Evaluating hosting speed is not a trivial task. And as you probably know, to answer the question “what it is to be fast for a hosting” I monitor 15 US-based hosts from US West and East Coast and publish monthly reports. It’s truly a great source of the objective information on hosting performance.
However, in some cases these reports may be considered as limited because there are only two testing locations used – East and West Coast of the US. It’s enough for general evaluating hosting speed for US visitors though. But I wanted to see how fast the hosts are for visitors from other parts of the US and the rest of the world.
And to cover this knowledge gap to some extent, I did this research. I tested the speed of two US-based hosts (the affordable and the more expensive one) from 17 different locations in the world. And in this article I will present you the results of more than 1800 hosting speed tests.
By the way, if you are short for time, you can go directly to the conclusions.
In other words, I get compensation if you click on the links and make a purchase. It does not increase price for you anyway. All such links open in new window/tab; no software/program will be installed to your computer. (This is a standard notice required by affiliate programs terms.) Please note that I mention the products not as an advertisement, but as my recommendation.
Methodology of testing hosting speed from different locations
Tested hosts
From the 15 hosts I monitor on a constant basis, I’ve picked out two hosts that I recommend. One is more expensive (A2Hosting) and the other one is a smaller company with much more affordable plans (HawkHost). Both hosts have advantages which make them preferred choice in different cases (more on this read in my reviews on A2Hosting and HawkHost).
Doing my research I wanted to see not only how the speed of hosting differs depending on a testing location, but also how the speed differs between a cheaper and a more expensive hosting.
Please note that I used the cheapest plans on HawkHost and A2Hosting to host my test websites.
For testing I used the following US server locations. My test website at HawkHost is hosted in Los Angeles (more on this location read here). And the other website is hosted at A2Hosting’s Michigan data center (some details about this server location are here).
Tested websites
The websites hosted on HawkHost and A2Hosting are identical simple WordPress websites with a dozen articles (about 1000 words each) with summaries including images on the main page.
No caching plugins are used.
Both websites have a super lightweight yet powerful security plugin (BulletProof Security). (By the way, you can read more about this security plugin as well as about other website security solutions here).
Testing locations
There are 17 testing locations worldwide including 6 locations in the US, 2 in South America, 2 in Europe, 3 in Africa and Middle East, 3 in Asia and 1 in Australia.
Here’s the map where you can see where the testing locations are located. The A2Hosting and HawkHost data centers where my websites are hosted are also displayed on the map:
Locations to test the speed of HawkHost and A2Hosting
How the speed tests were performed
Here’s the overview of the process how I measured the speed of the tested hosts:
- I bought the smallest hosting packages from A2Hosting (see the description of the Lite plan) and HawkHost (see the description of the Primary plan).
- I used webpagetest.org service to measure the speed of my websites hosted with A2Hosting and HawkHost. I used Full Page Load Time which is considered as speed.
- From each of the 17 selected testing locations I ran about 12 tests with 9 speed measurements each. Thus, from each testing location I performed about 100 test runs. A large number of tests allow getting more objective results on how really fast the hosts are. By the way, you can download the saved webpagetest.org test runs here (236 MB size).
- I put together the test results into the spreadsheet and for each location I sorted the hosting speed test results to build the charts.
- I also calculated the simple averages of speed data within each location.
- Besides, to present the results from a bit different angle, for each location I grouped the speed tests into clusters (speed intervals). And I built the histograms based on this data. These charts are also very indicative on the hosting performance.
Results of Testing HawkHost and A2Hosting speed from 17 locations worldwide
Average speed test results
Here’s the chart:
A2Hosting (a more expensive host) and HawkHost (a cheaper host) speed compared
The chart displays average values of Full Page Load Time (i.e. speed) of A2Hosting and HawkHost measured from different locations. It’s clearly seen that A2Hosting (a more expensive host) outperforms HawkHost (a less expensive host) in all cases except Sao Paulo (Brazil) location.
Also, on the chart you can see added notes how many times A2Hosting faster than HawkHost on average for each location.
The locations on the chart are sorted by the difference between A2Hosting and HawkHost speed. The locations at the bottom show not much difference between A2Hosting and HawkHost speed. Whereas the locations at the top show much better speed when checking the websites hosted at A2Hosting than HawkHost. Reminder: my both websites hosted at A2Hosting and HawkHost are located in the US.
Hosting speed comparison table
The table above contains the data used for building the chat above. In addition, the table represents the difference between hosting speed of the more expensive host (A2Hosting) and the cheaper host (HawkHost).
On average across all locations, according to these tests A2Hosting is 1.45 times faster (or by 1.4 seconds) than HawkHost.
In the next chapter you will read the speed test results by each location individually.
Speed test results of HawkHost and A2Hosting from each location
In the chapter below I will present two charts for each location:
-
Line chart named “Comparison of hosting performance (Full Page Load Time)”.
Each data point on each of the lines represents one speed test of a specified hosting (Full Page Load Time of my test website hosted on that hosting). I sorted the test results by speed, that’s why the charts look like the speed decreases. But this is just the sorted representation of the data sets for smoother perception. And the number of test (X-axis) is not the timeline, but just an ordered number. In other words, the data set is sorted not chronologically, but by speed values.
The smaller values on the chart, the better (i.e. the faster hosting). The steeper and longer the line tails at the right of the charts, the more occurrences that a hosting speed was slow during my tests.
An ideal hosting speed would look like a horizontal line with very low values (e.g. fast and stable).
-
Bar chart (actually, histogram) named “How it’s likely that a website speed is that much”.
This chart is a representation of the previous chart’s data which I calculated and summarized. The histograms can give you more insights on the performance of the hosts that I’ve tested.
Whereas the previous chart represents the whole set of hosting speed measurements, a histogram represents the groups of these speed measurements. Each bar on a histogram basically shows the number of speed measurements which got into a specific speed interval divided by the total number of measurements for a selected location for a specified host.
In other words, each bar on a histogram is a probability of a hosting speed getting into an interval.
Also, these histograms show in comparison which host is faster analyzed in each speed interval (e.g. upto 3 seconds, 3-4 seconds etc ).
An ideal hosting speed would look like one bar on the left (i.e. in 100% cases the speed is fast and get into the first speed interval).
Oregon (US)
More than 80% of all speed tests of the more expensive host (A2Hosting) is less than 3 seconds. And about 80% of all speed test of the other host (the cheaper one, HawkHost) is between 3 and 4 seconds. These are the typical speed values for these hosts tested from Oregon.
San Francisco (California, US)
In about third tests the more expensive hosting (A2Hosting) was faster than 3 seconds. The cheaper host (HawkHost) did not perform that well. And its fastest results were 3-5 seconds.
Denver (Colorado, US)
Nearly 80% of speed tests of the more expensive host (A2Hosting) were less than 2 seconds. Whereas the cheaper host’s (HawkHost) speed was in the interval 2-4 seconds.
Lincoln (Nebraska, US)
All speed tests of the more expensive host (A2Hosting) were less than 3 seconds. Whereas the cheaper host’s (HawkHost) speed was 3-4 seconds mainly.
New York (US)
About 90% of all speed tests of the more expensive host (A2Hosting) were less than 2 seconds. Whereas the cheaper host’s (HawkHost) speed was 3-4 seconds mainly.
Orlando (Florida, US)
Both hosts performed more or less the same from this location.
London (UK)
Almost all speed tests of the more expensive host (A2Hosting) were less than 3 seconds. Whereas the cheaper host’s (HawkHost) speed was 3-4 seconds mainly.
Prague (Czech Republic)
Almost 90% of all speed tests of the more expensive host (A2Hosting) were less than 3 seconds. Whereas the cheaper host’s (HawkHost) speed was in the interval between 3 and 5 seconds mainly.
Istanbul (Turkey)
About 80% of all speed tests of the more expensive host (A2Hosting) were less than 3 seconds. Whereas the cheaper host’s (HawkHost) speed was mostly 4-5 seconds.
Dubai (United Arab Emirates)
The most speed tests of the more expensive host (A2Hosting) were less than 4 seconds and a big part of the tests were 4-5 seconds. As regards the cheaper host’s (HawkHost), its speed was considerably slower from this location. The speed was distributed across several intervals. More than half tests were less than 15 seconds.
The reason why HawkHost was generally much slower than A2Hosting can’t be explained by distance (net latency) or general hosting slowness. However, it looks like during the tests HawkHost’s shared server experienced excessive server load for some reason.
Mumbai (India)
Almost all speed tests of the more expensive host (A2Hosting) were in the interval 4-6 seconds. Whereas the cheaper host’s (HawkHost) speed was distributed in the slower intervals. And about half speed tests of HawkHost were in the 5-6 second interval.
Singapore
The distribution of the speed of the both hosts look similar. The only apparent difference is that the more expensive host (A2Hosting) had significantly more results in the fastest interval (3-4 seconds).
Tokyo (Japan)
About 90% of all speed tests of the more expensive host (A2Hosting) were in the interval 2-4 seconds. Whereas the cheaper host’s (HawkHost) speed was mostly 3-4 seconds (about one third of the tests) and 4-5 seconds (more than half tests).
Buenos Aires (Argentina)
More than half tests of the more expensive host (A2Hosting) were in the interval 2-3 seconds, and about 95% of the tests were in the interval 2-4 seconds. Whereas the cheaper host’s (HawkHost) speed was mostly 4-5 seconds with a significant chunk (about 20% of the tests) in 3-4 second interval.
Sao Paulo (Brazil)
This is the only location where the cheaper host (HawkHost) outperformed the more expensive one (A2Hosting). Most speed tests of HawkHost were in the interval 4-6 seconds. Whereas A2Hosting’s speed was distributed more evenly in all intervals (from 2 to 10 seconds).
Rose Hill (Mauritius, Africa)
The speed tests of themore expensive host (A2Hosting) were in two intervals: 3-5 seconds and even more occurrences in 5-7 second interval. Whereas the cheaper host’s speed (HawkHost) was not faster than 5 seconds.
Sydney (Australia)
The distribution of the performance of the both hosts was quite similar. The only noticeable difference is that the more expensive host (A2Hosting) was more often faster than the cheaper host (HawkHost). More than half tests of A2Hosting speed was in the interval 3-4 seconds, whereas about half tests of Hawkhost’s speed was 4-5 seconds.
Conclusions
- The difference between speed of the cheaper host and the more expensive one from my recommended list of hosts is quite noticeable (1.4 seconds on average after being tested 1800 times in total from 17 locations worldwide).
- The more expensive host (A2Hosting, Michigan data center) shows really good performance from all locations worldwide. It’s one of my recommended hosts (you can read my short review on it here).
- The speed of the cheaper host (HawkHost, Los Angeles data center) is comparatively good from almost all locations around the world. It’s also one of the hosts that I recommend (you can read my short review on it here).
- It’s recommended to choose a hosting first. And only then choose the data center (server) location. The reason is that you will get more speed if you choose a better host than a closer location.
- It’s fine to choose data center in the US even if a part of your audience is in Europe. The reason is that your visitors from Europe will not notice much speed decrease. But choosing a great hosting does matter (i.e again, it makes sense to choose hosting first than looking for the nearest server location).
- Testing location matters. In other words, if you test a hosting from New York and from Los Angeles, it’s likely you will get quite different results. So, make sure you test from different locations to judge the hosting performance more objectively.
- The general conclusion is that it’s recommended to choose a more reliable host with the data center close to your visitors. But if your favorite host does not have a server in the desired location, then it depends. But in almost all cases a better host is more important than the datacenter location of some other hosting.
Some conclusions may look obvious in general, but if you look more carefully at the data behind them you will see that the conclusions are not that simple.
Anyway, if you have any questions or if you want my opinion on your specific case, feel free to ask me in the comments or by emailing me privately.
Generally, when choosing a hosting and data center location I suggest checking out the hosts that I recommend. Most of them have data centers in multiple parts of the world to satisfy practically any geographical requirements to hosting location.
BTW, I respect your privacy, and of course I don't send spam, affiliate offers or trade your emails. What I send is information that I consider useful.
For a website, Hosting is a crucial part. You have made a right post!
Thanks, Preet.
I’m glad that my articles help people make a right choice when choosing a host.