The 2016 Web Hosting Speed Test Results Have Arrived.
Understanding The Results
Page load speeds were tested from 100+ websites on each hosting provider (see caveats at bottom). Tests were done from a computer in Washing, DC (orange) and San Fransico, CA (blue).
The x-axes shows the time in seconds it took for a web page to load, the y-axes shows the number of websites that loaded a page in a particular amount of time.
Disclosure: I receive commissions on a referral bases from most of these web hosting companies. I spend a lot of time and work very hard putting together and keeping this data up to date, relevant, unbiased and accurate as possible.
Summary Of Results
A total of 16 shared hosting providers were tested - the fastest shared hosting for 2016 goes to iPage with 1&1 hosting trailing close behind. iPage was also the fastest in 2014 and seems consistently to be at the top. A few notable shifts from last year was Bluehost’s big drop in speeds going from .74 seconds to 2.4 seconds on average. Also, last year’s winner Arvixe fell off completely, finishing close to the end of the pack in 2016. The hosting giant Go Daddy performs ok, but has been slowing down a little since 2014.
I broke out cloud hosting/VPS providers from shared hosting this year, because it’s not really fair to compare VPS hosting services to shared hosting services. In the VPS hosting world - Linode edged out Amazon AWS while Rackspace and Digital Ocean tied at the respectable speed of .69 seconds.
Speed results are below, also here are some other interesting things I noticed and my hosting recommendations.
These results have been updated in April, 2016. You may see 2015 results here, 2014 results here or 2013 results here .
Shared Hosting Providers Results
VPS Hosting Providers Results
Interesting Things I Noticed
iPage is much faster on east coast
Interestingly, iPage is blazing fast on the east coast but not quite as fast on the west coast. While 1&1 is the otherway around. My guess is most iPage servers are on east coast while most of 1&1 servers are on west coast. Just looking at the west coast iPage averages .61 while 1&1 average .55 meaning 1&1 is a little faster on the west coast. Looking at the east coast iPage really crushes at .45 while 1&1 is at .56.
Shared hosting performs well compared to VPS hosting
Another fun thing I noticed is how well many of the shared hosting providers perform relative to VPS hosting providers. I expected a product like AWS to really blow all the shared hosting providers out of the water, but both iPage and 1&1 appear to be faster on average than AWS. 5 shared hosting providers come in faster than rackspace, who is industry leading VPS/cloud hosting provider.
Arvixe dropped off big time
Arvixe got a lot slower from 2015, they also seemed to lose quite a bit of market share (I could only find 27 sites to test). They were purchased by EIG not too long ago, and the now appear to share their hosting servers with A small Orange (Another EIG Company).
What on earth is bluehost doing?
Bluehost did pretty bad in 2014, then they looked up in 2015, however in 2016 they completely fell off. Not including bluehost, the range of the other 16 hosting providers was from .54 - 1.2 seconds. Bluehost manages to be almost twice as slow as the seconds slowest company Host Gator.
My Hosting Recomendations
I use to write long reviews on all features of web hosting, but have since gotten pretty tired of doing that. The bottom line is for most small businesses shared hosting is fine (if you're getting <10k visitors/mo - which you probably are if you're taking my advice) and iPage is currently offering the best shared hosting product (for the vast majority of businesses). They are weirdly cheap, have all the necessary features and are super fast. They are EIG owned (web hosting conglomerate), but they appear to operate as a separate entity, on different servers, and have less spamy looking website administration.
If you're scared of EIG, Inmotion Hosting is a safe bet. They have a good product, very robust features, good support and fast servers. 1&1 Hosting is also a fine choose, they offer a pretty snazzy website builder which can be a good option for those non-programmers looking to use a website.
If your site has reached over 10k monthly visitors and you're ready to move to a VPS server, Then I would recommend either AWS as it has the most robust product, or Linode as you can get a slightly better deal if you don’t use some AWS specific features.
There are tons of factors that go into website loading speed, many of which are outside the control of a web host. However some factors are very much in control of a web host and those are the factors that we want to analyze and compare between hosting providers.
The primary two factors are the servers processing power and then the bandwidth (transfer rate) of the website content. In testing I attempted to focus on these two factors.
Since hosting providers will set up websites on various servers I needed to test the speed of a number of different websites, so that we could get a good feel for how fast their servers are on the whole. I tried to test at least 100 hundred sites for every hosting provider (a few don't have 100, see caveats below).
The other variable I wanted to account for is physical location of the testing computer, since a website hosted on the east coast will take longer to load from a computer on the west coast than a computer on the east coast I wanted to test from a couple a locations.
I also wanted to test websites that were regularly used, but not huge (like google) because most big websites have special technologies for website loading. To meet these requirements I tested only the speed of websites getting anywhere from 300-10,000 monthly visitors.
Other Notes & Caveats
1. There were a few results that were over 10 seconds, but I threw them out as their may have been a network hang.
2. To measure variance I ran the same test a few times and saw no substantial changes.
3. Arvixe, A Small Orange, A2Hosting and IX Webhosting all had less than 100 sites tested. This means there could be more variance in a re-test , (however I noticed basically no variance on a re-test < .02s across all providers).
4. Bluehost results include JustHost and HostMonster which appear to run on the same servers.
5. Arvixe results are most likely only testing from old sites. New users to Arvixe may now be setup on either Arvixe, A Small Orange servers (or potentially some other ones that I don’t know about).
6. Yahoo Web hosting is now Aabaco web hosting, who appears to be slowing down the servers substantially from last year.
This data may be reformatted and redistributed as long as a reference to this document is provided. Please do not copy verbatim.