Major news broke in the web performance world last week following an announcement from Google. The company will discontinue Google PageSpeed Insights a popular service used by developers to analyze and optimize website pages for performance. Every site currently using Google’s PageSpeed Service will have to make the necessary DNS changes by 3rd August or the site will break. (Here are PageSpeed turndown instructions for developers)

Big Picture: How will this impact the web performance world?

The larger impact of PageSpeed’s closure will unfold as developers begin the search for replacement solutions. Google itself has suggested EdgeCast Edge Optimizer (Verizon), as a content deliver network (CDN) alternative. Other industry watchers have pointed to CloudFlare (also a free tool), as the natural landing place for former PageSpeed users.

As a basic tool, PageSpeed worked well to help developers to adhere to general Front End Optimization (FEO) guidelines. In fact, Yottaa has previously touched on this topic in the blog article, How to Get The Best from Google PageSpeed Insights, which details how to get the most out of the tool while avoiding certain pitfalls.

At this point, we should draw a distinction. Google’s PageSpeed is to be considered a ‘best practice’ tool only. Meaning, it is not adequate for holistic performance optimization — it’s more of a benchmark for checking whether certain tactical steps have been taken. Similar replacements will have the similar shortcomings.

How to Approach & Tackle Modern Web Performance

Mastering web performance is a difficult and complicated undertaking. It encompasses more than just FEO, and best practices do not always align with reality. Top web developers understand this and look beyond doing FEO for FEO’s sake. The best course is to think about performance in terms of providing a great end user experience, rather than knocking out steps on a static to-do list. That means considering everything from the backend servers and DBs, to the content of the web pages themselves, and everything in between.

To demonstrate this point here are just some of the advantages an holistic approach holds over a tool like Google PageSpeed.

• Effective Solutions: Certain PageSpeed recommendations are irrelevant or have minimal impact on real web performance, depending on the context. For example, domain sharding is largely irrelevant because improvements in browser technology have made it redundant. Similarly, parallelizing requests won’t move the needle for all but a few exceptional cases. By trying to apply general best practices to every site in an automated fashion, it’s easy to be misled. An adaptive, holistic approach would prioritize the steps that would have a significant impact on end user experience.
• Contextual profiles:Today, one of the most important features of a web application is adaptability, or how it responds to different user contexts. That includes a wide range of devices, screen sizes, and connection speeds. PageSpeed sees a more or less black-and-white world of mobile and desktop that doesn’t reflect the messy world of actual user experiences. Adaptive approaches use extensive testing and rules-based optimizations to accommodate a huge number of potential user situations.
• Rules Engine and Automation: GooglePage Speed’s open-source utility requires manual coding, implementation, tuning and management (in short, lots of additional coding) to deliver the sophisticated optimizations that are essential to improve user experience. With a rules engine and machine learning, this laborious task is made easier and scalable with an adaptive solution.

Trying to optimize a modern web application using only GooglePage Speed is equivalent to entering a boxing ring with one hand tied behind your back. Indeed, the tool was never intended to function in that manner. To tackle modern web performance developers must be well equipped with a purpose-built tool that is up to the task.