Tammy Everts - Catching Up With Web Performance

Catching Up With Web Performance

Tammy Everts

00:00 / 51:34

Video

Transcript

Tanner: Hello, everybody, and welcome to Catching Up With Web Performance, a podcast about stories of people in web performance. Today my very special guest is Tammy Everts. Tammy, how’s it going?
Tammy: Hey! It’s going really well, Tanner, actually. Like, I don’t know when you’re going to release this podcast, but for me, it’s like the week before Christmas, or the week and a bit before Christmas, so feeling Christmasy.
Tanner: Oh, Merry Christmas, Happy Holidays! All the things. Is there any Santa Claus stuff sitting in the background of the house right now?
Tammy: Oh, probably. Yeah, my office turns into Santa’s workshop over the holidays. It’s like, “Where can we hide this? Oh, put it in here!” So it’s a little bit nutty right now.
Tanner: Well then, speaking of gifts, I would love, my request from Santa this year is if I can have the gift of your story. I would love to know, yeah, great segue.
Tammy: Smooth.
Tanner: Tammy, like, this web performance thing, what is it and how did you get into it?
Tammy: I don’t know if I can answer the first question easily. The second question, we can start there. So my background is not in performance at all. I’m not a developer. I’m not an engineer. I am a person with an undergrad degree in English literature and a graduate degree in publishing.
Tanner: That’s fantastic.
Tammy: And I did my master’s, yeah, I did my master’s back in the late nineties, and I decided to specialize in “digital publishing” as we called it then. And I was strongly dissuaded from doing this by my thesis advisor, who thought that the Web was a fad and that I was totally just going down the wrong, like down the wrong alley. Like, strongly. I can’t tell you how vehemently he tried to persuade me not to do this.
Tanner: What’d you say back?
Tammy: I said, “I feel good about it, though.” So I kind of just persevered.
Tanner: “I have a hunch.”
Tammy: And so I did my master’s in digital publishing, and really glad that I did. And so from there I basically started working for a web agency, and kind of just, you know, it was the early days of the Web and the whole like web agency model, and so they just knew that they wanted kind of a “content person”. And I’m gonna do air quotes a lot because a lot of what I say feels ridiculous to me, so that’s how I signify it. So I was the content person, and what that kind of turned into was like, okay, words on the page, beyond design, how the pages related to each other. So I basically kind of taught myself how to do information architecture as we did it back in the day. Taught myself how to set up a usability lab in our agency, leaned heavily on the resources of Jakob Nielsen and the Nielsen Norman Group for just kind of like a lot of practical things. And like a lot of people in my generation are just really self-taught.
Tammy: So with that experience behind me, in I want to say 2009, I was contacted by a friend, like a former coworker, who was like, “I work for this company called Strangeloop and we do web performance optimization.” And those were all just like gibberish to me, I did not know what that meant at all. But they were like, “We need a content person.” So I was like, “Okay, I know those words, I’m familiar with that request.” So I went in and chatted with them and that was kind of the first time I encountered the idea that site speed mattered in any kind of fundamental business way. Like intuitively, you know, I’m a human being and I’ve been annoyed by slow websites and, you know, I had a lot of experience of being annoyed by slow websites. But just this idea that there was some business value attached to that, and there were things that you could do to make that better, was brand new to me, because my usability testing experience was very lab-oriented. We weren’t testing for speed. We had like, you know, all the sites were blazing fast, or the prototypes or whatever we were using, so that just never came into bearing. So we would build these websites and kind of push them out in my previous career and kind of be like, “Okay, well there your website’s done.” And so this was brand new.
Tammy: So I kind of just learned on the job at Strangeloop and kind of in other places that I worked. I worked at, Strangeloop was acquired by Radware, and then I moved over to Soasta, I worked with a lot of great people there. Soasta’s now been acquired by Akamai. It’s very, you know, obviously it’s tech, so things move around. And now I work for Strange, or not Strangeloop, another company with a similar name.
Tanner: We’ve come full circle.
Tammy: SpeedCurve. Not a strange loop, a speed curve. So I work at SpeedCurve, and I’ve worked here for the past five years. Yeah, so that’s my long answer to your question.
Tanner: Well, no, I feel like you brought that in pretty compact.
Tammy: Oh, good. Okay.
Tanner: There’s so many, there are so many things, though, to pull out here. Like, let’s just start with the gibberish. Because this fascinates me, I love talking to people who know absolutely nothing about web performance. Can you go back to, you said this is 2009, we had Strangeloop and they say, “Come write content, come write web performance content for us.” And you were like, “What that? What is web performance, I don’t know what to do.” Like, do you remember that moment? Or like, did they say “web performance” over the call? Or when was it that you had to actually write a web perf article?
Tammy: So web performance was right, like their tools were called WPO, web performance optimization tools, and you don’t really hear those terms anymore, so that was like right there in the name from the beginning. And it was in the blog that I helped create, Web Performance Today, which people who’ve been around for a long time might have encountered. And we had a podcast called Web Performance Today Podcast, and yeah, another web performance podcast.
Tanner: More podcasts!
Tammy: So that was always there in the name. Some other terms that were really new to me, and again I had to learn on the job, were just like, what is optimization? You know, learning how to use WebPageTest kind of in the very early days of WebPageTest, that was one of the first things that I was taught. And kind of coming full circle, SpeedCurve is built on a forked instance of WebPageTest. We’ve got a lot of our own bells and whistles. So it’s like, you know, that was a really good thing to know. How to read a waterfall chart, you know, that was…
Tammy: Again, fortunately, I worked with really, really great people like Joshua Bixby, who’s at Fastly, and Hooman Beheshti, who’s also at Fastly now. And they were really great about teaching me kind of everything they knew, and there was like no question was too stupid, and it was fantastic. So, you know, I think that’s kind of one of the nice things about this community, is that it’s very open to explaining things.
Tanner: Yeah, tell me about it.
Tammy: Because there’s a lot to explain. It’s, I mean, those are early days, and it’s gotten a lot more complicated since then. Like the metrics glossary that I used to rely on in 2009–2010, which was like “start render”, “load time”, “backend time”, seems so easy and basic compared to all the metrics that we’re looking at now. So it’s a lot.
Tanner: Yeah. I guess I’m so interested in what you’re interested in. Because you have this, gosh, how long is this now? You know, we’re hitting, you’ve been even doing this over a decade.
Tammy: Yep.
Tanner: Are there highlight moments or things, like, what has kept you going this long in web performance? And what are some of the things that excite you or interest you?
Tammy: That’s a really good question. It’s something that I kind of come back to a lot. At the end of the day, I don’t really care about technology for technology’s sake, and I used to be really kind of shy about sharing that, and now I’m like just shouting it off the rooftops. Like, I don’t care about tech. I love the Internet, I love content. You know, I like everything that the Internet does for me. The mechanics of how the Internet works or how to like, you know, you’ve got my attention from maybe like 2 minutes and then I’m gonna just start drifting off.
Tammy: What I’m interested in, and what I really find web performance is kind of a fascinating way to study this, is how people actually engage with the Internet, how people engage with websites, what the Internet does to our brains. So it’s not just how we use the Internet, it’s kind of how the Internet shapes us. So some of the research that I’ve participated in over the years, so kind of back in my Strangeloop and Radware days, I was really lucky in that I got to direct some research projects, where we did things like, a little bit harking back to what I used to do in my usability testing days, where we got people in lab settings and tested things like throttling their network connection and seeing how they behaved, putting EEG headsets on people and getting them to perform tasks on a slowed down website, and doing actual like A/B testing in these types of environments.
Tammy: So you could see, you know, if you have two cohorts of people who are given the same tasks, on the same websites, and some of them have a slower experience, and you kind of aggregate the data that we get from the EEG headsets for each group, and then you overlay them, what are the differences? What are the anomalies? And seeing everything that you kind of think, you know, or if it feels true to you, about how people use the Web, or how people feel when sites are slow, you can actually have a visual representation of this. Like we can, because the EEG data kind of mapped to the timeline of how people were going through their whole user path, you could see how frustration actually spiked in their EEG data when the pages got slow.
Tanner: That’s interesting.
Tammy: And on specific pages, you know, like on checkout or things like that. So actually getting to see what’s going on with people. So you kind of realize this isn’t something you have control over. Like I’m not a Zen master who can kind of, you know, control my sine waves through all stressful experiences. Like when something’s frustrating…
Tanner: I’m a human being!
Tammy: Yeah, exactly. I have a human reaction, and it’s hardwired and it is what it is. I can take some deep breaths and go to my happy place but, you know, that doesn’t scale. So that was really, really interesting to do. We did a few of those studies. And doing exit interviews of people who… Sorry, I’m going on and on because this is really interesting stuff for me.
Tanner: No, no, this is why we’re here, Tammy. This is why we’re here! Keep going on and on!
Tammy: Oh yeah, I meant to go on and on. Okay, good. Good, good. Okay, so… Yeah, so doing exit interviews with people as well where, and I should make clear that we were never telling people that they were part of a site speed experiment. They didn’t know that they were…
Tanner: What did you tell them?
Tammy: We told them that we were just doing regular usability testing, so we just wanted them to do specific tasks for us and let us know how they felt about the website and issues with doing the task. So in no way did speed come into play as something that we discussed with them.
Tanner: We weren’t priming them for this. We’re just saying come in and click some things and tell me how easy it was, maybe.
Tammy: Yeah, tell us how that felt for you. And one of the things that was really telling was when we did exit interviews, which is pretty standard for usability testing, you just do exit interviews with people where you kind of get their general impressions after you’ve done all the tracking during the course of their using the site, was we said, “What were your impressions of the site?” And we took the words that the people in the fast cohort used, the adjectives they used, and threw them into one of those like word cloud generator thingies, and we took all the adjectives that the people in the slow group used and put them in the word cloud generator thingy, and what came out of those word clouds, and I have visuals I can send you, they’re really, really cool…
Tammy: Yes, please.
Tammy: Where the word cloud for the frustrated people, the people with the slow experience, were huge compared to the other ones, like they wanted to talk more. And the words they used, some of them kind of said like, “Oh, it felt sluggish and slow,” which was interesting that they picked up on that, because they had no baseline to compare it to. They only got the slow experience. But they also said things like, “Oh, the navigation felt clunky, the design felt inelegant.”
Tanner: And it’s the same experience?
Tammy: Same website.
Tanner: The only thing you changed was the speed.
Tammy: Yeah, the sites and tasks were exactly the same. But the things that they were critical of were design, usability, navigation, structure. Like content, they thought it was boring. So it’s just really interesting how when something’s slow, you just get more critical overall. So that was really interesting. And that was with a fairly large study group. I think we had like 50 people that we tracked and kind of aggregated their data. So yeah, really interesting stuff.
Tammy: So it’s kind of nice to have that foundation to really feel like when we’re coming into using monitoring tools like SpeedCurve or other tools and telling people like, “Yeah, you really do need to be paying attention to this,” and trying to make your sites faster and all this stuff matters, is to feel like, yeah, it really matters on a very fundamental level.
Tanner: Right. And it feels like too, it can spin off in so many strange, indirect effects. They said it was boring, it had nothing to do with the speed. It was just, what? Now they have this completely different perception of me.
Tammy: Oh, and they said the design was tacky. That was another word that came up, “tacky”.
Tanner: Yes! Oh, wow. That’s fascinating. Are there any other things that surprised you during these studies or during this time?
Tammy: I mean, that was kind of the, I would say if I had to draw on the research that I’ve done, the correlation between speed and like tackiness and sluggishness and design and everything like that was probably the most surprising.
Tammy: I think one of the things that I find really interesting, I don’t know if it’s surprising, is… Again, I keep going back to Nielsen Norman Group research because I think they do a really good job of studying these things. And I think that people in the UX community are very familiar with the research but people in the performance community maybe not so much. So one of the things that they do is they revisit, like every ten years or so, what are people’s expectations in terms of how they perceive page speed and what their feeling is. And they have kind of this set of thresholds where like anything under 100 milliseconds feels instantaneous, under 500 milliseconds feels pretty fast, and so on all the way up to 10 seconds. And then after 10 seconds, there’s like a big kind of context shift for people, and it’s really hard, they’re shifting focus, and it’s really hard for them to get back on task. And what I like about the fact that they keep doing this and revisiting these old assumptions is that the takeaways are always the same. These thresholds don’t change for people.
Tammy: Interesting.
Tammy: And so it kind of speaks to the fact that we don’t learn to take kind of a different attitude, or have different expectations of site speed. Like our reactions are hardwired, probably to do with the fact that we’ve only been using computers for about 30 years, as opposed to everything that we did as human beings for the tens, hundreds of thousands of years before that, where you had to do everything quickly. Like, you know, oh, it was like I’m gonna go and, you know, stab a stegosaurus. I know that we didn’t have stegosauruses back then but…
Tanner: That’s gonna be my new favorite example. “So you remember the time you had to stab a stegosaurus? Yeah, that’s why we need websites to be fast.”
Tammy: I’m just making sure people are still paying attention. So yeah, everything that we had to do, you know, out in the fields, out hunting and gathering, doing all the things that human beings needed to do, are flow activities. This is what we’re designed to do. And a lot of people, maybe not everybody, is familiar with the idea of flow that’s been popularized over the last little while, in general, that you need a certain amount of flow activity in your life on a daily basis to be happy. And if you don’t have that, that’s not good. But if you use the Internet a lot, especially on your phone or with a choppy connection or something like that, you’re having the opposite of flow. You’re having like fits and starts…
Tanner: Which, real quick to backtrack, hit me with flow. What is this thing?
Tammy: Okay, so the idea of flow… And I’m not going to try to pronounce his last name.
Tanner: Is this Zen? Is this peacefulness? Am I in the zone? Is it I’m good at something?
Tammy: No, it’s a researcher. Wait a second. I’m going to pull up… This is the handy thing about having… Um, his first name is Mihaly…
Tanner: You’re gonna try and pronounce his last name? Hang on. I’m gonna see if I can do it.
Tammy: No, I’m not. Okay. It’s, it’s…
Tanner: Alright, let’s see.
Tammy: So it’s Mihaly, and now I’m not gonna pronounce his last name.
Tanner: Csikszentmihalyi? [Flow]
Tammy: Mihaly, and his last name starts with a C. Oh, there you go! Oh, you were trolling me.
Tanner: Hey, that’s all I do here. We’re how many episodes? This is like 15 episodes in, and I just don’t want anybody to know what I know. I want to not know anything. I’m here to ask the dumb questions and to keep you talking.
Tammy: And try to make me pronounce things I can’t pronounce. But he, so basically he’s a psychologist and he kind of identified and popularized this idea of flow as being something that we actually need. And so, you know, this is a measurable human need. And so the flip side of that, as I’ve said, is that if you are doing things that are kind of the opposite, kind of like making your brain operate in these little fits and starts, so you’re constantly trying to get back on task, it’s really dissatisfying and frustrating, it will make you very frustrated. So I’ve leaned on this idea in some of the other work that I’ve done.
Tammy: For example, gosh, I’m trying to remember which of the studies it was, it might have been one of the EEG studies, where we had our data and our findings reviewed by somebody, by a psychologist, who didn’t work on the study. And we were like, “Well, why do you think that this is?” And his theory, which is in the appendix of the study, is that our brains operate on glucose, and we use glucose throughout the day, and when we get things like hangry or, you know, we’ve made too many decisions over the course of the day, like work, we’re familiar, probably most of us, with that feeling of coming home from a busy day at work, and it’s like, “What am I going to have for dinner? I don’t know,” like you can’t make a decision anymore.
Tanner: “I don’t care, give me cake!”
Tammy: Yeah, like you just lost it. And so his theory is that when our brains are constantly having to come back on task, it’s forcing these extra little spurts of glucose that are making you have to kind of make the choice to come back on task. And we get frustrated because it’s basically our brain’s way of telling us, “Hey, you’re using glucose at an unsustainable rate. You need to stop this complete, like, right now. This is really dangerous. Like I’m going to starve and die.”
Tanner: That’s interesting.
Tammy: So I thought that was really, really an interesting way to think about it as well. So for me, it kind of comes back to first principles, of like how much of this is hardwired and how much of this hard wiring is actually physiological. So you can’t say, “Oh, well, people will get used to this,” or, “We’ll train our brains to be different.” Like our brains don’t train to be different over a lifetime. Like maybe our descendants 200 years from now will have different brains that allow them to use sputtery technology better, but it’s not going to happen in one lifetime.
Tanner: Yeah. I think it’s so cool that you got somebody else to come in and audit, or give that, you know, unbiased opinion on the work.
Tammy: Yeah. Well, it’s necessary…
Tammy: I’m trying to think of just, I’m just thinking, you know, like… In my day to day, like, how often do I actually get somebody who is out of, how often do I get somebody in my day to day work to look at what I’ve done and review it? I dunno, there’s something about getting that external feedback that I think is fascinating, maybe.
Tammy: Well, it’s very satisfying when you’ve gotten it, but it’s really nerve racking until you get it. Because basically, like part of the process when you do this type of research is you do have to send it away and get it vetted. And it’s like, what if they say that, oh, there was some kind of error in how we did our our research, or something like that, and we just spent like tens of thousands of dollars on a project that is kind of like ehh, you know?
Tanner: Yeah, like could you imagine redesigning a website and spending months working on it, and then launching it, and all the sudden it’s slower, and somebody tells you you did a bad job?
Tammy: Yeah, yeah. Well, and it’s funny because I was just thinking about this this morning. And I was thinking somebody should, I would be really curious to see a survey result, like how many people working on websites today, or apps, does any kind of usability testing?
Tanner: What’s a usability test? I’ve never done that.
Tammy: Yeah, that’s the thing, right? And it was a hard sell 20 years ago and I don’t think it’s gotten easier. Like I think, you know, the desire to build new features and things like that is, it far outpaces or outstrips people’s desire to slow down and test what they’re doing.
Tanner: Tell me about it.
Tammy: Bummer Thoughts 2000.
Tanner: Bring it on! I’m here for it. I just, one of the things that fascinates me, picking off of this, what smooth segue, off of having somebody else come in and audit your work is… I mean, performance feels like it’s been, and maybe it’s not a just, has it objectively been a niche topic for years? It feels like any time I want to evaluate the performance of my work, I need an expert, I need to get somebody else’s opinion. I need to go ask Tammy, “Hey Tammy, can you look at my site? Did I do a good job here? I don’t really know how to do this myself.” Is my experience unique in that way? Or I feel like you talk to a lot of people, maybe?
Tammy: Yeah.
Tanner: It’s like, what’s it like out there? Is that what performance is for people? That, “No, you need to go get an expert to come talk to you.”
Tammy: It’s kind of, yeah, I absolutely get that. I mean, the world that I live in now, like working at SpeedCurve, and this isn’t going to turn into a product shill or anything like that, is like that’s kind of the point, what tools are supposed to do for you, is they are meant to be that validator. So, you know, the very hand-wavy sort of solution is, okay, well figure out what metrics you need to focus on, get a tool that lets you measure, like, track those metrics over time and, you know, set thresholds or performance budgets or whatever you want to call them against your current kind of status quo, and make sure you don’t regress. And then kind of have some goals, and things like that, that you want to get to that might be based on industry case studies or benchmarking yourself against your competitors.
Tammy: So the tools could do all these things, but kind of the flipside of that—and also there are consultants out there, like we have a consulting practice at SpeedCurve. There’s lots of really good independent consultants out there who can help you get set up with the tools, teach you about all of this, validate your work, let you know what you can still focus on, all that kind of thing. But I think the thing that’s challenging now that’s different from ten years ago is on one hand the tools are better and there are better metrics, so that’s really great and a lot where people are using real user monitoring, which is basically actually measuring performance kind of in the wild and actually measuring, you know, how people engage with their sites, but also how their sites perform out there in the real world, alongside synthetic testing, which is like simulating how your pages would perform in different contexts and on different browsers and things like that.
Tammy: So the tools are definitely more sophisticated, the metrics are much more nuanced, which is great. But alongside that, just means it’s that much harder to learn. So I feel like we’ve solved one problem and kind of necessarily by doing that created a new problem. So when I think of challenges in web performance right now, it’s just teaching people how to use the tools.
Tammy: And kind of a corollary to that is the people who want to learn how to use the tools, they’re excited, they kind of get that this is really important. But the tools require an investment of time and money because, you know, the good ones definitely cost some money, but more importantly, they cost time. Like onboarding to use a new tool that you want to have be integral to your organization is a big deal. Like the money is almost incidental. It’s like when you go to get a dog or a cat and I’m like, “It costs that much to get a dog or cat?” And it’s like if you think the upfront cost is the biggest investment in getting a dog or a cat, you shouldn’t get a dog or a cat. It’s kind of the same way with any kind of actual important tool. Like your real investment is in the time that you’re going to spend getting really good value out of it.
Tammy: So before those people who want to use the tool can invest in the tool, they’ve got to convince other people in their organization who don’t know about web performance, don’t care about web performance, or feel like, “Oh, we’re probably fast enough. We’re about the same as everybody else.” They’ve got to get them to care. So there’s this whole kind of like daisy chain of effects, where it’s, you know… Teaching people to teach other people how to care about performance is a big part of what I do as my day-to-day job.
Tanner: Man… I have so many of these moments where I want to jump in and just dive off on some complete tangent. I’m gonna cut this, but I came really close to going, “Oh, and Time Is Money, am I right?” Maybe I’ll throw that in there anyways.
Tammy: Thanks.
Tanner: But so let’s, I’m gonna wrestle this back into stories here. Help connect a little bit for me. I’ve got Strangeloop 2009 and I’ve got SpeedCurve today. How did you get from one to the other? What was the journey in between?
Tammy: Oh, for me? Okay, now I have to really think.
Tanner: Here we go.
Tammy: So it’s kind of like, just working in tech, it’s a little bit of a tale of acquisitions and happenstance, like kind of just doing this, so…
Tanner: Like, were you in through that whole Strangeloop, Radware, Soasta, and then who acquired Soasta, Akamai?
Tammy: Akamai acquired Soasta, but at the time I kind of had… They were lovely, they invited me to join Akamai as well. But I’m also very good friends with Steve Souders and Mark Zeman, who are the basically partners of SpeedCurve, and when I had an opportunity at the same time to join SpeedCurve, it was like, how do you say no to working with Steve and Mark?
Tanner: Steve Souders, the man himself. How can I say no?
Tammy: Yeah, so for your audience, who might not know who Steve is, like some people sort of jokingly but not jokingly call him the godfather of web performance. So he’s kind of the reason why a lot of us have jobs and why a lot of these tools exist. Like he wrote the original web performance books, explaining the rules as they were back in 2000 and… I want to say 2006.
Tanner: 2007! Hang on, we got it…
Tammy: Are you gonna go to your library?
Tanner: I gotta do this. This is like my parlor trick, that I can just pull books out of the air.
Tammy: Nice.
Tanner: Oh, you mean this one? [High Performance Web Sites]
Tammy: That one right there, yeah.
Tanner: Yep. And don’t worry, there’s Even Faster Web Sites on the shelf.
Tammy: There’s the sequel, yeah. So, and Steve worked at Yahoo! and then at Google and he led web performance in both those places, did some of the earliest work on RUM, like real user monitoring, that’s out there, and so really kind of like has set the stage for a lot of things. And he was a co-chair of O’Reilly Velocity, which was a long running conference, it was around web performance. And he’s just one of those like connector people. He just knows everybody and tries to really, like, make the tent as big as it needs to be so that everybody can be inside it. So yeah, so he and Mark are wonderful. So I had to come to SpeedCurve, so I’m really happy that I did.
Tanner: Around what time is that?
Tammy: So it’s actually coming up on 6 years in 2023.
Tanner: Happy anniversary.
Tammy: Yeah, 2017. So long ago. And honestly, it’s the longest I’ve ever worked at the same company, which is kind of a…
Tanner: No!
Tammy: Well, you know, it’s tech. Like one year is actually three years.
Tanner: But that’s a pretty cool, I guess, superlative there. It seems like there must be something special going on at SpeedCurve then if it’s the one, the place that Tammy stayed the longest. Like this place kept your interest.
Tammy: Okay, so in fairness, I will say, like, I’m gonna touch wood here, I have been very, very lucky, #blessed, to have in my entire working life I have had, I’ve worked at really great places with great people, so I’ve not had like a scarring, you know, terrible tech experience or like any bad bosses.
Tanner: No war stories.
Tammy: So I’ve, like, great bosses, great coworkers everywhere I’ve been. So some of my shorter durations at different companies are because they got acquired or, you know, something happened, and so I really just kind of went from one company to the other company, and so those were actually big spans of time.
Tammy: But no, SpeedCurve is really special. Like we’re a small team, everybody has tons of experience, very flat kind of organization, a lot of autonomy, and everybody cares about web performance, like a lot. So we’d, and I mean, a lot of other teams could say the same thing, so I’m not, you know, we’re not unique.
Tanner: But it’s literally what you guys do.
Tammy: It’s literally, but it’s like what we do, like we have conversations about metrics that are like impassioned.
Tanner: I picture, is there ever a standoff between you, like the group standing on tables and shouting at each other?
Tammy: No, no. It’s no, we’re all really nice. So that’s really helpful.
Tanner: Nice but passionate.
Tammy: We get excited and we start talking over each other in Zoom calls and it’s like, “And what if we did uh, eh, ah?!” But fortunately, Mark and Steve keep us really focused. And yeah, so it’s just like we really care about performance. We come from really diverse backgrounds. So Mark’s background, and he’s somebody you might want to consider interviewing as well, because he actually comes from a design and agency background.
Tanner: Yeah, that’s me too.
Tammy: He built SpeedCurve on top of WebPageTest because he was using WebPageTest to show things to his clients, because he wanted his clients to understand that the sites they built, it wasn’t just about having them look really pretty, which he’s good at doing, but it was about having them actually be very fast and performant. So he was using WebPageTest a lot, but he wanted a different interface, or different outputs, because, you know, everything needs to be a bit glossier and things like that. And so he built kind of the first iteration of SpeedCurve to do that, and then partnered with Steve and kind of helped build on a lot more nuance and things like that. So that was kind of the forking that happened with Pat Meenan, WebPageTest founder’s blessing. And Pat’s been really instrumental in helping us and make our tools better as well.
Tanner: So they start this thing, and then at some point they call in Tammy. They say, “Tammy, we don’t know anything about digital publishing. Can you write content for us?” Or was it, did you get the pitch? Or what was the pitch this time? How’d you get roped in?
Tammy: Well, we knew each other through Velocity. So I had been to a bunch of Velocity conferences and so had Mark and Steve. And so one of the things that I love about going to the same conferences, I’m kind of an introvert, but you after you’ve sort of seen the same people a few years in a row, it’s a little bit like summer camp. And so even an introvert will start to feel like, “Oh, I kind of belong here. This is good. I think these are my people.” And so that’s how I met them. I did a talk at Velocity, and Mark has spoken at Velocity, and Steve was the co-chair, so that’s how I met them. So I was kind of, I guess that employee one at SpeedCurve, which is… Usually, like this is sort of my experience, it’s like company builds a really good product, and then they need a content person, and then if I’m lucky I get called to do that.
Tanner: Wow. So I just want to know, like, what was, you come in to this new thing, SpeedCurve, it’s this tool on top of WebPageTest, like what do you start writing? What are you doing? Do you come up with all these ideas and just pump it out?
Tanner: Yeah, the tricky thing is just finding, you know, I think everybody can relate to this, like trying to find time to make content. Because when I first started at SpeedCurve, kind of what I knew and what I came in knowing how to do because I’d been doing it pretty much exclusively at the previous places I’d worked, was just writing articles, doing interviews, talking about performance, doing research and that kind of thing. So it was very much in the research, you know, content, speaking at conferences, kind of vein. And so I still do that at SpeedCurve, but what my job has kind of morphed into, and I’m actually really glad that it has, because I think that’s one of the things that I think at this point when I’ve been doing the same things for so long, you need to have a little bit more, augment it with a little bit more, is I talk to a lot of customers.
Tanner: Yeah.
Tammy: So previously, I was kind of in this researching/writing bit of an ivory tower. Whereas now, like I talk to people. I talk to hundreds of SpeedCurve customers a year at really big companies and small companies, and I show them how to use our tools, and I talk about what their performance issues are, and it’s really educational. Like you realize just what the actual pain points are. And there are a lot of them are kind of cultural and organizational as well as technical. And it gives you just a lot of empathy for people, like kind of what they’re going through, and the fact that the problems don’t really change. So that’s been really great. So I do a lot of talking to customers now.
Tanner: Yeah.
Tammy: On top of everything else. So it really has influenced how I create content, you know, but also just how I think about performance.
Tanner: Yeah. Well, how so? What are some of the ways? Count the ways! Are there any things that come to mind when you think of, you know, I think it’s interesting that you said, “I used to be in an ivory tower writing my articles, and now I actually, I talk to people.” Yeah, how is it changed the way you think about performance? How has it changed, how has it changed you? How has talking to people changed you?
Tammy: So one of the things that I’ve realized is that if I had, if I were still in the ivory tower, I think I would have assumed that everyone is on more or less the same trajectory and knows more or less the same things about performance. And what I’ve learned from talking to people is that at any given time, a huge chunk of people are new to this. So very smart people who are learning everything for the very first time. And so I’ve learned that… I think if I didn’t have this knowledge and this like firsthand awareness, I would probably be like deeper in jargon land than I am. And it’s something that, you know, you have to really fight. When you’ve been doing anything technical for a while, it’s just like this kind of accretion of assumptions and jargon that you so internalize that it doesn’t sound like jargon to you anymore. And so even now, like where I really strongly subscribe to plain language and plain writing in everything that I do, I’ll still catch myself, like, “I just said a thing that probably didn’t mean anything to whoever heard it,” and I have to catch myself and go back and explain things. And so a big part of being in my world and talking to people is just going, always sticking with first principles and making everything as simple as possible. And it’s not because people are stupid, it’s because people are new.
Tanner: Right.
Tammy: So I’m just trying to kind of take the frustration out of learning about this, to make it easier and faster for people to onboard themselves. So yeah, that’s a big thing for me. And so I’m always kind of going through our support docs, like we have a really good support hub at SpeedCurve, and kind of just like, “What’s missing?” Like what’s the thing that we’re not explaining? Or how could we be explaining this better? And I’m not the only person at SpeedCurve who feels that way. Like we’re all pretty big proponents of easy UI, easy copy, just like how can we… You know, it’s a product. It’s a smart thing to build ease into your product. It kind of sounds like a no brainer. But I just, as a human being who recognizes the humanity in other people, I just don’t want to, I want to help. I want to make life easier for people.
Tanner: Yeah. Yep, yep, yep. I’m particularly interested in words, in the ways that we talk about performance. In addition to my book parlor trick, I also like collecting metaphors or analogies about how we describe performance. I feel like you said one earlier and I may have missed it. But I’ve been trying to keep a list of the different ways I describe performance. One of the first ones I go to is like, “This is how a car works.” “What if you were in a restaurant?” I recently started doing sports analogies for performance. So I’m curious, in your digging up, or what’s the word for getting rid of accretion? As you’re like, as you’re…
Tammy: Dissolving?
Tanner: Yeah. As you’re tilling the soil of your language, what words do you catch yourself saying and how are you replacing them to make sense to other people?
Tammy: I think… I don’t know about like other terms.
Tanner: Have you caught yourself recently, this moment that you described where you said, “I’ve had to catch myself, they probably didn’t understand what I said”?
Tammy: Yeah. So talking about the tools I’ll say like, “Oh, it’s just synthetic testing,” and a lot of people don’t know what synthetic testing is. Or real user monitoring, RUM, you know, we throw out these acronyms. The metrics are probably a big one, because sometimes we get so caught up in what the metrics are teaching us that we don’t think to slow down and explain what the metrics actually are.
Tammy: So like thinking about Core Web Vitals, for example. It’s a set of three metrics that Google is popularizing and they’ve made part of how they measure user experience and they’re also a ranking factor for search ranking. So these three Core Web Vitals are now kind of on a lot of people’s tongues, if you’re an SEO person or a marketing person and a performance person. So that’s really nice and I applaud that effort, for sure. But what the web vitals are are First Contentful Paint, or sorry, Largest Contentful Paint, First Input Delay, and Cumulative Layout Shift, or, you know, LCP, FID, “fid”, or CLS. And suddenly everybody’s talking about CLS and FID and, you know, if you’re new to that and you’re just hearing it, it really just, it’s just…
Tanner: Disorienting.
Tammy: It’s just letter salad. Like it doesn’t mean anything. And so it’s kind of like…
Tanner: That’s a good TLA.
Tammy: Yeah, another TLA, Three Letter Acronym. So just kind of saying like, oh okay, well Largest Contentful Paint is just like what’s the most meaningful content, how is that rendering, and Cumulative Layout Shift is like how janky is your page, and First Input Delay is like how quickly does a page respond. So those types of things, like actually making sure that we’re talking about metrics in terms that people can actually understand and feel they can trust, or feel that they’re actually relevant.
Tammy: I think that for me, coming from a usability background, I like the fact that metrics exist. I think metrics are a really heroic attempt to apply proxies to humans, human behavior. But it’s really easy to lose sight of the fact that they’re just proxies. And the only way to actually measure human behavior is if you were to follow every one of your users around and watch them use your app and measure how they use your app and see what page… Like having seen this in a lab, I know that what you learn in a lab is so profoundly more nuanced than what you’re going to learn from a number in a spreadsheet or on a chart. So labs help teach you empathy, and kind of really internalize empathy for users, numbers on a chart don’t. And the problem with empathy, as I’ve said in a few talks, is it doesn’t really scale well.
Tammy: So the idea of numbers on a chart is really, really abstract from an actual person using your site. And you’ve got to constantly work to recognize that your number is helpful but it’s not real, in the sense that, you know what I mean? It’s just, it’s just an aggregated bunch of numbers that is giving you a general sense of a trend. And I think it’s safe to say that and still be a proponent of using these tools, because they’re the best thing we have. They’re the only thing we have. And it’s better to use them than nothing, because if you use nothing, I mean, there’s a reason why companies that invest in performance, companies that use tools, tend to perform better than companies that don’t.
Tammy: If you look at companies that are, like large companies, that are currently struggling, and you can find, I’m not gonna name any names, but this is the kind of thing I track on the side out of interest. Because there’s a business component to all this as well. Like I care about users, but I get that to get other people to care about users, we have to make them, we have to convince them that this actually affects their business. If you look at companies that are hurting financially right now, companies that are in receivership or whatever, you could say, “Oh, it’s the economy, blah, blah, blah.” Except that we also know that so many companies in this economy are doing really, really, really well. And one of the biggest differentiators between the companies that are doing well and the companies that aren’t doing well, if you were to do a bunch of synthetic tests through any publicly available synthetic testing tool, create your own benchmarks dashboard, pick the ten best performing companies financially and pick the worst ten performing companies financially that you can find, and measure the performance synthetically of some of their home page and a few key product pages or whatever, you would see, as I’ve seen, that the slower sites are the ones that are struggling. Like it just, it matters.
Tanner: Oh man, we’re gonna get off on a whole ’nother thing about correlation now. Let’s talk about correlation, Tammy.
Tammy: Oh boy.
Tanner: Oh man.
Tammy: Yeah, so correlation charts are something that we do at SpeedCurve, and you can get them in other RUM tools. So correlation charts are a really great way to kind of… And we call them “correlation charts” for a reason, because we’re not trying to say there’s a cause and effect here. But what a correlation chart can show you is how the performance metrics for your site correlate to user engagement metrics like bounce rate, or time on page, or things like that, session length, or cart size, or conversion rate, or whatever it is you’re tracking for your real users. And so what we typically see in correlation charts is that if you, depending on your page, or depending on your site and what metrics you’re tracking, often performance does correlate, like faster cohorts of users have higher conversion rates or lower bounce rates, that sort of thing. And I’ve created a gazillion charts that show this mostly, and there’s always going to be some exceptions, but the trend is it mostly shows this. So yeah, there’s a really high correlation factor as well.
Tammy: If you want to do more of a, like, actual testing this to see how it actually plays out on your own site, because I wouldn’t rely on a correlation chart. I mean, they’re helpful if you are trying to teach people in your company, like, “Look, there’s a correlation here. We should care about this and it merits more investigation.” But if you actually want to see the impact of performance changes on your business metrics, then you need to do an A/B test or a multivariate test. Where you, it’s kind of what I was talking about earlier with the usability studies, you need to, and we used to do this at Strangeloop as well, you know, if our customers wanted to, is you optimize most of your traffic and then you don’t optimize like a small cohort of your traffic and it’s the same site, just one is optimized, one is the other. Same site, it looks the same, it’s just one is slower, one is faster. And you could see it every time we did this type of A/B testing. In fact, some of our customers were like, “Okay, stop it, we get it.” Like they’re supposed to be, it’s supposed to be a 12-week test, and they would stop it after 2 weeks because they realized how much money they were losing.
Tanner: Right.
Tammy: So you can probably, if you do an A/B test, see that this affects your business as well.
Tanner: Man. I am so sad that our time is coming short.
Tammy: Ohhh.
Tanner: Because I have a million other things that I want to talk about. Tammy, it has been an absolute pleasure. I would love if we could wrap up here in the last minute or two, like, do you have any parting words of wisdom? Maybe something you would give your Strangeloop self, some words of comfort or wisdom. And then after that, how can people follow you? Stay up to date.
Tammy: Oh yeah.
Tanner: Know what’s going on in the world of Tammy.
Tammy: So parting words of wisdom. I think the reason why so many people are still doing web performance, you know, 10–15 years down the road is because it feels like it’s this onion that just has infinite layers that you can peel. And I think it’s very, it’s fascinating. And the metrics keep evolving, and the tools keep getting better, and our understanding of user behavior keeps evolving, so it’s just not static. And that can be frustrating in some ways for people who are just like, “I just want it to stay still so I can measure it.” But as somebody who works in this space, it’s really, really interesting. And I advise people to keep pursuing it if you’re just getting started. There’s also lots of great people in this space, so I would say people are really awesome as well.
Tammy: Find the thing that you care about that’s kind of like your, that’s your main thing, and that makes you feel passionate about this, and I think that there’s something in that for everyone. I think that where I feel really lucky is that the thing that I care about, which is mostly just user experience, actually everybody does care about. Like the most technical person doesn’t just love spreadsheets and numbers, you know, they want to see how it connects to actual human beings as well. Because at the end of the day, we all want to feel like our work actually has benefit, like it’s helping in some way.
Tammy: What was the, oh my goodness, what was that show? I’m truly drawing a blank. It was about web performance, it was the comedy. It takes place in Silicon Valley…
Tanner: Is it Silicon Valley?
Tammy: It’s Silicon Valley! Yeah, that’s it.
Tanner: It took place somewhere, I can’t recall.
Tammy: What was it? So I actually did watch the show religiously, I thought it was pretty funny. And it was like, “Wow, they made a sitcom about web performance, how weird is that?” And the guy who’s the main character, Thomas Middleditch plays main character, he’s actually from the town that I live in, Nelson, which is pretty funny. Yeah, which is like this little ski town in B.C. He’s from here. So anyways, but they always would say, “We’re here to make the world better.” And every tech company was basically saying, “We’re here to make the world better.” So I think we all do kind of want to make the world better, and so seeing yourself as like kind of a, you know, somebody who’s part of that is really helpful.
Tammy: And where people can find me. So I don’t know if people are still using Twitter? It’s like @TamEverts.
Tanner: Yeah, we’ll see when this gets published where things are at.
Tammy: But I’m on LinkedIn too, of course. And then there’s also a Mastodon called WebPerf.Social and I’m there and I’m just @Tammy@WebPerf.Social.
Tanner: Well that’s fantastic. Tammy, thank you so, so, so much for coming on the show today.
Tammy: Thank you.
Tanner: It’s been a blast talking. I would love to keep talking another 7 hours, but maybe someday. Maybe, if I’m lucky enough, I’ll get to see you at one of these conferences.
Tammy: Yeah.
Tanner: Well, until then, thanks. Until next time.
Tammy: Bye.

Recent Posts

Archives

Categories

Links

Video

Transcript

Leave a Reply Cancel reply