Jobs, jobs and jobs. This non-performing economy should be all about jobs. So how are job web sites doing? How is their performance?

By performance I meant “web performance”. Yottaa’s “benchmark” feature allows anyone to create a benchmark to track and compare the performance of various sites over time.

Coming back to the topic of job sites, a Yottaa user named “testfoobar”(who are you, testfoobar? how about a slightly less geeky name such as “foobar”?) has created a web performance benchmark on yottaa.com to benchmark the web performance of top job sites. These sites are Indeed.com, SimplyHired, HotJobs(a Yahoo property), Monster.com, theLadders.com, theLadders.co.uk(which leads me to suspect “testfoobar” is from UK), and CareerBuilder.com. The benchmark was created on October 25th and is available at Employment Web Sites Performance Benchmark. The benchmark answers a few interesting questions from a web performance perspective: Who has the best performance and who are the bad performing ones? How bad are the bad performance sites? What are the problems with the bad performance sites? and what we can learn from this?

Who Has the Best Performance and Who Has the Worst Performance?

Indeed.com is the clear winner while CareerBuilder.com, monster.com and theLadder.co.uk are the bad boys.In the last two months, indeed.com clearly outperforms all other web sites significantly. The following shows the Yottaa Score of all these sites tracked over two months (Yottaa Score is a user experience score for measuring the web performance of a site, taking into consideration of  the various factors that impact user experience, such as Time to Title, Time to First Paint, Time to Display, etc).

The following snapshot shows the point of time differences in page load time, reachability, YSlow score and page size. Notice that indeed.com has the smallest page size (34.9KB), best YSlow score (92, which means it is well optimized) and best page load time (1.1 second). The only area that it lags behind is global reachability (420ms). Global reachability is a measurement of a web site’s infrastructure. It is not surprising that yahoo.com (who hotjobs belongs to) has the best infrastructure.

[

How Bad Are the Bad Boys?

Looking at Time to Display, Indeed.com takes about 1 second to display a page (“time to display”). All other sites take at least 2 seconds. The bad boys (monster.com and careerbuilder.com) take 3-6 seconds. The difference here between different sites is a whopping several hundred percent. For “Time to Interact”, the numbers are even worse.

The real trouble for the bad boys is “Time to First Paint”. “Time to First Paint” is the amount of time from when the user clicks the URL to the moment the user sees something show up in the browser window. During the entire “Time to first paint”, the user is staring at a blank browser screen and wondering whether this thing is working or not. For monster.com, careerbuilder.com and theLadders.co.uk, a user is staring at a blank screen for 2-3 seconds. I bet that they are losing a significant percentage of visitors right there.

What’s Wrong With the Bad Boys?

Further investigation of the benchmark would reveal the reasons why the bad boys are bad boys:

1. Infrastructure Issues: Infrastructure issues include both bad data center connectivity(or location) and bad DNS service.

The following diagram shows that it took 500ms to 600ms in average to connect to CareerBuilder.com’s Web Servers while all other sites took 200ms to 300ms. CareerBuilder.com might want to re-evaluate their data center location, connectivity or both:

The following diagram shows the DNS resolution time for these job sites. While all other sites took less than 500ms, Monster.com and theLadders.co.uk took 1000ms to 1500ms:

2. Page Complexity Issues: Issues related to page complexity significantly impact web performance. These issues include page size, asset count, number of domains, number of third party assets, number of requests, number of Javascript files, etc. Yottaa web performance tool collects this data and visualizes them.

The following diagram shows CareerBuilder.com and Monster.com have the biggest page footprint:

Page Size – Job Sites Web Performance Benchmark

The following diagram shows CareerBuilder.com and theLadders.co.uk have almost twice as many assets as other sites, which degrades performance:

3. Lack of Optimization: There are a wide range of techniques that can be applied to make a web page load faster. The best known tool is called “ySlow” from Yahoo. Google also produces a tool called “PageSpeed” that is similar to ySlow. “ySlow” produces a score called “ySlow Score” which ranges from 0 to 100. The higher ySlow Score is, the better a web page is optimized. The lower the score is, the more room there is for optimization. Yottaa web performance tool currently embeds ySlow in its service in evaluating the degree of optimization for various pages and provides recommendations.

Looking at the ySlow score tracked over time in the diagram below, CareerBuilder.com and theLadders.co.uk both have ySlow scores below 50 which means both sites have lots of room for optimization that can improve the performance:

What Can We Learn From Here?

The take away from this web performance benchmark are some useful guidelines for how you should approach your web site performance.

First, you should measure and benchmark your web performance. Yottaa Web Performance tools are designed as free online services to do so. It enables you to assess and benchmark the performance of your site, openly and freely. When it comes to optimization, you should look at three different aspects:

1. Infrastructure

Is your data center well connected? Is your data center located in the right location? Do you need more data centers? Do you have the right DNS service? Do you need CDN service? etc. Yottaa’s web performance tool has a metric called “Global Reachability”. This metric is designed to measure infrastructure aspect of a web site and allows you to compare different sites and make informed decisions.

2. Content complexity

Can you reduce your page size? Do you really need so many third party widgets? Do you really need so many assets? etc. Yottaa’s web performance tool uses a metric called “Page Complexity” to measure this aspect, giving you a number that you can use to compare and make decisions.

3. Optimization

Did you optimize your content? Did you follow the best practice? Yottaa embeds tools like “ySlow” to measure this aspect. Again, this number allows you to compare and make informed decisions.