Test Website Speed From Multiple Locations

Posted on Aug 21, 2021

Share

PageSpeedPlus allows you to Test Website Speed From Multiple Locations. 11 different places are available including London, California, Virginia, Montreal, Frankfurt and more. Just add a URL and we will automatically scan it from there. Most importantly the data is presented nicely in our UI so it can be analysed with ease. This feature is a native part of our tool and compliments the PageSpeed Insights scans that we perform.

Try PageSpeedPlus Now

Contents

Why we built it

PageSpeedPlus has traditionally been focused on scores from PageSpeed Insights, Lighthouse and CRuX dataset which report on Page Experience metrics like Web Vitals (FCP, FID, CLS). We’re very good at measuring these and the quality of our product is demonstrated in the customer testimonials and feedback we get. Sometimes though, users want to see raw loading time in seconds rather than the score out of 100 from Google’s tools. Based on feedback from our audience, they also want to know what that loading time is from various locations.

We know there are existing tools on the market that offer load time testing from multiple locations so we analyzed them but found they have limited automation or require tests to be done manually, offer a small number of locations and the data is difficult to visualize.

Dissatisfied with 3rd party offerings, we decided to apply our Web Performance expertise in building a PageSpeedPlus feature to Test Website Speed Globally.

How it works

PageSpeedPlus runs synthetic tests. We have infrastructure in all of our test locations, which allows us to execute code from that physical place and get a true measurement of loading time. When a test fires, a job is dispatched to every region, which initializes a fresh Chrome instance using Puppeteer and visits the URL in question. Metrics are extracted from the Navigation Timing API and stored against that particular scan.

navigation-timing.jpg The Navigation Timing API

Info
As our tests are synthetic they are executed in consistent conditions every time, which makes monitoring and debugging easier.

What we measure

Time To First Byte
This is the amount of time between creating a connection to the server and the first piece of data being returned to the browser so is a measurement of the server responsiveness. TTFB is very important because it has a knock-on effect to later events that occur when a page is being rendered and a good score means subsequent stages in the page rendering lifecycle can execute sooner. Google’s advice is to aim for 600ms or less to keep users engaged.

Start Render
Start Render is the amount of time between creating a connection to the server and the browser displaying the first asset on the screen that a user can see. Lots of things happen when a page is being downloaded but this is the first noticeable event to the naked eye so it greatly influences our perception of speed. It tells us the website works so displaying something quickly holds user attention and decreases the likelihood of a bounce. Aim for 1000-2000ms.

Total Loading time
Total loading time as the name suggests is the amount of time between creating a connection to the server and the web page being fully downloaded and ready for the user to interact with. This is the single metric that encapsulates loading speed above all others and a good score means you have a fast site with a quality user experience. Aim for 2500ms or less to keep users engaged. 

metrics.jpg

Downloading a webpage is a multi-step process. Measuring the three key moments in the page rendering lifecycle allows bottlenecks to be indented so you exactly know where to investigate further.

Comparison with other other tools

Other Tools
  • Manual task to run tests or Automation not frequent
  • Limited locations
  • Data difficult to visualize
  • Expensive
  • Long queues on occasions
PageSpeedPlus
  • Automated tests every hour
  • Multiple URLs at once
  • 11 locations
  • Easy To Use. Set and forget
  • Fast
  • Intuitive charts
  • Analyse data over time
  • Part of existing PageSpeedPlus pricing plans

As you can PageSpeedPlus is superior in many ways.

Popular Websites Loading Time Around the World

Here are some examples of some of the largest websites in the world and their loading time from 11 global locations captured by PageSpeedPlus:

chart.jpg

As you can see, loading times are not consistent around the world. This could be due to factors outside the control of the website like telecommunications infrastructure in the target location but a large amount is within their control like server configuration, caching, API calls and on-page javascript. This chart helps to track speeds from multiple regions and identify subtle differences that can have a big impact on user experience.

Locations and plans

Global Load Times testing is available on all paid plans without additional charges.

Basic
2 locations
Standard
5 locations
Pro
11 locations

The locations available are:

  • Dublin
  • Tokyo
  • California
  • Montreal
  • Sydney
  • Frankfurt
  • London
  • Virginia
  • Ohio
  • Oregon
  • São Paulo

This is the current list of locations. More will be added in the future.

Why Measure Website Speed Worldwide?

In short, loading time differs for users depending on where they are located and a fast experience is required for everyone who visits your site to avoid abandonment and high bounce rates.

When someone visits a web page, the request typically has to make a round trip to the server and back to get the content no matter where that user is located. The greater the distance they are from your server, the further the request has to travel geographically and therefore the longer the loading time will ultimately be.

If your server is in London, users in Manchester won’t have to wait very long to get a response because the distance is only 180 miles. On the other hand, users from Melbourne will have to wait for their request to travel to the other side of the earth and back, which will take much more time. This latency is unavoidable as humans have no control over the speed of light.

This is an extreme example but is not an uncommon scenario. Check your Google Analytics and there will likely be traffic from places you didn’t expect.  

Even if you only have traffic from a single country, loading time can still vary greatly. For a website hosted in Virginia on the US East coast, a user in California still has to wait for the request to travel thousands of miles across the county. SEO conscious readers will know that Google crawls from California so you should consider them a user and pay attention to loading times from there.

Quick Tip
To reduce the latency of server response times consider having multiple distributed servers, loading assets from a CDN, caching and compressing pages

Conclusion

People visit sites from every corner of earth so loading times will differ a lot and you need to monitor them globally. Having a robust tool to regularly test from multiple locations will give a true understanding of user experience from around the world and identify if there are infrastructure changes required to serve portions of your audience better.

Testing the speed of a webpage from multiple locations is made easy with PageSpeedPlus. Information is presented in a clear way that is easily understood by all stakeholders to ensure everyone is on the same page when talking about performance. Just add your URL today and we will automatically test it on a schedule with elegant charts to help visualise the global performance.

Try PageSpeedPlus Now