Testing the speed of comment systems for blogs
For blogs and many other types of sites, getting an interaction going with readers is essential. But it would seem that many believe that installing a third-party comment system on a website, something like Disqus, IntenseDebate, or LiveFyre, drags down the site’s speed considerably.
Instead of spurring on debate, discussion, and interaction, a slower site could discourage users to take part, and users could instead end up leaving the site. But is it really the case that adding these comment systems slow down sites?
We put five comment systems to the test and found out that there’s less difference in speed than you might think.
Comments are important
First, let’s establish that comments are important for many websites, not just blogs. We think comments are a critical component of our blogs at Pingdom, and we have posted before asking for feedback on which third-party comment system we should go with, if any. As you can see, we still use WordPress’ built-in comment system, but we are right now looking at alternatives.
For example, in December we noted that Disqus has a commanding lead in the Technorati Top 100 blogs, where it had a 22% market share. In the comments to that article we got many good opinions and suggestions as to what we should do.
But if installing a third-party comment system will slow a site down, then it’s something that many webmasters will not want to do. Since users like speedy sites – customers are often won or lost in mere seconds – webmasters obviously take this very seriously.
Number of files loaded (requests)
So we put five of the more common comment systems to the test:
You can read more about how we did the testing at the end of the article.
First, we looked at how many requests were made when using the different comment systems. With our testing site, we saw 6 files being loaded in with just plain WordPress, without any third-party comment system. Disqus and LiveFyre each loaded 9 files, IntenseDebate 11 files, and Facebook 14 files.
You would perhaps expect that this difference in number of files that are being downloaded mean a big difference in the total size being requested when the site is loaded. However, there is not much difference between plain WordPress and three of the third-party systems: Disqus, IntenseDebate, and LiveFyre. For example, the difference between WordPress and Disqus, was only about 4 kB.
The only comment system that really stuck out in terms of page size was Facebook, which required almost 47 kb extra to be loaded compared to plain WordPress. Although this, in and of itself, may not seem like much, it’s very likely a website will also have other plugins and additions installed, each requiring something additional to be loaded, so the page size can then quickly increase.
But of course page size and number of files don’t necessarily accurately reflect how fast or slow a website is as perceived by a user. Here’s how the sites fared when we looked at the time it took to load them:
What this showed us is that again there is not that much difference between the five different comment systems, with a few exceptions. WordPress is, in each instance, the fastest one to load, but it’s only when testing from Amsterdam it crushes the others.
When we then finally combined the results from each location, and looked at an average loading speed for each of the comment systems, this is what we got:
Again, as you can clearly see, there was not that much of a difference between using WordPress’ built-in comment system and any of the other four, third-party systems. We would even go so far as to guess that there was less of a difference than you would have thought before reading this.
Not a big drag on site performance
The conclusion must be that there is not such a big difference in website performance between running just plain WordPress and having one of these third-party comment systems installed. That is, with the possible exception of Facebook.
Of course this is just a very limited test, which doesn’t necessarily reflect real-life cases, where multiple users access a site running multiple plugins, and other modifications. Where servers are located also play a role, as does where content is located (on a CDN, for example).
That said, we think it indicates that installing a third-party comment system does not necessarily drag down a site’s performance. At least there is not a big difference in performance in our tests.
Could there be other things adding to the issue of users perceiving sites with these add-ons installed as slow? Let us know what you think.
Take our poll
We created a simple, one-question poll, trying to gauge what our readers would prefer to see implemented here on the Royal Pingdom site.
How we tested: We tested on a WordPress site, which was hosted in Sweden. It had the latest WordPress, 3.3.1, and all plugins were up to date. However, we disabled all plugins other than the ones needed for each comment system to work. We also used the default WordPress theme with no modifications, just so we could get a test that was as fair as possible. For each comment system we used the corresponding official WordPress plugin, with the exception of Facebook, for which we used the Simple Facebook Comments plugin. We tested ten times for each comment plugin from all three testing sites that our Tools uses (tools.pingdom.com/fpt): Amsterdam, The Netherlands, Dallas, Texas, USA, and New York City, New York, USA. From the ten tests we then took the average for requests, page size, and performance.
Speech bubble image via Shutterstock.