Home > seoexpert's blog > Web 2 Point Slow - Slowcial Communities

Web 2 Point Slow - Slowcial Communities

Web 2. SlowThe Web is one of the fastest changing media, but one aspect seems to remain unchanged: the Web is slow. Back when I started using the Web in the mid 90ies I either accessed the Internet from a computer in the university or from home with a 14K modem. It was a nuisance to download a page with more than 50KB.

Today, I use a DSL2000 connection to access the Internet, which is theoretically 140 times faster, but it still takes long to download many pages.

Of course, new technology provides for much faster access, but today many pages require more than 500KB of data to be downloaded. There are more images, videos, flash, JavaScript etc., which would have been impossible more than 10 years ago.

Recently, I installed the new YSlow extension for Firefox which is actually a plug-in for the great Firebug extension. On the YSlow website it reads:

YSlow analyzes any web page and generates a grade for each rule and an overall grade. If a page can be improved, YSlow lists the specific changes to be made.

Yesterday, I measured 20 Web 2.0 community websites, that are ranked among the Alexa global top 100 sites, with YSlow. To be included in the test the following conditions had to apply:

  • The site is in English.
  • The front page displays content without the need to be logged in.
  • The site offers community functions such as a forum, sharing content, messaging or others for registered users.

Test Conditions

I used Firefox with the latest versions of Firebug and YSlow. The Firefox pop-up blocker was enabled and I was not logged in to any of the sites. I measured the speed of the start pages, the only exception was Youporn.com, where I needed to hit the enter button before content was displayed.

YSlow Results

The following table displays the websites tested, their Alexa Rank from 10.08.2007, and the YSlow grade from the same date:

Alexa RankWebsiteYSlowgrade

4 sites stand out: yahoo.com and craiglist.org with a grade higher than 90 as well as dailymotion.com and icq.com with a grade less than 50. The average grade is 65,75 (D) and the median is 65.

I also created the following chart to better visualize the measured data:

Web 2 Point Slow - Slowcial Communities


These descriptive statistical data show that successful Web 2.0 communities are pretty slow. This is not necessarily a problem of slow Web servers or Internet connections, but of the amount of data, the number of HTTP requests, too much JavaScript, Flash, images and other media, HTML structure and the time it takes for the browser to render the pages.

I also checked some popular blogs just for curiosity to see that the situation seems even worse in this sector. Maybe I will do a top 20 blogs test in the future. By the way, SEO Expert Blog has a grade of 75C and the fastest site I have measured so far is to no surprise Google.com with a grade of 99A.

Interesting site you have there - youporn.com. I try to keep an eye on Web2.0 sites, but I'm not familiar with this one. Wonder if it's work safe ;)

youporn.com seems to be very popular but I guess most people visit it at home when they are alone ;)

Interesting to take this data and compare it against other tools that measure site performance and then rate those against a study group of real people and their "apparent" speed scores.
There's a logical error in your argument. You say "These descriptive statistical data show that successful Web 2.0 communities are pretty slow". That is not a valid conclusion from your experiment. YSlow is not a measure of performance. It is a measure of "the level of conformance to a set of rules that might make your site faster". Of course there is a correlation between optimised sites and performant sites but server performance/load and content size are equally important. Without actually testing it, I guess a single 50Gb image served from a CDN will achieve an A score (n/a in most areas) but is not a good web page or very fast to render. The best you can achieve with this type of analysis is to graph average download speed against YSlow score to show which sites can achieve performance improvements. I would also add that Yahoo sites are bound to score a really good YSlow score because the YSlow test is, when all is said and done, "Do you use Yahoo's optimisation techniques?"

This is not a scientific analysis and I nowhere claimed that it is one. I'm aware that you can't prove anything with descriptive data only, and I think that you can prove almost anything you want with inferential methods.

The benefits for Yahoo by providing a tool like YSlow should be taken into account, nonetheless I think it's a useful tool to get an idea of what can be optimized.

The server side is one thing the client side another. A server cluster with lots of powerful machines can serve pages fast, but it makes no difference to the user if he accesses a site with a slow computer and has to wait for a minute until the browser finished interpreting and rendering the code.

sorry, I didn't mean to criticise too harshly, just clarify and help understanding. For the record, Sonopia use YSlow to check our optimisations. Yahoo are really leading and helping the industry in general here.

Never mind Roger. With my reply I wanted to make some things clear that may be misunderstood. Criticism is justified and I am glad that there is some discussion on the topic.

YSlow is good for development. From one hand we have tons of popular ajax, from other fast net. Thing is to keep balance.
YSlow is cool, but IMHO there is nothing better than WebDeveloper extension...
this tool is great. i used fasterfox before to measure the content loading.. but this tool is really better and has further features. thanks a lot!