How to Measure Website Performance:
Now that we know what to measure we need to know how to actually go about it. This is actually a bit more complicated than you might expect, as there are several types of tools that are better in certain circumstances than others. Remember that every time a user interacts with your website, the conditions can be very different.
One user could be within a couple miles of the origin server, while another could be across the globe. Web speed could be impacted by the traffic hitting the server while testing. Maybe you are preparing for holiday website traffic and want to know how your website’s performance will be impacted by a surge in volume.
There are several ways to measure your website, and they are more or less useful depending on your situation. We will outline these below with some recommendations for tools that we like.
Synthetic Testing
Also known as active monitoring, synthetic testing is performed by creating scripts that imitate an actual user accessing a site and measuring how long it takes for various elements to load. Synthetic testing is a fantastic way to quickly gather website performance data and identify obvious issues without having to wait for larger volumes of data.
Most synthetic testing tools provide data in the form of a waterfall, a graph that breaks down exactly when each element of a page starts loading and how long it takes for each element completely reach the user’s browser. This is a great view for identifying bottlenecks, whether it be a slow loading HTML document, a third party widget, or a particularly large image.
Unfortunately, synthetic tests have some significant limitations. First off, the metrics they provide only represent 1 browsing session from 1 specific location, so you can’t assume these metrics are applicable to all, or even most, of your users. Furthermore, a synthetic test must be scripted to take certain paths through a website, so they are not necessarily a good indicator of how an actual user would navigate through a site. The best use of synthetic testing is to help identify larger issues, or catch problems with website performance in a situation where you don’t have a lot of traffic to your site yet to collect data from.
Recommended tools: WebPageTest.org, Pingdom
Load Testing:
Like synthetic testing, load testing is a simulation meant to help you get an understanding of how your website will handle specific situations. Where it differs is that instead of simulating the experience of a single user accessing your website, it attempts to simulate how your website performs with different volumes of users hitting your website at the same time.
Load testing can help identify issues that you might encounter when your site reaches higher volumes of concurrent users. The accuracy of the data it provides, however, is questionable. It is nearly impossible to simulate real user behavior, and website performance may be impacted with a smaller volume in real users than was revealed during a load test. Load testing can also be very expensive, and with it’s limited accuracy it might not be a worthwhile exercise for every business.
Real User Monitoring:
In many ways, Real User Monitoring (RUM) is the holy grail of accurate website performance measuring tools. RUM tools measure your website’s performance for actual users and aggregates the data into the metrics you care about. You can see how your website performed during different levels of volume, from different geographic locations, how performance varies between different paths through the website, or the difference in page load speed for new visitors compared to repeat visitors. Advanced RUM tools will also break down metrics into front- and back-end load times to quickly show you what areas are in need of improvement.
The major drawback to RUM is that is requires actual data from users on your site. If you are launching a new site, or have a small number of users, RUM might not be very useful because it will take too long to start producing meaningful information. In these cases, synthetic testing might be more valuable.
Unfortunately, not every RUM tool is created equal. Google Analytics, for example, has a RUM tool that can be a useful free option, but it samples a very small number of sessions on your site, making the data susceptible to large margins of error. This tool also only shows page load time, and doesn’t break it down further. Other RUM tools are expensive and provide an extraordinary amount of data that may be useful for sophisticated developers but would be overkill for the average marketer. For those looking for these advanced tools, we recommend New Relic, which will give you detailed information on front-end and back-end performance in addition to scalability analytics and more.
Improve Your Website Performance Today
To improve your website performance, measure your page speeds using the tools above and then consider implementing a caching strategy such as Varnish Cache to immediately improve page load times. Section makes it easy for you to install and configure Varnish Cache on your website, plus we provided added website speed benefits through our CDN. Sign up for a free CDN trial today and you’ll get 14 days of free service, no credit card needed.
For ecommerce sites looking for detailed information on measuring and improving website performance, we also have an eBook available to download for free:
