What trends are you capitalizing on right now? The interesting thing is that the fundamentals of a fast website aren’t trends—just as the fundamentals of SEO in general remain constant. Take content, for example: keep writing for the user instead of going overboard. Of course, things like PWAs and SPAs come and go, or Google introduces a new metric (such as INP in March 2024). On LinkedIn, I mainly try to provide information about these topics, so I wouldn’t really call it capitalizing on a trend. But if I had to name one thing: Google is promoting “Speculation Rules,” and that could certainly be interesting for static sites. At the same time, however, more than 90% of my cases involve e-commerce, and in those situations, a page speed best practice that works for a static website might not work for an online store—or it might involve many more nuances.
In your opinion, what do you think will be the biggest challenge for page speed in the future? Frameworks. The amount of JavaScript used in frameworks is constantly increasing. Moreover, the ease of getting started with a framework can lead to developers becoming less familiar with the fundamentals of the web. Here’s an example: as browsers continue to evolve, you need less and less JavaScript these days—whether it’s for lazy-loading images or even building an interactive image gallery. If a developer isn’t aware of this and blindly relies on what a framework does, the site may continue to serve unnecessary JavaScript for years to come.
How do you determine which tasks should be prioritized during the optimization process? I base that primarily on knowledge and experience. A Lighthouse report aims to give the reader an idea of the impact of a recommendation. However, Lighthouse (which is a synthetic test) and Core Web Vitals (which are based on a subset of real users) are not the same. As a result, you might spend days working on a Lighthouse recommendation that, in practice, yields no visible results. Before I actually share my recommendations, I always start by analyzing RUM (Real User Monitoring) data. This gives me a clear picture of factors such as internet connection and device type, since the target audience always varies from one website to another. I combine this data with the values shown by each individual metric to assess which best practice or current anti-pattern should be given higher priority than another. By combining this with my knowledge of how browsers work and what has worked in other situations, I can prioritize my recommendations appropriately.
SEO and page speed optimization are hard work—how do you measure the results of your efforts? I’ve actually already given that away: with RUM data. Google’s free Core Web Vitals data can be a starting point. But it doesn’t measure everything, lacks nuance, has no filtering options for the data, and is 28 days behind. When a development team implements my recommendations and the marketing team opens the floodgates to third-party tools a week later and unleashes them on the site, it becomes difficult to prove with Google’s data that our technical efforts led to improvements. Real-time RUM data strengthens the evidence, which is why I consider real-time data indispensable.