5 seconds to search results reduced traffic by 20%

Methodology

Marissa Mayer presented data from Google's experiments with intentionally slowed search results.

Source

Marissa Mayer, Web 2.0 Summit presentation

Modern Context

This finding helped establish the principle that even sub-second delays have measurable user behavior impacts. Google's entire infrastructure philosophy is built around speed.

Walmart

2012

Every 1-second improvement in page load increased conversions by 2%

Methodology

WalmartLabs engineering team measured conversion rates across different page load times during a performance optimization initiative.

Source

WalmartLabs engineering blog

Modern Context

Also found that for every 100ms of improvement, incremental revenue grew by up to 1%. At Walmart's scale, these percentages translate to billions.

Deloitte + Google

2020

0.1 second of mobile speed improvement increased retail conversions by 8.4%

Methodology

Joint study analyzing real-world performance data across multiple retail sites. Controlled for other variables.

Source

Deloitte Digital, 'Milliseconds Make Millions' report

Modern Context

The most recent large-scale study. Notably, the impact was even larger than earlier estimates. Also found 8.4% lift in page views and 5.7% increase in time on site.

COOK

2017

Reducing load time by 0.85 seconds increased conversion rate by 7%

Methodology

UK meal delivery service measured conversion rates before and after a performance optimization project.

Source

Web performance case study, Nerd Summit presentation

Modern Context

One of the clearest small-business case studies. The 0.85s improvement cost relatively little engineering effort but delivered a measurable 7% conversion lift.

AutoAnything

2012

Cutting page load time by 50% increased sales by 12-13%

Methodology

Auto parts ecommerce retailer measured sales before and after a major site speed optimization.

Source

Radware performance report

Modern Context

Clean A/B result from a mid-market ecommerce company. The 12-13% sales increase came from halving their load time, not eliminating it entirely.

Mobify

2016

Every 100ms decrease in homepage load speed increased session conversion by 1.11%

Methodology

Mobile commerce platform analyzed conversion data across many client sites with varying load times.

Source

Mobify engineering team analysis

Modern Context

Also calculated the annual revenue impact: $380,000 per year from a 100ms improvement. One of the few studies that gives a specific dollar figure per millisecond.

Bing (Microsoft)

2009

A 2-second slowdown reduced revenue per user by 4.3%

Methodology

Microsoft ran controlled experiments on Bing search, deliberately slowing pages for test groups.

Source

Eric Schurman, Microsoft presentation at Velocity conference

Modern Context

Also found that a 2-second delay reduced clicks by 3.75% and queries by 1.8%. User satisfaction decreased by 3.8%. Effects persisted even after speed was restored.

Shopzilla

2009

Reducing load time from 5s to 1.2s increased revenue by 12% and page views by 25%

Methodology

Shopping comparison engine completed a major site rebuild focused on performance.

Source

Philip Dixon, VP Engineering, Shopzilla

Modern Context

One of the most dramatic transformations. Also reduced server costs by 50% because faster pages required fewer servers to handle the same traffic.