#SAM2013 Ribbon
Slow Websites cost the US Ecommerce Market $504 Billion in 2011
Posted in: Ecommerce   -   August 10, 2012

Most ecommerce companies are well-versed in the first lesson in web performance: nobody likes a slow website; faster is always better. Many ecommerce companies are probably even familiar with data revealing the impact a slow site can have on sales, bounce rates, average page views, and other critical web KPIs. However, even more significant to retailers than these facts and figures alone are the implications revealed when performance data is combined and synthesized.

At Rigor, we want to show e-retailers what all of these numbers reveal about the total potential web performance possesses to increase ecommerce sales for the individual retailer and the ecommerce market as a whole.

The following points provide the pieces needed to put together the big picture revealing what the ecommerce market would look like in a utopia of optimization and lightning fast sites:

  1. Google engineers found that users begin to get frustrated with a site after waiting just 400 milliseconds – literally the blink of an eye – for web pages to load. (Source)
  2. Amazon found that shaving 100 ms off of load time results in a 1% increase in sales (Source)
  3. The average load time for ecommerce sites is 10 s (Source)
  4. Total US retail sales for 2011 were estimated at $4,155 Billion; Ecommerce constitutes $194 Billion, or 4.67%, of those sales (Source)
These data points are widely publicized and probably common knowledge to many ecommerce experts. However, what industry leaders don’t talk about as much is what a synthesis of these facts can show us about the massive impact performance optimization could potentially have on the growth of ecommerce and the sales potential the industry is forgoing by allowing poor performance to permeate the market. Below we find that some simple math reveals some astonishing conclusions.

What the Average E-Commerce Site Could Achieve
Let’s start with the 10 s average load time statistic. If the average site takes 10 seconds to load, and users begin to abandon pages after 400 ms, this leaves room for 9.6 seconds of improvement to the average site.

 10 s – .4 s = 9.6 s
Using the data from Amazon revealing that each 100 ms decrease in load time results in a 1% increase in sales, we find that a 9.6 s improvement to a site’s load time would increase revenue by 260%.
(1 + 1%)9.6s/100ms = 260%
In short, this number crunching shows us that for the average individual e-commerce retailer with a load time of 10 s, achieving ideal site speed would result in an astronomical sales increase of 260%.

What Performance Optimization Could Mean for E-Commerce as an Industry
With a bit more basic math, we can determine the sales gains the ecommerce market as a whole could achieve if ecommerce sites were as fast as users would like them to be.

Using the Census data from 2011, we can see that total ecommerce sales were $194 Billion. If ecommerce sites could theoretically increase sales by 260% as previously calculated, that means, as a whole, the ecommerce industry loses $1/2 Trillion / year due to poor performance.

$194 B x 260% = $504 B
If 2011 were a utopia with no latency, the e-commerce market would have generated $698 Billion in revenue and would constitute 15% of total US retail sales.

$194 B + $504 B = $698 B

$698 B / ($4155 B + $504 B) = 15%

Now, I realize that we do not live in a site speed utopia and that this model is idealistic in assuming zero tradeoffs, but it is interesting to consider what the ecommerce industry would look like if the average load time for ecommerce sites were 400 ms–the load time that reportedly would be satisfactory to impatient internet users. Moreover, these numbers show us the great potential the web performance industry possesses to help ecommerce companies exceed revenue and growth objectives by satisfying the universal need for speed.


Craig (@craighyde) is responsible for product management, marketing, sales, and operations at Rigor. Craig was previously a Solution Consultant for Compuware/Gomez where he designed and implemented application performance solutions for the Fortune 100. Before that, he provided the Fortune 1000 and government agencies with systems to test and manage their networks as Regional Sales Manager for Network Orange. Craig is active in the Advanced Technology Development Center at Georgia Tech and can be found on the pitch with Atlanta Old White Rugby Football Club.

Tags: , , ,

  • http://blog.justindorfman.com jdorfman

    Great post Craig. Hope all is well at Rigor.

  • http://paulirish.com Paul Irish

    How much are slow browsers costing the market?
    https://twitter.com/paul_irish/status/226332033054539776

  • Der Meister

    The article misses the “to the power of” sign: (1 + 1%) ^ 9.6s/100ms = 260% (the calculation is correct).
    But with this calculation you get 260% of the _current_ sales, which is “just” 160% increase.
    Thus the e-commerce market would just generate 504 B (instead of 698 B), which would be 11% of total sales (idealistic, as you say yourself).