How Fast is Your Website, Really? Learning More from AlertSite

We all understand how important delivering fast and feel-good experiences are to the success of our applications and the business goals they support. Zona Research introduced us all to the 8-second rule in 2001, while current research about customer expectations suggest the new benchmark is a page load time of 2-3 seconds.  More importantly, application slowdowns have twice the impact on business results as outages.

One of my favorite comments about speed comes from Fred Wilson, a New York based tech investor and active blogger, while speaking at the annual Future of Web Apps in 2010 about the 10 Golden Principles of Successful Web Apps.  He said, “First and foremost, we believe that speed is more than a feature. Speed is the most important feature. If your application is slow, people won’t use it.”

But what is Website speed?

I talked about website speed in the following blog: Website Performance - Only ONE Aspect of Website User Experience.

The three sides of the prism for understanding speed as it relates to Website performance and user experience are:

  1. The traditional view of Web performance (a.k.a. “the waterfall report”) which measures how long it takes the network to deliver every resource referenced in the website HTML to the browser from the network.
  2. What the browser thinks about how fast the page is loading (a la DOM Load and Page Load metrics).
  3. What a user would see with their eyes as the page is rendering on-screen. This is the ultimate empirical measure of speed.

To that end, we are excited to announce the release of our perceived user experience metrics in our Web performance monitoring service, AlertSite.  We have essentially developed, in software of course :), a set of robot eyeballs that:

  • examine the Web page as it’s rendering in the viewable portion of the browser;
  • measure the point at which it first paints;
  • measure the point at which the page stabilizes above the fold; and
  • captures images of the screen at these milestones to provide a visual reference.

I’ve collected some summary data from the home pages of the top 5 online retailers per Internet Retailer.

Perc. MetricsBrowser MetricsTrad'l Metrics

First Paint

AFT

DOM

PAGE

Response Time

Amazon

 0.83

 2.34

 1.20

 2.03

 4.40

Apple

 1.05

 1.58

 0.86

 1.49

 1.22

Dell

 1.04

 1.75

 1.39

 1.75

 1.89

Staples

 1.11

 3.84

 2.29

 4.61

 4.71

Walmart

 1.39

 2.13

 2.90

 4.76

 4.76

In some cases, there are some very significant differences between the new perceived performance metrics and the traditional network oriented measures; in other cases, not. This all depends on exactly how your Web application is constructed.

Let’s look together at two different examples from the Top 5: Amazon and Walmart.

Amazon, from the technical perspective, took almost 4.5 seconds on average to load all of the page resources for Amazon’s home page. And this is a valid measure of how long it took the network to get all those assets to the browser, but how does this compare to what a user would perceive?

In less than one second, the majority of Amazon users see a page that looks like this (clipped from the captures saved in the reports):

Amazon

A fully rendered page without the ads on the right loaded into those iFrames.

Before 2.5 seconds, users see a fully rendered page including the promotional content on the right side:

Amazon1

For Amazon, I would say the First Paint time represents the best measure of perceived user experience as all relevant content is typically displayed by then just not promotional right-side content.

Walmart shows more than 4.5 seconds to load from the traditional waterfall perspective yet the first paint moment happens around 1.4. About half the time, that first paint moment shows just the header like this:

Walmart1

And the other half of the time that first paint moment shows the bulk of the page like this:

WalMart2

Walmart.com (the page) was visually stabilized above the fold in a little more than 2 seconds typically.

When briefing the analyst community, as we were preparing for launch, a really interesting bit of feedback was about how the whole purpose of measuring Web performance is to use that data as a proxy for end-user experience and that by providing all three perspectives we help customers choose the measure that best approximates user experience for their application.

So the real question is…which of the newer user experience metrics that we now provide are most closely correlated with satisfaction, bounce rates, productivity, conversion or any other business KPIs for your Web application?

Visit my follow up blog to see the new AlertSite perceived user experience metrics in action.

 

 

 

 

 

 

 

 

 

 

 

 

 


Close

Add a little SmartBear to your life

Stay on top of your Software game with the latest developer tips, best practices and news, delivered straight to your inbox

By submitting this form, you agree to our
Terms of Use and Privacy Policy

Thanks for Subscribing

Keep an eye on your inbox for more great content.

Continue Reading