Working with developers to align on technical and SEO priorities is a challenge faced by many in-house SEOs, and by SEO agencies offering recommendations. How do we start conversations and support initiatives that get developers and SEOs all working towards the same goal? Is Core Web Vitals the common ground we need?
In this conversation with Moz Developer, Lucas Rasmussen, we explore his recent project aimed at improving our A/B testing experience and how it overlapped with Core Web Vitals.
Question: What is your role at Moz?
Lucas: I’m a Web Developer at Moz. I manage the Moz website and content management system (CMS).
Question: What was the main objective of your recent Cloudflare project?
Lucas: I started a project based around making an A/B testing suite for the Moz website that focused on improving split test results and a better more consistent visitor experience. The problem we had to solve was to run client side A/B tests without a different customer/page experience. When someone loads the page as part of an A/B test, the page flashes white for a split second and it affects the experience, which affects the overall validity of the test. We wanted to do better for Moz.
We chose to create a system using Cloudflare, where Cloudflare automatically shows two different types of pages. This way we could build a system where the A/B test page loads just as fast as if it wasn’t an experiment.
I had an ambitious goal of getting average page load time across the whole site down to two seconds.
Question: How long did it take from ideation to completion?
Lucas: All-in-all it took around 2-3 weeks to complete with an additional two weeks of planning. This also involved changes with our CMS, and a few misplays along the way.
We needed some help from our engineers, learning how Cloudflare workers actually work. They are very powerful!
The core work took one week in its entirety, working out what needs to be done — getting feedback, responding to that, and actually doing the work.
Question: How are you tracking the results?
Lucas: I’m tracking my results in the Cloudflare dashboard specific to Web Analytics. We are currently limited to 30 days of tracking, I’d love to see more to see changes over time.
It might be worth noting that if you want more data, Moz Pro Performance Metrics section of Site Crawl displays historical data for up to 90 days for tracked URLs.
I’m keeping an eye on what’s going on with the page load time, especially the request time. When the timing goes up, that’s a flag that there is a problem somewhere. It indicates to me that something isn’t cached.
Looking back at our ambitious goal of getting average page load time across the whole site down to 2 seconds. We have currently plateaued at 2.6 seconds. But we are tracking a large portion of users across the whole site.
Question: What was the most enjoyable part of the project for you?
Lucas: Turning it on and seeing the impact and change to page load time — l Ioved being able to see real-world results. And in this case IT WORKED. There are so many changes you can make and you think they are going to change something, and even if you know they are going to make a difference, you might not see the impact… When I changed users to cached the difference was significant, from around 1,500 milliseconds to 200 milliseconds.
Question: What do you know about the importance of Core Web Vitals?
Lucas: I do have visibility into Core Web Vitals as a concept. LCP (Largest Contentful Paint) in particular is a metric I track in the Analytics dashboard.