Every 100ms Costs You Money.
Calculate How Much.
The cost of latency, in one line: Amazon found every 100ms of added page delay cost about 1% of sales (Greg Linden, 2006), and Akamai's April 2017 retail report, across about 10 billion visits, found a 100ms delay can cut conversion rates by 7%. Use the calculator below to put a monthly figure on your own site.
Sources re-read November 2026: Linden's “Make Data Useful” talk and Akamai's 2017 press release.
Our model: about 7% fewer conversions for each second above 1s. That is far milder than Akamai/SOASTA 2017, which measured 7% from a 100ms delay. Full calculator with more options
Where does the “every 100ms = 1% of sales” figure come from?
The Amazon stat traces to Greg Linden, a former Amazon engineer, who shared it on his blog Geeking with Greg and in his 2006 Stanford presentation “Make Data Useful” (the “Speed Matters” slide). In Amazon’s A/B tests, deliberately delaying pages in 100ms increments cost roughly 1% of sales per 100ms. The separate ecommerce conversion figures come from the Akamai/SOASTA State of Online Retail Performance Report (Spring 2017), which found a 100ms delay could cut conversion rates by 7% and a two-second delay could raise bounce rates by 103%.
The Headline Numbers
100ms = 1% of sales
Internal A/B test showed every 100ms of added latency cost 1% of sales revenue.
1s = +2% conversions
For every 1-second improvement in page load, Walmart saw a 2% increase in conversions.
0.5s = -20% traffic
Adding 0.5 seconds to Google search results reduced traffic by 20% (Marissa Mayer).
0.1s = +8.4% conversions
Deloitte/Google joint study found 0.1s mobile speed improvement lifted retail conversions 8.4%.
Latency Cost by Industry
Milliseconds mean money in every industry. The specific cost model differs, but the principle is universal.
Ecommerce
7% fewer conversions per second (our model)
Our calculators assume about 7% fewer conversions for each second beyond 1s. Akamai measured a 7% conversion drop from just 100ms of delay in 2017.
Learn moreAPI / Microservices
P99 latency compounds across chains
A 6-service chain with 20ms P50 each has 120ms total P50, but P99 with retries pushes past 300ms. Tail latency is where the real cost hides.
Learn moreFinancial Trading
$4M per millisecond for brokers
A 1ms latency advantage is worth about $100M a year to a major trading firm. In Moallemi and Saglam's model, a trader 500ms slow lost over 20% of the spread on Goldman Sachs stock (4 Jan 2010).
Learn moreCloud Gaming
100ms playability threshold
Below 100ms, most players cannot distinguish cloud from local. Above 100ms, satisfaction drops sharply and subscriber churn increases.
Learn moreThe Compounding Problem
Latency does not just lose revenue once. It compounds across four dimensions simultaneously. A slow site loses direct conversion revenue, drops in search rankings (reducing organic traffic), discourages return visits, and damages brand perception. One source of slowness, four channels of revenue loss.
Akamai (2017): a 100ms delay can cut conversion rates by 7%. Immediate, measurable revenue impact on every visit.
Failing Core Web Vitals pushes you down in search results. Position 3 gets 4x the clicks of position 8.
Akamai (2017): 53% of mobile visitors leave a page that takes longer than three seconds to load.
Users associate slow performance with low quality. The halo effect works in reverse.
Load Time vs. Conversions: Our Model
What the calculator assumes: about 7% fewer conversions for every second above 1 second, compounding. Conversions are indexed to 100 at a 1-second load. This is a model, not a measurement; the published measurements are listed under the table.
| Load Time | Conversion Index | Change vs. 1s |
|---|---|---|
| 1s | 100.0 | Baseline |
| 2s | 93.0 | -7.0% |
| 3s | 86.5 | -13.5% |
| 4s | 80.4 | -19.6% |
| 5s | 74.8 | -25.2% |
| 6s | 69.6 | -30.4% |
| 8s | 60.2 | -39.8% |
| 10s | 52.0 | -48.0% |
The published measurements
- 100ms delay: conversion rates down 7%. Akamai/SOASTA, State of Online Retail Performance, April 2017 (about 10 billion visits).
- 2-second delay: bounce rates up 103%. Same Akamai report.
- Over 3 seconds on mobile: 53% of visitors leave. Same Akamai report.
- 1s to 3s load: probability of bounce up 32%. Google/SOASTA research, 2017, published on Think with Google.
- Every 100ms of delay: about 1% of sales. Amazon A/B tests, reported by Greg Linden, 2006.
The model index is 100 × 0.93 raised to the number of seconds above 1. Our 7%-per-second rate is far milder than Akamai's 7% per 100ms, so the calculators tend to understate the cost.
Frequently Asked Questions
How much does website latency cost in lost revenue?
Akamai's April 2017 retail performance report found that a 100ms delay in load time can cut conversion rates by 7%. The calculator on this site uses a far more conservative 7% conversion loss per extra second. On that model, an ecommerce site earning $1M per month loses roughly $70,000 monthly for each extra second of load time. Amazon estimated that 100ms of added latency would cost 1% of sales. At their current scale, that translates to billions per year.
What is the source of the Amazon '100ms = 1% of sales' statistic?
The figure comes from Greg Linden, a former Amazon engineer, who described it on his blog 'Geeking with Greg' and in his 2006 Stanford presentation 'Make Data Useful' (the 'Speed Matters' slide). In Amazon's A/B tests, deliberately delaying pages in 100ms increments produced measurable, costly drops in revenue, with every 100ms of added latency costing roughly 1% of sales. The related '7% conversion loss' figure traces to the Akamai/SOASTA State of Online Retail Performance Report (Spring 2017), which found a 100ms delay could cut conversion rates by 7%.
What did the Akamai study find about website speed and conversion rates?
Akamai's State of Online Retail Performance Report (Spring 2017, published April 2017) analyzed roughly 10 billion anonymous user visits from top online retailers, using data gathered by SOASTA (acquired by Akamai in 2017). It found that a 100-millisecond delay in load time can reduce conversion rates by 7% and a two-second delay increases bounce rates by 103%. The report also found that 53% of mobile visitors abandon a page that takes longer than three seconds to load. Note the Akamai 7% figure is per 100ms of delay, distinct from the '7% per second' assumption used by the calculator above, which is our own, deliberately milder model.
How much does 1 second of latency cost?
It depends on your revenue. At $100K monthly revenue, 1 second of unnecessary latency costs about $7,000 per month. At $1M monthly revenue, it costs $70,000. At $10M, it costs $700,000. Use the calculator above to model your specific situation. The 7% per second figure is our conservative model, set well below the Akamai/SOASTA 2017 finding of 7% for a 100ms delay.
Does latency affect SEO rankings?
Yes. Google has used page speed as a ranking signal since 2010 for desktop and since 2018 for mobile. Core Web Vitals (LCP, INP, CLS) are explicit ranking factors. Google's current 'good' thresholds are LCP under 2.5s, INP under 200ms, and CLS under 0.1, measured at the 75th percentile of real Chrome users. Slow pages lose revenue twice: once from lower conversion rates and again from reduced organic search traffic.
What causes website latency?
The main causes include slow server response time (TTFB), large unoptimized images, render-blocking JavaScript and CSS, no CDN or misconfigured CDN, excessive third-party scripts, large DOM size, and serving content from geographically distant servers. Most sites have multiple contributing factors that compound. A CDN alone often delivers 15-30% improvement.
What is P99 latency and why does it matter?
P99 latency is the response time at the 99th percentile, meaning 99% of requests are faster than this value. It matters because averages hide tail latency. A service with 20ms average and 900ms P99 is not fast for 1% of users. At 1 million requests per day, that is 10,000 slow responses. In microservice architectures, P99 latency compounds across the service chain.
How is the latency cost calculated?
The calculator assumes conversions fall about 7% for each second of load time above 1 second, our own conservative assumption, set well below Akamai's 2017 finding of 7% for a 100ms delay. It calculates the expected conversion lift between your current and target load times, then applies that lift to your monthly revenue. This is an indicative model, not a published benchmark. Actual impact varies by site type and audience.
Managing Cloud Infrastructure Costs?
Reducing latency often means adding CDN edge nodes, which increases egress cost. Platform engineering teams own latency budgets and SLOs. See our related guides.
You're using a tool we built.
Digital Signet builds custom software and AI automation for platform teams. Performance dashboards, alerting tools, cross-vendor integrations, workflow automation.
See what we build →