5 Questions for Tammy Everts
by @tammy@webperf.social and @j9t@mas.to (@frontenddogma@mas.to) on , tagged interviews, performance, metrics, user-experience, economics (toot this?)
Tammy Everts has spent more than two decades studying how people use the Web. Over the years, she has worked on groundbreaking UX studies that have involved EEG headsets, facial action coding, and Google’s machine-learning system.
As CXO at SpeedCurve—a UX monitoring platform with customers from Airbnb to Zillow—Tammy’s focus over the past 10+ years has been the intersection of UX, site speed, and business metrics. Her book, Time Is Money: The Business Value of Web Performance from O’Reilly, is a distillation of much of this research. She is co-chair of the annual performance.now() conference in Amsterdam. She also co-curates WPO Stats, a collection of web performance case studies.
Jens: Welcome and nice to meet you for a brief interview on Frontend Dogma! You are particularly focused on performance. How has this developed over the years, both in terms of our field, as well as you personally?
Tammy: In the early days of web performance, measuring page speed was pretty basic. There was a small number of tools that could run one-off synthetic tests—essentially rough simulations of how a page would perform in a specified browser/connection/geolocation context. The metrics that were available were simple and mostly just told you how long it would take for a page to fully render. Because of this, the learning curve for performance newcomers was short and sweet, though it’s arguable how much insight we had into actual user experiences.
Over the years, everything exploded. There are so many tools available now. Real user monitoring (RUM) tools, which measure how pages perform for actual users, arrived on the scene to complement synthetic tools. There are literally hundreds of potential metrics—some of which are pretty arcane—that you can track across your entire site. All this complexity helps us get closer to understanding actual user experiences, but it also makes for a very long, steep learning curve.
Jens: What are you working on these days?
Tammy: Interestingly, I’ve looped back to where I was at the beginning of my performance journey around 15 years ago. Back in the early days of web performance, there was very little understanding of how page speed intersects with user experience and—importantly—with business success. And no one was talking about performance in terms that a non-developer, non-engineering audience could digest. So a lot of my early research and writing and conference talks focused on educating people and trying to make performance a mainstream topic.
Five years ago, I assumed that portion of work was done and everyone had gotten the message, but I was wrong. I’m finding huge new audiences who need to be educated. They genuinely have no idea that web performance is one of the biggest indicators of online success. So I’m dusting off my original research, revisiting old assumptions, and exploring fresh ways to reach new people.
Jens: Is there anything that has you particularly excited about performance right now?
Tammy: I love talking about performance. When I think about performance, what I’m really thinking about is time, which is the most precious thing we have. Humans are not hard-wired to sit and stare at screens while we wait for things to happen. One of the beautiful things about our brains is that they crave seamless, elegant experiences. Pursuing those seamless experiences—whether it’s playing music or baking a cake—is hugely integral to our mental and physical health.
So it stands to reason that non-seamless experiences—such as waiting for a janky web page to load—have a negative impact on our health. When people get impatient about waiting a few seconds for a page to load, it’s too easy to be dismissive of their feelings, but those feelings are valid. All those precious seconds add up!
When I’m talking about performance to new people, I love to lead with the human angle. We can’t all relate to abstract performance metrics, but we can all relate to the feeling of being frustrated by technology.
Jens: What’s one particular challenge around performance you think we could pay more attention to?
Tammy: As I mentioned earlier, it’s easy to feel overwhelmed by all the tools and metrics out there, but that’s not an excuse to ignore the performance of your site and hope it’s good enough.
I encourage everyone to start small and simple. A good monitoring tool should allow you to get everything you need to not just track metrics, but also get alerts when you regress, get specific recommendations about what to fix, and see the impact on your business, whether it’s bounce rate, conversion rate, or whatever success metric you care about.
It doesn’t have to be hard. Start with just one metric, like Start Render, which measures when content starts to be visible on the page. There’s no one perfect metric, but the good things about Start Render are 1) it’s easy to understand, 2) it’s supported by almost every browser (unlike Core Web Vitals, for example, which until recently were only supported in Chrome-based browsers, though now they’re also supported in Firefox), and 3) there’s a lot of research that shows improving Start Render correlates to improved business and user engagement metrics.
Track Start Render over time on just your key pages. Baseline your best and worst times, and then identify a threshold that you don’t want to exceed. For example, if your worst Start Render time is 3 seconds, your threshold would be 3 seconds. Some tools will allow you to create a performance budget, so you get an alert when your threshold is violated. Then drill down into your synthetic test details to see what you need to fix, fix those things, and keep on monitoring. Building a fast page from scratch isn’t that hard. Keeping it fast is where the real work lies.
Jens: What are you most concerned or excited about in development in general?
Tammy: Looking at the really big picture, I’m concerned about the massive tech layoffs that are happening across so many organizations. Yes, sometimes layoffs and restructuring need to happen, but there’s more than that going on these days. I see companies with record year-over-year profits shrinking their teams in order to generate short-term profit, and it worries me. I hate seeing hard-working folks lose their jobs, of course. That’s always gutting.
But beyond that, those hard-working folks were making the web stable for all of us. It’s kind of like that game Jenga, where you slowly remove blocks from the tower until eventually the tower collapses. The tower is fine until suddenly it isn’t. That tipping point applies to complex systems as well. It takes a lot of people to make the web reliable. How many of those people can we remove before we hit the tipping point? Nobody knows. So looking long term, what I really fear is the destabilization of this vast entity that underpins so many of our essential systems. We’ve become too reliant on the web to let it crumble.
Having said all that, when I talk with young developers (like my teenager) and other people who are new to tech, it always lifts my spirits. They’re so hopeful and excited, and it seems like they see through a lot of the nonsense and noise. Given the right opportunities and support, I believe they can make the next version of the Web something we can feel good about using.
Jens: Thank you so much for your time and insights, Tammy!