Performance is one of our key goals in any web project. It’s also something we think we’re pretty good at, and we use it as a selling point for our services. However, it can be something that’s tricky to measure objectively, despite the fact that it’s almost universally acknowledged that page load speed is one of the most significant drivers of user satisfaction (see this nice infographic by Kissmetrics).
Page load speed, of course, can vary enormously even for a single page depending on factors such as the device and browser used, the location of the user and server, caching performance and the amount of traffic on the server, to name a few.
So it isn’t much use to tell somebody that their site loads in two seconds, or half a second, because it’s largely circumstance specific. We really need more static metrics. With this in mind we generally start with page weight and server request quantities as key metrics, but these don’t tell the full story either, nor is it necessarily always apparent to clients what they mean, or what is a ‘good’ weight or number of requests.
We try and take a range of page speed measurements as well using one of the many tools available online, and provide comparisons to other sites and give general targets, but they don’t necessarily help to explain why a stunning page full of retina-display ready images mightn’t be the best idea to somebody on a new Mac Book with 50 Mbps broadband.
We’ve spent some time, therefore, coming up with a range of metrics we use as standard when undertaking site audits, which taken together help to give a reasonably coherent picture of a site’s performance without requiring technical knowledge to interpret:
We also include a few recommendations that would improve performance as part of the audit to show how working with us could add value to a company’s digital resources. Performance is a crucial factor in the return on investment for any website, and it’s worth investing time in helping clients to understand its impact and how it can be measured and improved.