Home > Designing, Others > How Web Hosting Can Impact Page Load Speed

How Web Hosting Can Impact Page Load Speed

November 29th, 2016 Leave a comment Go to comments
web hosting page load

Page load speed is a very important factor when it comes to the overall success of a business’s website or blog. And it is something that should not be taken lightly. It can affect almost everything; from your conversions, SEO and rankings, time on your site, bounce rate, and the entire user experience. If you dip in all of those areas, this could easily result in loss of revenue and customers. Today we will explore just how much page speed is impacted by certain web hosts and why you might want to reconsider using a cheap shared host.

Speed is a Ranking Factor

Google announced way back in 2010 that page load speed is a ranking factor. What that means is that sites with a faster page load speed could in turn rank higher in search engine results pages (SERPs). In Google’s Site Performance for Webmasters video, Developer Programs Tech Lead Maile Ohye, states that “two seconds is the threshold for e-commerce website acceptability. At Google, we aim for under a half-second.”

And fast forward to today and we now have Google launching the mobile-first index. This means that the ranking factor is also heavily looking at mobile optimization. You can almost guarantee that whatever weight they placed on the speed ranking factor before that it will be increasing. So you need to ensure that you have a fast responsive loading site, along with possibly an AMP implementation.

You can use free tools like Google PageSpeed Insights or Pingdom to test your website’s page load speed. They will inform you how fast your website is and offer suggestions on how to improve it, such as leveraging browser caching or minimizing redirects.

pingdom speed test

Page Load Speed Can Affect Conversions

A slow page load speed can also affect your conversions and sales. For an e-commerce site, the last thing you need is to be losing revenue simply because your website is too slow. According to research from Kissmetrics,

“A 1-second delay in page response can result in a 7% reduction in conversions.”

2-speed-affects-conversions

Example: Let’s say your website loads in 3 seconds, and you have a product that sells for $200. On average you have 1000 conversions per week, equalling $200,000 in total sales. If you were to migrate to a faster web host and bring the page load speed down to 2 seconds, at 7%, that means you could potentially increase your weekly profits by over $14,000 a week!

Kyle Rush from the 2011 Obama for America campaign site showed through A/B testing that a 3-second page time reduction (from 5 seconds to 2 seconds) improved donations by 14%, resulting in an increase of over $34 million in contributions.

There have been countless studies proving that speed does affect your conversion rates.

Time On Site / Dwell Time

In Google Analytics, average time on site is a type of visitor report that provides data on the amount of time (in minutes or seconds) visitors have spent on your website. This is sometimes also referred to as “average session duration” or “dwell time.” The higher this number, generally the better.

average time on site

A couple ways to keep your time on site higher is to write better and more engaging content that will keep the visitor on your site. Another way to structure your site in an organized fashion so people kind find what they are looking for. But speed is definitely a huge factor in average time on site as well. If a visitor gets frustrated with how fast it is to navigate your site they will most likely leave immediately. Ensure not only your homepage is fast, but your entire site.

Bounce Rate

Bounce Rate is the percentage of single-page sessions (i.e. sessions in which the person left your site from the entrance page without interacting with the page). According to another study by Kissmetrics, a whopping 40% of people abandon a website that takes more than 3 seconds to load!

Here are some typical bounce rate averages according to types of content and industries:

  • 40-60% Content websites
  • 30-50% Lead generation sites
  • 70-98% Blogs
  • 20-40% Retail sites
  • 10-30% Service sites
  • 70-90% Landing pages

One of the easiest ways to keep people from hitting that back button in their browser is to simply ensure that it loads fast! This can include even diving into more advanced optimization methods such as improving your TTFB and initial HTTP request. A content delivery network (CDN) is a great way to improve TTFB as it will help cache your static assets closer to the visitor. But one of the primary factors in this is the configuration of your web host’s servers.

How Premium Hosts Stack Up Against Shared Hosts

There are countless ways to improve page load speed on your website, such as:

  1. Using a CDN to cache content closer to the user
  2. Improving TTFB
  3. Optimizing images
  4. Lazy loading content
  5. Caching
  6. Being smart with web fonts
  7. Optimize CSS code and delivery

But the number one factor that is going to drive the overall load times of your pages down is choosing a fast web hosting provider. And to be honest, it is one of the easiest. While it is great to save money, hosting is one area in which you should be very careful as choosing a bad provider could actually affect your revenue going forward. We recommend looking at hosting as an investment in your business/website and therefore skip the shared hosting route and go for a premium or managed host. You will be much happier in the long run.

Premium web hosts typically use a much faster technology stack based on the latest and greatest in the industry. Some things for 2017 that you will want to make sure they have are PHP 7 or HHVM, NGINX, MariaDB, HTTP/2, and premium DNS options. This combination has been shown to be very fast when it comes to loading any type of website, from a static site, to Drupal, and of course WordPress.

Speed Tests

To show you just how much web hosts differ when it comes to performance, we decided to run a few tests. Here is a comparison of Kinsta, a high-performance managed WordPress host, vs a 3rd party shared host. Both are configured in the exact same way and are approximately the same distance from Pingdom’s testing server.

Kinsta Speed Test

We did a fresh install of WordPress on Kinsta, nothing is configured other than SSL. It is running the default 2016 theme. We ran 5 speed tests on Pingdom and took the average. As you can see the fresh install clocked in at 386 ms, which according to Pingdom is 98% faster than all other tested sites.

kinsta fresh install page load speed

3rd Party Speed Test on Shared Host

We then replicated the exact same setup on a 3rd party shared host. We did a fresh install of WordPress and configured SSL. It was also setup with the default 2016 theme. We then ran 5-speed tests on Pingdom and took the average. As you can see the fresh install clocked in at 1.00 second.

shared host fresh install speed test

So in this scenario the premium host, Kinsta, had 61.4% faster page load speeds! Why is there such a large difference? Here are a few reasons:

NGINX

Kinsta uses NGINX whereas the 3rd party shared host was using Apache. NGINX has been proven to be lightweight and fast. Chris Lea said, “Apache is like Microsoft Word. It has a million options but you only need six. NGINX does those six things, and it does five of them 50 times faster than Apache.”

Here is a quick look at the top million busiest websites (see below). As you can see in the trending graph, Apache is losing market share while NGINX is rapidly growing. Large websites like upwork, themeforest, wordpress.org, quora.com, and dropbox all use NGINX.

websites using nginx

Img src: NGINX vs Apache

Server-Level Caching

Kinsta, like many premium WordPress hosts, use server-level full page and object caching to deliver pages lightning fast. The shared host we were using had no such technology in place. We could fiddle with caching plugins but then you should be looking at time involved hassling with this. With premium hosts it is a lot easier to just install and go! Joe from Human Made did a great comparison of the Batcache caching plugin vs Varnish, a server-level caching approach. As you can see under heavy loads server-level caching won by a mile!

server level caching

Premium DNS

Kinsta had Amazon Route 53, which is a premium DNS provider. The shared host was simply using their free included DNS. DNS works like a phonebook for the internet, as it is a mapping of your domain to an IP address. There are slower and faster DNS providers out there. Amazon of course is one of the faster ones since they have an enormous infrastructure and network. Faster DNS lookup times can further increase the speed of your site.


And don’t forget DNS Doomsday, besides performance, a premium DNS provider is also important when it comes to reliability and uptime for your site.

HTTP/2

HTTP/2 is a new protocol designed to speed up connections with better multiplexing, parallelism, HPACK compression with Huffman encoding, the ALPN extension, and server push. Because of browser support, SSL/TLS is required to utilize HTTP/2.

The performance of HTTP/2 is a lot better than that of HTTP/1.1. For example, with server push capability HTTP/2 allows servers to respond with a page’s full contents other than the information in the browser’s cache. Efficient compression of HTTP header files minimizes protocol overhead to improve performance with each browser request and server response.

http2

In our tests above, Kinsta was utilizing HTTP/2 on their servers which the shared hosting provider was not. You can easily see this by checking the header requests in Chrome Devtools network panel (as seen below).

HTTP/2 Supported (Kinsta)

http/2 kinsta

HTTP/2 Not Supported – Still Running over HTTP/1.1 (Shared host)

http/2 not supported

HTTP/2 is simply a must now for secure connections!

Overcrowding Issue

Another issue with shared hosting is that they overcrowd their servers. Typically shared hosts don’t make any money on actual hosting customers, due to the overhead of support. For example, the shared hosting plan being tested above only cost us $10 a year!

crowded servers

They make most of their revenue on upsells and 3rd party add-ons such as domain registrations, SSL certificates, etc. Due to the fact they have very low margins they will tend to put as many customers on the same servers as they can until it starts to slow down. You can guarantee that at some point your site will be affected. And unless you have a performance monitoring tool running you will never know at what time it happens.

With premium hosts they normally go a different route. For example, with Kinsta above, they use Linux containers (LXC), and LXD to orchestrate them, on top of Google Cloud Platform which enables them to completely isolate each WordPress site. This is much more secure and when it comes to performance the Google Compute engines allow them to sale the sites automatically. The overcrowding and scaling issue is simply not an issue at all.

Summary

As you can see, page load speed is super important, as it can affect everything from your SEO, time on site, bounce rate, and conversions. Choosing a premium or managed hosting provider can be one of the smartest decisions you make for your business or blog; as it plays one of the biggest roles in improving your page load speeds. The old saying “You get what you pay for…” comes into play here. Think of hosting as an investment and choose wisely, as they greatly differ in terms of performance and technologies being offered.

Read More at How Web Hosting Can Impact Page Load Speed

Categories: Designing, Others Tags:
  1. No comments yet.
  1. No trackbacks yet.
You must be logged in to post a comment.