As an SEO company that offers website hosting, it’s our responsibility to make sure that our client’s websites not only load fast as lightning but also have consistent uptimes.
Staying true to this commitment, this past January Mad Fish invested time and resources upgrading our hosting infrastructure. We had wanted to make this upgrade for some time, as we were estimating that we would reach capacity on our old server sometime in June 2014.
I know what most of you are thinking.
Why in the world would we want to put ourselves in the position of host for client websites?
We found early on in our company’s history that regardless of the host that a client chose, there were still at least one to two hosting issues per year, and regardless of the problem, we were the client’s first phone call when issues arose. We constantly ran up against poor (or non-existent) customer support while helping clients troubleshoot issues. We also constantly ran up against performance issues with websites and spent time and energy trying to work with unresponsive hosting companies or educating clients on what they needed to communicate to their hosting company in order to make improvements.
All big wastes of time.
It didn’t take us long to realize that to truly be responsible for a client’s entire web presence, we needed to have our own hosting services.
Why the upgrade?
Our old hosting server was doing a great job. But newer technology became available which allowed us to put together a faster server at a lower operating cost. Also, as browser technology improves, many of our client sites are more media rich, requiring increased disk space. When our old server was assembled, we made sure to utilize all of the hard drive bays and available ram slots thinking we would never exceed it. Based on our client’s (and our own) current needs, it became quickly apparent that in order to stay true to our hosting commitments, it was time to upgrade.
Freeway lanes, traffic jams, and load averages… oh my
The need for an upgrade did not solely center on the need for more disk space. In 2013, we saw a lot of traffic growth to the websites we host. While this is a great bi-product of our client’s seeing on-going success on the web, we started to see higher load averages during peak times of days.
What’s a load average you ask?
The load average can be thought of as the number of cars driving on a freeway. Typically, traffic is moving at an even pace throughout the day, but during rush hour, the freeway starts to get congested.
In the server world, if your freeway gets gridlocked, the server can crash and shut down. This frequently happened when mom and pop companies were mentioned or endorsed by celebrities like Oprah back in the day. Oprah would include them in one of her ‘favorite things’ shows and the next thing they would know millions of visitors would use their website. This flood of traffic all at once would cause traffic congestions that would crash the website.
In the SEO world, there are two key traffic types that we want to keep happy when it comes to avoiding gridlock and downtime.
- Visitors/Customers: If you don’t have reliable uptime for visitors they won’t buy your product, or comeback to your site.
- Search Engine Crawlers: If you don’t’ have reliable uptime for crawlers, then your rankings could begin to suffer.
Again, another reason why, as an online marketing agency, hosting our clients sites makes sense.
So how do we insure that our client’s won’t be affected by gridlock?
We’re big believers in the Litespeed web server software over the standard Apache software for serving websites. This is because Litespeed handles traffic “better” than Apache. It also serves web pages faster and more efficiently. Litespeed runs on top of Apache, but has a monthly licensing fee, which can deter many hosts.
By using Litespeed, we’ve found that during traffic jam situations, website traffic speeds still function normally and website performance remains continuous. This is great for clients but doesn’t change the fact that the freeway is still getting clogged. To avoid impending traffic delays, and keep the freeway clear, we made the call to upgrade our hardware.
How’d it go?
Total time for the migration took us about a week (originally planned for three days), and without going into all the details, we did manage to avoid hardware downtime. The day after the new shiny hardware began being used in production, we did have several client site’s who needed edits made to their code to work in the new environment, however in the end the sites were all better for the code changes.
Did this impact Results?
Our hosting client’s sites all load faster and we at Mad Fish can sleep easier knowing that our hosting reliability is covered for some time to come.
While most websites on our server loaded well under 3-5 seconds prior to the upgrade, we saw load time improvements to most sites from 30% to upwards of 70%. This means if a website loaded in 3 seconds, now it loads in just under 1 second. This is without the use of a CDN for most sites, so you can imagine how much quicker websites end up loading when they utilize one.
If you’re an existing client reading this, I really appreciate your patience during our upgrade in January. If you’re not yet working with us yet, and want to find out how we can help you grow your business online, and improve website speeds, feel free to reach out.