Our newest quarterly performance state of the union, released earlier this week, analyzes the load time of the top 500 ecommerce sites and compares their current performance to their performance one year ago. Among other things, we learned that the median web page takes 9.3 seconds to load — a 21% slowdown in the past twelve months.
Every quarter, we measure the load times of the top 500 ecommerce websites (as ranked by Alexa.com) with our eye on a number of performance metrics, including load time, time to interact (aka TTI — the moment when a page’s primary content loads and becomes interactive), page size and composition, and adoption of performance best practices.
The purpose of this research — now available in our Ecommerce Web Performance State of the Union (Winter 2013-14) — is to obtain a current snapshot of how websites perform for real users in real-world scenarios, as well as an understanding of past performance patterns that allows us to predict future trends.
Highlights of our key findings include:
1. The median page has slowed down by 21% in just one year.
The median top 500 ecommerce home page takes 9.3 seconds to load. A year ago, the median page took 7.7 seconds to load. The majority of online shoppers will abandon a page after waiting 3 seconds for it to load.
2. The top 100 sites are slower than the rest of the pack.
We compared load times for the top 100 ecommerce sites to load times for the top 500 sites. Among the top 100 sites, the median page load is 10 seconds – making the median top 100 site 7.5% slower than the median top 500 site (at 9.3s).
3. At 5 seconds, the median time to interact (TTI) does not meet user expectations for sub-3-second load times.
Only 20% of the top 100 sites had a TTI (defined as the point at which a page displays its primary interactive content) of 3 seconds or less. 26% of the top 100 sites took 8 seconds or more to become interactive.
Why These Findings Matter
When it comes to performance, spectacular site outages grab headlines, but slow-loading pages don’t. Much of this has to do with public visibility into the problem. When Facebook or Amazon go down, it doesn’t take long for the entire Internet to know about it. When a site is slow, it may earn a few cranky Twitter comments, but that’s about the extent of the reporting.
The advantage to very public outages is that companies know about the problem — fast. While damage control isn’t fun, the arc of an outage is straightforward — site goes down, public outcry ensues, site goes back up, usually within hours — and has a clear beginning, middle, and end. Site slowdowns have no such tidy arc. There are a number of reasons for this, ranging from lack of real-time performance visibility at the site owner’s end to the fact that slowdowns can be erratic and difficult to pinpoint.
While outages carry a higher cost penalty per hour than slowness, website slowdowns occur 10 times more frequently than outages. According to one study, website slowdowns can ultimately have twice the negative impact on an organization’s revenues as outages.
The key takeaway from this report is that site owners must not only ensure that their pages are fast enough to meet user expectations, but they must also have ongoing real-time awareness of how their sites perform for users in the real world in order to ensure that slowness is not creating a slow revenue leak that is ultimately more damaging than even the most dramatic outage.
As a former senior researcher, writer, and solution evangelist for Radware, Tammy Everts spent years researching the technical, business, and human factor sides of web/application performance. Before joining Radware, Tammy shared her research findings through countless blog posts, presentations, case studies, whitepapers, articles, reports, and infographics for Strangeloop Networks.