Pingdom Website Speed Test

Pingdoms Website Speed Test is probably the best way to measure the speed of your website. Not only does it incorporate most of the things Google PageSpeed does, but it also adds additional features and the ability to test performance from different locations around the world.

To use it all you need to do is point you browser to Pingdoms Website Speed Test, enter a URL, and select where to test from. And that is basically all there is to it.

What's Good

• Ability to test from different locations.
• Tracks actual load time
• Tracks number of requests made
• Tracks download size
• Shows an in-depth breakdown of which resources are loaded and what amount of time was spent doing so.

What's Bad

• Performance results are sometimes very inconsistent between runs, indicating that system load on Pingdoms end has more of an effect than you'd probably like.
• Encounters "errors" with some frequency which prevents it from displaying the results.

Sample Results

Bellow you'll find sample results testing from Dallas Texas

Site	Grade	Loaded in
Dev.to	88 (B)	0.859 s
Apple	74 (C)	0.944 s
This Blog	100 (A)	1.01 s
Android	98 (A)	1.26 s
Microsoft	75 (C)	2.07 s
The Verge	73 (C)	5.34 s

And from Stockholm Sweden

Site	Grade	Loaded in
This Blog	100 (A)	0.123 s
Dev.to	88 (B)	0.305 s
Android	98 (A)	0.588 s
Apple	74 (C)	1.32 s
Microsoft	75 (C)	2.52 s
The Verge	73 (C)	3.00 s

As we can see there is no difference in the Grade, which is basically based on the same rules as Google PageSpeed but weighted differently.

We can also see some variance in load speed based on where in the world you are. This highlights the importance of server locality, and what effect the use of a CDN to spread static resources around the globe can have on performance. The results are most pronounced for this blog where accessing it from Stockholm in Sweden (which is close to where this is currently hosted) loads the entire page in 0.123 seconds, while doing so from Dallas Texas brings that up to 1.01 seconds. A massive difference.

Testing from all the different locations

Here we have a table with timings for this site from various locations:

From	Seconds	Distance
Stockholm, Sweden	0.123	20 km
New York, New York	0.827	6 300 km
Dallas, Texas	1.01	8 100 km
San Jose, California	1.29	8 640 km
Melbourne, Australia	1.83	15 500 km

Again, we see a clear correlation between the distance from the server to the client, as we'd expect. Do note that distances here are travel distances and may not correspond to actual network distance which is basically guaranteed to be greater.

And with all resources moved to Amazon CloudFront

CloudFront serves your content via different servers around the globe seriously cutting down on the time to load for all places with the exception of Stockholm:

From	Seconds
Stockholm, Sweden	0.121
New York, New York	0.331
Dallas, Texas	0.42
San Jose, California	0.428
Melbourne, Australia	0.882