You can’t improve what you don’t measure. But what if you’re measuring the wrong thing? There is a dizzying array of metrics and tools to measure website speed, many of which are contradictory, some of which are ideal for static content, and others for dynamic content/database-driven websites.
At the same time, modern web technology has enabled a new breed of app-like websites, such as eCommerce progressive web apps (PWAs) and single-page applications (SPAs), which deliver amazing speeds. However, these websites appear to load slower than they actually do when measured by most of the popular speed metrics, which were created to measure legacy, multiple-page websites. As a result, you could be missing out on critical business growth simply because you are using the wrong tools or metrics to measure website speed.
If you are running a revenue-generating eCommerce website, picking the right metrics may be the most important way to align your teams around the goal of creating a fast site.
This article will tell you exactly what metrics and tools to use to measure eCommerce website speed in 2019/2020 for maximum revenue impact. Consider this your how-to cheat sheet.
Four tips to measure website performance for maximal revenue impact
Here are four simple tips to filter out the noise and get you going:
1. The single best metric is Largest Contentful Paint (LCP)
Did you know waiting for a page to load creates the same psychological discomfort as watching a horror movie? You can significantly decrease the risk of scaring customers away by measuring and optimizing for the right perceptual metrics.
When the largest image above the fold has loaded (i.e. on a homepage, category page, or product page), the more a consumers perceives a page as fully loaded and are more likely to make a purchase. A few page load metrics measure this milestone, including Last Painted Hero (LPH), Largest Image Rendered (LIR) and Largest Contentful Paint (LCP), however LCP has become the standard metric per Google's 2020 Page Experience Update. LCP is one of three Core Web Vitals, and is now included in SpeedCurve.
LCPs up to 2.5 seconds are considered fast and rank in the 75th percentile for this metric. Optimizing for this metric has been found to be better correlated with eliminating the psychological stress associated with slow page load, than Time To Interactive (TTI), First Meaningful Paint (FMP) and other common metrics.
SpeedCurve and Google do a great job explaining why these are the best metrics to measure website speed.
2. Measure browsing speed, not only the first page load time
Most benchmarks focus on the first page load of a website. While that metric is very important, especially now that Google ranks faster sites higher in its search results, most shoppers spend much more time waiting for pages to load while browsing through products and categories. In fact, Wolfgang Digital found that the average shopper spends just over 3 minutes browsing through six pages on an eCommerce site, with 21% of this time spent waiting for pages to load.
Fast first loads increase search ranking and thus organic traffic, while faster browsing transitions improves conversion rates. Optimizing and measuring both is critical. Fortunately, two free tools can be configured to measure browsing speed: WebPageTest.org for synthetic tests (email firstname.lastname@example.org to get test scripts that show you how) and Firebase for real-user measurements (RUM).
3. Use 4G settings, not 3G to measure speed
Many tools still default to measuring performance under simulated 3G conditions. But according to the Chrome User Experience Report (CrUX), the most comprehensive RUM data in the market, most consumers are probably on 4G, even outside the US. The pitfall of using 3G is that you may end up optimizing different things and slowing down or making little improvement to your 4G speed, which will actually hurt your overall revenue.
4. Use both synthetic and real user measurement (RUM)
Synthetic tests determine the performance of a website by simulating a page load inside a testing tool and can run "on demand" whenever you need them. RUM measures page load speeds actually experienced by real shoppers on your website but you have to wait to aggregate these metrics from the browsing activity of your site visitors. Both must be tracked. Synthetic tests are important as part of your continuous integration and build processes so that you can ensure no release causes your site speed to regress. RUM is the ultimate measure of what site speed your shoppers’ experience.
In late 2019, it is possible to create instant eCommerce websites with median LCP as measured by RUM on the mobile web of 0.3 seconds (both first loads and browsing transitions), which is literally the blink of an eye.
Don’t leave money on the table, schedule a consultative conversation to learn how the Moovweb XDN can help your company achieve sub-second load times, increase organic traffic by an average of 32.1%, and lift conversions by 15-30%.