CDN performance
07 Sep 2020

Performance Metrics – What to use and when

Performance metrics are a set of tools that you can leverage to ensure your site delivers the best user experience possible. As everything we do these days requires an online interaction, be it a search for a specific item or listening to your favorite song, the digital services we interact with have become an integral part of our daily lives. Creating a first-rate user experience in this competitive market is vital, and developers need to factor this narrative into their solutions. 

User experience (UX) is a semantic variable. As each person’s perception differs, their emotions and attitudes about a particular product or service vary. However, the one factor that is a universal constant in creating a positive sentiment is speed. No one likes waiting for a web page to load. There are several ways to optimize your website. You could leverage a Content Delivery Network (CDN) platform to cache static resources such as text, images, stylesheets, and JavaScript files. However, before you implement any services to increase your site’s performance, you must gather metrics. 

What Are Performance Metrics?

The adage you cannot manage, what you cannot measure, is more relevant than ever in today’s cloud-first, mobile world. Performance metrics bring science into the art of web development. They allow you to create a baseline and then measure any speed improvements in an iterative manner. Without this foundation, you cannot evaluate the effectiveness of any performance improvements you implement. For example, if you enhance your site by optimizing images, you cannot scientifically gauge any progress unless you measured it before applying the change.  

In addition to creating a performance benchmark during the development process, metrics also help you manage your online sites and services. As the digital realm is a dynamic environment where the rate of change is exponential, continuous monitoring of vital user experience statistics is crucial. However, not all metrics are equal. The key to unlocking the value in performance data is relevance. Depending on the scenario, the metric you need to monitor varies. In some instances, the site’s user experience relies on getting content to the page as quickly as possible, such as an online shopping service. In other cases, getting a significant amount of content displayed is of greater importance. There will also be sites that require a combination of several metrics. In those instances, you would need to rank the importance of each measure aligning it with your target audience.

Performance Metrics Sets

Categorizing metrics that measure your online site’s performance can help you focus on the statistics vital to your service. Below are four useful metric sets that together can help you create a consolidated view of your website’s user experience score. As mentioned previously, one set of metrics is often insufficient to create a complete picture. You need a 360-degree view to measure your site’s performance effectively.

Milestone Metrics

Milestone performance metrics monitor events during the page loading process. They measure the time taken to complete the particular sequence of tasks needed to load a webpage. These metrics are incredibly beneficial when it comes to measuring user experience. 

Time to first byte

Time to first byte (TTFB) measures the responsiveness of your background services. It calculates the time taken for your webserver to respond with a connection to a browser’s initial request. A high TTFB could indicate that your origin server needs some performance tuning. It could also highlight inefficiencies in your back-end coding. The geographical distance from a server can also have an impact on the TTFB. In such cases, leveraging the services of a CDN will improve your TTFB score. 

Time to interactive

The time to interactive performance metric measures the time taken during the loading process from the TTFB until the web page is fully interactive and responding to user input. Time to first CPU idle is a similar measure. However, that metric measures the time taken to the first available interaction. On the other hand, time to interactive calculates how long it takes for the page to allow a complete user experience responding to every click without delay. 

Quantity-Based Metrics

Quantity-based metrics measure the number of requests and the size of each component. These factors typically contain elements and services you can count. Although quantity-based metrics do not provide any direct user experience measures, they are vital for long-term monitoring. They also help reveal trends and can assist developers in identifying problems with third-party scripts. Quantity-based metrics are incredibly useful during the development phase of a web application project. As they count the number of requests made to a web server, they can help you increase code efficiency by eliminating redundant calls that increase wait times and reduce page load speed. 

Rendering Metrics

Rendering metrics are a set of data points that indicate content load speeds on a particular web page. These measures can help you measure and tweak rendering performance. Their only disadvantage is that they do not take the relevant importance of different content on your page into account. 

Paint Timing

Paint timing refers to three particular rendering metric performance measures that examine the user experience as content loads on a web page. First paint is the time the browser first renders after navigation. First contentful paint is the point in time when the site loads any text or image. The third and final metric, first meaningful paint, is when the primary “above-the-fold” load occurs.

Other Rendering Metrics

Paint timing gives you one view of your site’s rendering performance. Understanding the whole picture requires a few other metrics that together provide a consolidated overview of the user experience. Start render measures the time taken from the beginning of the initial navigation to the complete load of the first non-white content. On the other hand, visually complete is the moment in time when visual progress reaches 100%. Note that this metric only measures visible progress. Other components, such as JavaScript, could still be loading. However, as it has no impact on the user experience, it is excluded as a rendering metric measure. 

Custom Metrics

In addition to the milestone, quantity-based, and rendering metrics mentioned, you may require a custom metric that calculates a performance measure unique to your site. Twitter’s Time To First Tweet is an excellent example of a custom metric tailored to meet a particular service’s user experience goals. It measures the time taken between someone navigating to their site and seeing the first tweet. 

Aligning Performance Metrics

Performance metrics are an invaluable resource in helping you deliver an exceptional user experience. They can assist you in calculating inflection points as they render in a user’s browser. However, as with any other process or framework, you need to align this testing approach with your unique objectives. You also need to measure your site’s performance regularly. As the online realm is dynamic and competitive, continuous monitoring is vital in ensuring your user experience is ahead of the curve.