Transcription
Thank you everyone, thanks for that introduction, Marc.
Before I get started, I sat through every talk yesterday and they were all so amazing and compelling and interesting and moving.
And I just, before I even get started, I wanted to maybe ask everyone to give a huge round of applause for all the speakers who spoke yesterday.
It was amazing.
It was amazing.
So even though that's a really hard act to follow, I know that the speakers coming after me today are also really amazing and I'm really excited about watching all of their talks as well.
So without any further ado, let's get started.
This is my very dirty, this is an actual photo of the stereo in our family car.
It's very dirty and dusty because we do a lot of camping and hiking and that's what it looks like in the summertime.
That's not part of this story.
The part of this story that is relevant to what I'm about to talk about is the part where my husband left the house a couple of weeks ago to go and pick up our youngest child.
And when they came back home, our youngest child, his name's Will, Will came in the house and he came right up to me and he was like, mom, and his eyes were huge, dad broke the car stereo.
And I was like, what?
It's a car, how can you break a car stereo?
And so my husband told me that what happened was that he was just trying to get some music started and he was hitting the same button on the screen and nothing was happening and he was hitting it over and over again, not with incredible force, but this just sort of happened.
And my husband's actually a very mild-mannered, kind of non-aggressive person.
He is very kind to children and animals and the elderly and usually to inanimate objects, but this happened.
And don't worry, we can still play the stereo, we just have to use our phones to activate it so we're okay.
But if I were to write a paper about what happened and actually do a little bit of a study on this and if I had been fortunate enough to have, say, an EEG headset on my husband while he was trying to activate the car stereo, I might have entitled the paper The Impact of Latency on Human-Computer Interaction.
But in our house, it's just come to be known as that time dad broke the stereo.
That's the story.
So a little bit about me and why I'm leading with a story about my broken car stereo.
So this is where you can find me.
I am currently the CXO of a company called Speedcurve, where we help lots of companies, big companies, small companies, to monitor the performance of their websites, basically to understand how fast they are and what impact the speed or slowness of their website might be having on their users and on their business.
And how I got here, oh, sorry, this is where I live.
This is going to become relevant later on in the story.
I live in a very small ski town in British Columbia, Canada.
Like I said, this will come up later, so I'm showing you right now.
That arrow more or less points to where my house is in the neighborhood, and this is kind of what it looks like in the autumn.
I didn't take this photo.
From a very, very early age, I have been kind of obsessed with time, and I don't think I realized that this was a bit of a personal obsession.
I think I thought that maybe everybody cares about time, everybody's interested in the passage of time.
For me, it happened late one night, the year that I was seven years old, and I was lying there, maybe having a hard time falling asleep, and it suddenly occurred to me, and I can actually take myself back to that moment with very little effort.
It occurred to me that I was in a time stream that was moving inexorably forward to my ultimate demise.
And that was a pretty heavy thought to have.
I had to go wake my mom up and tell her, and she was very surprised.
But ever since then, that sort of stayed with me, this idea that time is really all we have.
It's our most precious commodity.
I'm not telling you anything you don't already know.
For some of us, maybe this thought is a little more omnipresent in our daily lives than others.
It is for me.
Maybe ten years ago, I would have been comfortable being interviewed, or sorry, introduced, people saying I was a usability expert, I've been doing usability things for about ten or so years.
But now it's more than ten years later, and I've learned enough to know that I don't know very much.
This is classic sort of Dunning-Kruger, I know what I know, I know how what I know is just scratching the surface.
If we think that we have a complete understanding of how users use the web, then we're kind of kidding ourselves.
If we think that the current tools and technologies that we have in place give us a complete 360 degree view of how people use the web, we're also kidding ourselves.
So I consider myself a student of usability, and I will be to the end of my days.
I'm fascinated by how people use the web.
Web performance was a brand-new term to me about 12 years ago.
Like I said, I've been doing usability things for about ten years, mostly involving doing usability tests in lab settings with real users.
And I was kind of, I had stopped working to have my children, and I was kind of looking for work again.
And this is testament to the power of LinkedIn, and I'm in no way affiliated with LinkedIn, but I decided I need a job, I guess.
So I created a little profile for myself on LinkedIn, and literally the next day, somebody who was in the set of connections that LinkedIn suggested for me, people that I probably knew based on their email addresses or something, contacted me, and to this day I'm embarrassed to say I have no idea how I know this person, but he seemed to remember me, and he said, you should come and work for the company that I work at.
And I was like, what do you do?
And he was like, we develop web performance optimization tools, and those were all brand-new words.
I mean, I knew the words, I had never heard them in that set, you know, that particular combination.
So what is web performance?
What does it have to do with usability?
And it was kind of a revelation for me, as it was explained to me.
When I got started in web performance, it was very much about measuring server backend time or measuring availability and downtime and things like that, and the impact that that might have on users or on your business.
And at the time, this is being around 2010, 2009, there was this kind of emerging idea that, like, actually, maybe performance can be measured in other ways, and maybe we can correlate those to user behavior.
And the revelation for me, as I became involved in this industry, and I did end up working for this company, they wanted to move beyond just looking at availability metrics and uptime metrics and things like that, to actually measuring performance from a user experience perspective.
And so the revelation for me was that time was this missing vector for all of the things that I had been investigating and researching in my career up until that point.
And I would say that if you don't consider time to be a crucial usability factor, you're missing out on an extremely huge aspect of the user experience.
Something that a lot of people may not know, I certainly didn't know it until somebody told me, was that end user response time, 80 to 90% of it happens at the front end, not the backend.
What that means is that the way that the browser renders your images and your third parties and everything else is more important from a usability perspective than just server response time.
So if you are a developer, if you're building code on a page, if you're a designer and you're responsible for the design framework and everything else on the page, if you're the content person who's responsible for the words and images and everything else that appears on the page, and if you're a third party vendor or somebody who adds those helpful third party tags around social media and analytics that we like to add to our pages, then all of these things are web performance issues.
All of them are user experience issues.
So you may not know it, but by virtue of working in any of these areas, adding any of these things to your pages, you are a user experience person, you are a web performance person.
So I've done a few things since then.
I wrote a book for O'Reilly a few years ago called Time is Money, because I am very interested in the intersection of performance, usability, and business metrics like conversions and time on site and bounce rate and things like that.
It's sad to say that this is the capitalist world that we live in.
Performance and usability didn't really become or start to become mainstream issues until we figured out that there was kind of a monetary aspect to it.
So blame capitalism.
I also co-curate a website called WPO stats, which is a repository of case studies that you can go to if you want to see, oh, you know, what's the impact of page slowdowns or page speedups on bounce rate for media sites.
There are case studies on WPO stats that you can find.
Things like this.
I plucked a few examples from WPO stats.
Like for example, Vodafone improved performance metrics by one second, improved their bounce rate by 14% and their conversion rate by 13%, which is really, really significant.
So there's a lot of other great case studies that you can find there.
And that's kind of what I do as part of my day job.
But for a lot of people, case studies aren't really enough.
You can argue that, well, you know, maybe the site got faster or people converted more or stayed on the site more because the content improved or you had some kind of sale or special offer.
You made some design change or UI change that made people want to do whatever it is you want them to do.
And so part of what I've been doing for the past 10 odd years is to strip away that aspect of the business KPIs and actually just focus on what is the impact of performance on people at a real and kind of fundamental human level.
And I've been really very lucky to have had opportunities to direct some research.
For example, there's using EEG headsets.
I kind of mentioned those earlier, where we put EEG headsets on people and kind of monitored how they felt as a site slowed down or sped up.
And I'll be talking more about that later.
Similarly, doing interesting things with latency and studying people's faces, actually.
So something called facial action coding, where we actually can look at microexpressions in people's faces and see how they're responding to pages in real time.
And we record these things and we analyze them.
So I've done a lot of these studies.
And the really, really exciting and super surprising conclusion is that waiting is hard.
Let's go home.
Waiting is real.
And passive waiting.
Waiting when nothing seems to be happening is even harder.
Our perception of how time passes is more important than reality.
And I would go so far as to say that perception is reality.
So let's talk a little bit about that.
How do we perceive time?
If you were here yesterday morning and participated in the activities that Emily suggested for us, you know that 25 seconds can feel like a really, really long time.
Like a really long time.
If you've ever done a talk at a conference like this, you probably know that the last four or five minutes before you have to come out and do your talk is probably the longest five minutes that you'll experience.
If you've ever given birth and had an unmedicated labor, you'll know that time definitely has its own way of passing.
So lots of different ways that we can kind of appreciate in the real world how our perception of time is very, very subjective.
But there's some interesting studies around this as well.
So we can, as academics, we can kind of like take this into a lab and look at it.
And we can look at real world examples.
So this is an interesting case study from an airport in Houston.
So a major airport in Houston, where they were getting tons and tons of complaints about people having to wait at the baggage claim for way too long for their bags to arrive.
And so what the airline did was a very logical thing.
They hired more baggage handlers.
And so they, by doing that and kind of having people hustle to get the baggage to the claims area faster, they were able to reduce the time for baggage to arrive at the carousel down to like just under eight minutes, which was acceptable according to kind of industry best practices, industry standards.
And so what happened?
Nothing.
People kept complaining.
Complaints didn't decrease at all.
And so the airport did some additional investigation.
And what they found was that people were spending about one minute walking on average from the gate to their baggage carousel and about seven minutes waiting.
And so they could have thrown more baggage handlers at the problem.
And maybe that would have, you know, kind of cut that wait time a little bit more.
But they did something different.
Instead they moved those two areas further apart.
They moved the baggage carousels to the end of the baggage claim area.
And they moved the gate farther away from the baggage claim area.
They reduced the amount of time that people were waiting at the baggage claim area to one minute and increased the walking time to seven minutes.
And guess what?
No more complaints.
People were happier because they were walking and it was no longer passive waiting.
This is a very interesting study.
It gets discussed a lot.
There's actually, if you're interested in this topic, and it's super fascinating, you can look up queuing theory.
And there's a lot of really interesting queuing theory.
This is a whole area of academic study.
You can look up a lot more interesting stories as well.
Another thing that comes up when you read about queuing theory is why do elevators have mirrors?
Why does the waiting area where elevators are often have mirrors?
This goes way back to when elevators were really slow, had to be operated by humans, and people had long wait times.
And so to give people something to do, they put mirrors there.
And sure enough, having mirrors in elevator wait areas is something that still exists today in most areas.
And it cuts down on complaints.
It's kind of this convention that we've gotten used to.
So how we perceive time is really interesting.
Normally we perceive, when we have to wait, the wait time that we perceive on average is about 15% slower than it actually is.
When we're asked to recall that later on, we actually remember it as being 35% slower than it actually is.
Like we hate to wait.
Human beings hate to wait.
That's just when we're passively waiting.
When we're enduring something where pain is involved, it gets much worse.
And it affects our perceptions even more deeply.
So there's a show of hands, who here, and you don't have to, like, hopefully this isn't from personal experience, who here knows what a colonoscopy is?
Hopefully not from personal experience.
Okay, so I'm not going to go into the details.
It's a very invasive and painful process.
I've not experienced it, just to be clear.
But by all reports, very invasive, very painful.
And so there's a study into the colonoscopy effect, where patients were, you have to be conscious during a colonoscopy, and they were given a device that allowed them to record when they were feeling more pain.
And less pain.
Tapping less.
And so this is comparing two different patients, patient A and patient B. And if you were to guess, looking at these two charts, which patient reported having the, oh, so sorry, this is for people in the back who might not see this.
This axis represents the intensity of pain, and this axis represents the amount of time that we were tracking, that people, sorry, not we, that they were tracking people.
So who here thinks that patient A had the more horrible experience?
Pain peaks are about the same, but fewer of them.
Who thinks that patient B had a worse experience?
Okay, so it will come as a surprise, perhaps, to many of you, it certainly would to me, if I weren't hearing about this for the millionth time, that patient A actually believed that their experience was more intense, and that it was of the same duration as patient B's.
And the reason for that is because, not because of the duration of the pain spikes, not because of the number of the pain spikes, but because of the intensity of the final pain spike before the experience ended.
So how an experience ends actually hugely colors our perception of how it was overall, and also how long it took.
Some other really interesting things, and I could go on all day, I'm not gonna, I'm gonna move on eventually, but when the span between our heartbeats is longer, time feels slower.
Generally, our heartbeats kind of go, do-do, do-do, do-do, and it's about the same.
But basically, it's measurable that when people report perceived time, and this is in milliseconds, so I'm not kind of saying like, oh, go and work on your pronatomic breathing and really kind of slow your heartbeat down, and you know, time's gonna feel like it lasts forever.
That would be great.
It's not true.
But it's just interesting that this is a measurable effect, that our physiology and how our bodies are performing actually affects how we perceive time.
Another interesting study around just sex and time perception.
Women tend to underestimate perspective time estimations compared to men, and the suggestion here, and this is theoretical, is that women tend to perceive time as passing a little bit more slowly.
And to be clear, please don't take any of these studies away and weaponize them against your female friends or partners.
Just say like, oh, well, you just think time's moving more slowly, and that's why you're always late.
This is definitely not why I'm on this stage.
So this is a red screen.
I don't know how, it's kind of red for you.
It's very red on my screen.
So if you were a man who was looking at that red screen and it was redder, and you're looking at it for 10 seconds or a minute, measurably, you might have experienced that period of time as being up to 40% longer.
So for some reason, time seems to slow down for men when they're looking at the color red, and there's all kinds of interesting theories about this.
So just really interesting stuff.
Again, this is our physiology.
These are things we can't change, and we can't control.
Have you ever thought time is speeding up as you get older?
You know, like a single summer day when you were a kid, you can recall it in perfect detail, you know, like from dawn till dusk.
It kind of is, sorry.
Or at least our perception of time.
And remember I said earlier that perception is basically reality.
It's not because, as some people speculate, there's some Einsteinian sort of relativity effect where because each year of your life is proportionately a smaller portion of your life, it feels like it's shorter?
No.
It's because over time the rate at which we process visual information slows down.
And so it feels like time is speeding up simply because we're processing less visual information, we're processing it differently.
On the flip side of that, what's interesting is that as you get older, users age 65 and up are actually much slower at using websites than younger users.
And so it's kind of this unfair double whammy that your time feels like it's going by faster and you're using things more slowly.
How time is measured, the units of time that we use are really an interesting, they have an interesting effect on us. 86,400 seconds. 1,440 minutes.
It sounds like a lot of time.
That's a day.
Just this is kind of getting back to what I was talking about before.
I live in a very small town.
It's in the, as people in other parts of the world would like to say, it's in the middle of nowhere.
Well, it's the capital of nowhere then.
It's really nice.
But it means that this green is mountains and forests.
This is where I live.
And this is the, all of these little names, they look like town names, they're a lie.
These are 12 houses, five houses.
The next town is way down here and it's a 35 minute drive away.
And when my kids were younger, I certainly know, I experienced this as a kid myself, a 35 minute drive feels like we're never going to get there.
This is, we live in the car now, we're car people.
So there was a lot of complaining.
We listened to a lot of music in the car, as you might have guessed.
And so one thing that we noticed was that the song, My Sharona, it's a great song, it's pretty much exactly five minutes long.
So we invented a unit of time called a Sharona.
It's a five minute span of time.
And we started measuring time during car rides or other kind of long events in Sharonas.
And believe it or not, this is very anecdotal, it'd be interesting to study this.
This really helped.
It especially helped my youngest because he was very, very impatient, busy little kid.
And so 35 minutes, it worked.
Seven Sharonas, I can do seven Sharonas.
I mean, we didn't actually listen to My Sharona seven times, but it's an experience you can think of, you can contextualize that and think I am capable of listening to My Sharona seven times.
This works for me.
So all of these things have an impact on how we perceive time.
Space, age, boredom, pain, heart rate, color.
So it's really cute that so many of us tend to think that we, just because we invented clocks, we invented the concept of clock time, that we have clocks in our brains.
And like, no, we're able to make clocks and we're able to look at clocks and read clocks, but we do not actually have clocks in our brain.
And that's great.
Our brains are so much more than clocks.
How we perceive time is so fascinating and interesting.
It's part of what makes us, you know, it's like the fingerprint of who we are.
So I don't want a clock in my brain.
Let's talk a little bit kind of back up and talk a little bit about how memory works, what's going on in our brains.
Because memory, as we kind of already alluded to, is a big part of how we perceive time.
So we take in information, we perceive new information constantly.
And within the first second, our brains have decided what they're going to keep.
This is sensory memory.
What they're going to keep that might be kind of useful for decisions that you have to make and what they're just going to flush.
And that transfer happens between your sensory memory and your working memory.
And then the memories that your brain decides to keep, those are your long-term memories.
And you don't really have a lot of control over what's happening.
This is all happening kind of on the fly.
We can't hold on to all of our memories, all of our sensory memories.
If we tried to take in all of our sensory memories and keep them in our working memory, you would walk in front of a bus.
It's just simply too much information.
And this is why we forget.
It's also our sensory memory, the information, specifically visual information that we take in happens in like 100 millisecond increments.
There are studies going back a couple hundred years into persistence of vision which found that yes, being able to rotate a wheel of fire, so long as you're doing that at 100 milliseconds per rotation, you're creating this perfect circle.
So this persistence of vision is kind of linked to this idea of 100 milliseconds.
And that's an average.
Some people actually do perceive time, just because of their physiology again, as they see more frames.
They have that gift.
Some people see less.
But these are averages.
You've heard of the doorway effect.
It's like most people have heard of this, this idea that you go up a flight of stairs, you go through a door and just the act of moving.
It's also called location effect.
Act of changing locations, flushes memory.
This has kind of been called under fire a little bit recently, where there have been some other studies saying, it doesn't always happen or it's not always replicatable.
But it's still, we've all experienced this phenomenon enough that it feels real enough that we can allude to it.
So this kind of ties into the idea of flow.
And we work in tech and design and software.
How many people here are familiar with the concept of flow, flow time?
So a fair number of people.
It was named and identified and studied by a psychologist named Mihaly Csikszentmihalyi.
And he basically described flow as a state in which people are so involved in an activity that nothing else seems to matter.
The experience is so enjoyable that people will continue to do it, even at great cost, for the sheer sake of doing it.
So this is somebody who's in a flow state, believe it or not, it's not always pretty.
You're doing something, the idea with flow is that you're doing something, and what I found actually interesting, I'll just sort of back step for a second, was so many of the speakers who spoke yesterday about the activities they do that make them passionate, Shirley, Rob, others, what they described were perfect flow state activities for them.
They're things that were challenging, but which were within the thresholds of what they were capable of, but they were pushing themselves to the edge of their capability.
They were absorbing, they were energizing.
So an ideal flow state activity, it's not something passive like, you know, reading a cheesy novel that you've read a whole bunch of times before, or playing a video game that's like pretty easy to play, and you're just kind of doing it.
Those are escapist activities, and they're not to be disparaged, they're good things to do, but flow state activities are something else.
They also require focus, they absorb you, you lose your sense of the passage of time when you're engaged in flow state activities, and they're extremely important for us to do, not just for our mental well-being, but even for our physical well-being, because the two are actually basically the same thing.
Why a flow state activity, the opposite of a flow state is an interrupted state.
It can take up to 23 minutes to regain focus after you've been interrupted when you've been doing something.
And so again, we live in a capitalist world, and so this research has sort of been, you know, sort of started around the areas of productivity, workplace productivity, how can we monetize our understanding of interruptions in the workplace.
But more important to me is the fact that being constantly interrupted when we're performing a task is actually really bad for our emotional health and our well-being.
People who are interrupted constantly during tasks, whose flow state is never allowed to be achieved, or they're not allowed to stay in it for a significant period of time, eventually become demoralized, they lose interest in the task, even if it was a task that originally absorbed them, so it's really important to eliminate or reduce interruptions as much as we can.
So how do we apply all this to how we use technology?
Well technology is everywhere.
We're always on devices and things like that, and that's not a judgment, that's just a function of the world that we live in and the types of jobs and rules that we've decided to take on.
But the thing about all of these gadgets and computers, phones, whatever editing software you might use or anything like that, is they're all glitchy, they're all laggy, none of them or very few of them give you the sense of you're able to work seamlessly with them.
So I love technology, I also find technology to be one of the most frustrating things in my life, and it's possible to hold both of those ideas in your head at the same time.
Our brains simply were not designed to accommodate the fits and starts of human-computer interaction.
Our brains were designed to do seamless things like, I don't know, kill a buffalo or milk a cow or plant a field, things like that.
We're good at those things, those things are easy activities for us.
Using computers is not.
And we can think that we can learn to accommodate these fits and starts, but it's always going to be work, it's always going to be a huge cognitive burden for us.
Believe it or not, the average web user believes they waste two days a year waiting for websites and apps to load.
Two days.
That comes out to about nine minutes per day.
I don't actually think I wait nine minutes per day, and I use the internet a lot, for, I don't think I wait nine minutes per day, but the fact that it feels like that, that that's a reported feeling that people have, is something that we shouldn't make light of even though it's probably not actually tied to reality.
Other things that we say, kind of going back to the 90s, when people first started getting interested in this idea of wait time or lag and how fast do we expect websites to be, back in 99, yeah, eight seconds.
In 99, eight seconds for a page to load was like, yeah, if you go back a little bit further, probably longer than that, assuming you're, I was just happy my modem stayed working for more than half an hour. 2006, we said four seconds.
And now we say about two seconds.
These numbers are interesting because they say kind of, they're a reflection of what we feel we should be accommodating.
They're not actually a reflection of what we actually need at a fundamental level.
Robert Miller, going way back to like 1960, oh, wait a second, I'm getting ahead of myself.
Robert Miller wrote this really great paper, and don't worry, if you're interested in any of this research, I've got citations at the end of my slides, so I'll be sharing those.
Basically said that a wait time longer than two seconds breaks your concentration and affects productivity.
And he said this in 1968.
So this is going back to these kind of early seminal studies.
And how this kind of affects technology today, well, Urs Hölzel at Google, Senior VP of Engineering, said basically, we want you to be able to flick from one page to another as quickly as you can flick a page on a book.
So we're aiming very, very high.
Something like 100 milliseconds. 100 milliseconds, huh?
It kind of goes back to this.
What Google wants, their stated goal is for every web interaction, every page to kind of just feel, have that seamless feeling.
Do websites achieve that?
That's another story.
So what do we know about how people perceive speed on the web?
Again, these are some early studies.
Jacob Nielsen from the Nielsen Norman Group has done this research a couple of times, and the data remains the same.
Actually looking at how people use the web, we need under 100 milliseconds to have this experience of instantaneous response.
Anything under 300 milliseconds is like we become aware of, like, there's a delay.
Anything around a second is like, okay, well, the machine's working, I can do this.
After a second, you're starting to actually think about other things.
That's how quickly our brains can start to shift focus.
Our brains want things to do.
So after a second, we're already thinking, should I go to another tab?
I have so many.
Maybe I could check one of them out.
So you're already starting to do that kind of bargaining with yourself.
And then after 10 seconds, you're basically starting to abandon the task.
And maybe you're sticking around with it if it's something you absolutely have to do.
You're really committed to the next page as part of whatever flow you're doing.
But this is basically, these are expectations.
And he did this study in 2000.
He revisited it in 2010.
The numbers are the same.
So when do we start to interact with the page?
This is a really interesting question.
Because you would think that we start to interact with the page the minute anything starts to appear on the page.
That's what I used to think.
Well, actually, no.
Most people or when I say most, I mean just over half people wait for almost all or all of the images on the page to render.
So this is the difference between what we might think as people who build websites and want people to use our sites.
As long as we're giving them something, within a couple of seconds, they'll be fine.
But actually, people are waiting for images to render.
And this is another Nielsen study, which looked at an 8-second delay in the key image right there, the hero image on the page.
And what they found was when the page rendered optimally and the image rendered optimally, the focus stayed on that hero image, which is arguably the most important content on the page.
Or else why would you put it there?
When there was an 8-second, am I on?
Okay, great.
When there was an 8-second delay in that hero image, it was really hard to get people's attention back up.
And this is all done with an eye tracking study, where we can actually see how the eye moves around on the page.
So knowing what the important content is on your pages and making sure you're delivering that first, because you're going to lose people's attention otherwise.
Web stress is a really great phrase that I wish I had coined.
I didn't.
I definitely did my part to help popularize it.
So Foveance did a study that was commissioned by CA Technologies, where they looked at introducing latency delays onto web pages and testing, kind of doing some A-B testing, where they had a control group that had optimized experiences and then a test group that had the slower experiences.
And they hooked EEG headsets up to them.
And they found that when the apps or the sites were slow, people had to concentrate.
Look at this on charts.
People had to concentrate up to 50% harder to stay on task with the tasks that they were asked to do.
So I was really excited when a few years later, in kind of the early days of mobile, when we were actually just trying to make the argument that mobile mattered.
I'm really dating myself.
Yeah, mobile.
We should really care about user experience on mobile devices.
We don't worry about that.
So I wanted people to worry about it a little bit.
And so I convinced the company that I was working for at the time, Radware, to commission a study that I was involved in to look at mobile web stress.
And we did the same thing as what the Foveons folks did.
We put some EEG headsets on people.
We didn't tell them that they were doing a speed and usability study.
We just said they were doing regular usability testing of some travel sites and some e-commerce sites.
And we gave some groups optimized experience.
We gave some group a slower experience.
And what we found was a few things.
Gratifyingly, the results were very, very clear.
And so I was able to go back to my company and say, like, phew, we didn't just spend tens of thousands of dollars on something that was completely pointless.
People experience slowness in the moment.
So just at a cognitive level, we found that frustration peaked of up to 26% on key pages like browsing pages and checkout pages.
And these were on John Lewis, which is an e-commerce site, EasyJet, and Ryanair.
And I feel like Ryanair, like, this is an older study.
Like, I don't know that I would choose Ryanair now, to be honest.
I would maybe choose someplace else.
But interestingly, they actually had some of the highest frustration spikes for people.
And that might have been a little bit to do with some preconceived ideas that people had about the brand.
We also found, and this was kind of interesting, we did exit interviews, as you do with usability subjects.
And we just asked them, like, what were your impressions of the sites?
And we did a bunch of different things with that data.
But one of the things that we did was we extracted all of the descriptive words that they used, and we put them in word cloud generators.
And what we found was really interesting.
For the fast people who experienced the fast version of the site, the word cloud, A, didn't have as many words.
People were pretty happy, and they didn't talk.
They didn't want to talk to us as much.
I mean, it's not like they were trying to get away.
They just had less to say.
And they found the site easy to buy from, easy to use, user-friendly, good.
A few people said slow, so I thought that was interesting still.
Better than others, quick.
And there was a few negative words, like confusing or busy, but whatever.
They sort of sneak in there.
But the people who experienced the slower pages, and these weren't radically slower pages.
They were, you know, we introduced, like, small latency delays.
But they said slow, boring, clunky, confusing, frustrating, inelegant.
So a lot of words, and what's interesting about these words, if we want to kind of group them into categories, content, boring, visual design, tacky and confusing, usability, frustrating and hard to navigate.
None of these have anything to do with speed.
The perception of slowness on the site affected people's perception of non-speed-related elements on the pages.
And I thought that was really, really interesting.
Slowness also affects long-term behavior.
So in another study where we looked at real user behavior, we did a similar kind of throttling where there was, like, 500 milliseconds of throttling for some users, 1,000 milliseconds for others, and then our baseline kind of optimized experience.
And what you can see, this is measuring return rate, or, like, user retention.
And what we found was that return rates generally were lower for the slower sites during the time of the experiment, so up to about 12 weeks.
But then we tracked people for six more weeks.
And what we found was that even when we had returned all of these cohorts to the same optimized experience, the people who had experienced the slower versions of the site, the same site, took longer to return to their previous levels of visiting.
And so there was a long-term impact.
So what we think we want doesn't always make us happy.
And this is an image study that I did that involved facial action coding.
So baseline images versus progressive images.
People get really, really passionate about this subject.
And I want to point out, this is just one study.
You're welcome to go ahead and do your own.
How many people here prefer to use baseline images in their projects?
Baseline JPEGs?
How many people prefer progressive JPEGs?
Progressive JPEGs are kind of the norm.
They're the rule.
Well, I was working for, still working for Radware.
They actually were really dynamite in supporting a lot of the research that I did.
And we're actually working on a brand new, highly optimized image format.
But that doesn't come into this study.
So I'm not going to talk about it.
But for this study, we had real users using different versions of site pages that we mocked up with different image versions in them.
And we didn't tell them, you know, that they were, that you're the baseline group and you're the progressive JPEG group and you're the experimental image format group.
We just had them use the different pages.
And we tracked how they seemed to respond via facial action coding.
So we didn't really trust people to use, to give us just words.
We wanted to actually see how their faces responded to these different image formats.
What was really interesting, I actually went into the study thinking that the order of preference for people was going to be the highly optimized experimental image format that we had created, followed by the progressive JPEGs, followed by baseline.
That's what I thought.
But in actuality, when, oh, and we also, people have to heart rate monitors.
So we could, like, tracking their heart rate alongside the facial expressions.
And what we found when we looked at the data was, well, actually, the experimental format was the preferred format.
But baseline seemed to make people happier.
Progressive JPEGs seemed to make people less happy.
This was very surprising.
This actually became very controversial whenever I talked about it at conferences back when we did this research.
We were so surprised that we actually got an outside third party consultant to look at our data, look at our research, and kind of give us some theories about why this could happen.
And so our consultant was Dr. David Lewis.
He's the chair of MindLab International.
And he said that basically image rendition is a two-stage process for progressive JPEGs, or a multi-stage process.
So first you get this initially coarse image, and then layers of image.
I'm explaining to you what you already know.
But cognitive fluency is inhibited.
Your brain is basically, each time a new layer of the image comes in, you're having to do the work once again of constructing that visual image.
And it just basically, your brain uses glucose.
And your brain likes to conserve glucose.
This is why we get things like decision fatigue at the end of a busy day, and we can't figure out what to have for dinner, so we just eat popcorn because our brains are just exhausted.
Basically as it was explained to us by Dr. Lewis, your brain's just using a lot of extra glucose to do this.
And so it's registering as kind of unpleasant.
Is it horrible?
No.
Am I telling you that you should switch over to baseline JPEGs?
No.
I'm just saying that there are other reasons why we might want to use progressive images.
Maybe it's because it's giving people something stable and placeholding in the browser, so they kind of feel like, okay, something real is happening.
Maybe it's giving them the ghost of an image, so it's like, okay, that's the product image I'm looking for.
I can wait.
There are other reasons why progressive images can be helpful, but at a cognitive level, they're not.
They make us a little bit unhappy.
Our brain's basically like a big easy button.
So slowness, as I think it's come across at this point, it's not a thought.
We can't have cognition about slowness.
We can't control slowness.
It's a feeling.
And we don't control our feelings.
An interesting survey by Tea Leaf and Harris Interactive basically was about phone rage, how people react to slow mobile sites.
And I thought this was really interesting, that like 62% of people behave more or less normally.
That's this large blue cohort.
Thank goodness.
But then amongst everybody else, it was like 23% curse at their phone, 11% scream, and 4% throw their phone.
This is a stock image.
I don't actually know anybody who would do this.
It's like in an office environment.
And I love that the woman in the background is just like, Carl's throwing his phone again.
Whatever.
I guess it's Tuesday.
But this is a measurable effect.
This idea that we can actually measure when people are unhappy and they're interacting with pages in an unhappy way.
And these are called rage clicks.
And so it's just a series of clicks.
People are hitting their mouse buttons in frustration.
And this is measurable.
There's a formula even that you can use.
I certainly didn't come up with.
But if you want to learn more about it, again, you can look it up.
It's a friend and former colleague, Nick Jansma, kind of pioneered this bit of research in this study.
It's really interesting.
And so Steve Souders, I don't know how many people in this room are familiar with Steve's work.
He's basically the godfather of web performance.
He wrote the original book about performance called High Performance Websites and the sequel, even more high performance websites.
And I'm lucky that I get to work with Steve at Speedcurve and have for the past six years.
But he wrote this really great thing that I love.
I love this quote.
The real thing we are after is to create a user experience that people love and they feel is fast.
And so we might be front end engineers.
We might be devs.
We might be ops.
We might be designers.
We might be writers.
What we really are is perception brokers.
But how do you measure perception at scale?
Oh, my time's up, but that's okay.
I'm getting to my conclusion.
Some of the biggest measuring mistakes that I see people make over and over again is not well, first, the biggest one is not measuring at all.
You need to track even periodically how fast your pages are.
You can't fix it if you can't if you're not measuring it.
Assuming that your own experience is universal.
It's like, well, it's fast on my desktop.
It's fast on my phone.
Your experience is not universal.
Your experience is interesting, but you are one data point among millions.
Not monitoring continuously.
Things can change really suddenly.
You can have server issues.
You can have issues with your third party suddenly having server issues.
You can have an issue because somebody has suddenly uploaded a really huge image or animated GIF to your CMS and suddenly a page's performance is just exploding.
So things can change suddenly.
Not monitoring real users.
Synthetic measurement tools are really great.
And if you're interested in learning more about synthetic tools, please come up and talk to me.
I'm happy to talk about them.
There are great free ones out there.
If you're just getting started, there are paid tools.
The basic synthetic tools are great because they let you actually analyze individual pages.
But what they don't do is tell you how your real users are behaving.
They just give you little snapshots.
And you need to monitor your real users as well.
Or not focusing on the right metrics.
Things like load time are archaic.
They don't mean anything anymore because there's so much stuff that happens behind the scenes that you can have a load time of 30 seconds because of the things that are happening behind the screen on your page.
But your key content is still rendered quickly.
Core Web Vitals, which some of you might be aware of.
They're kind of a Google trifecta of metrics are good.
But they really only measure a few things.
And they only measure them in Chromium browsers.
We're just looking at averages.
You need to understand what users are experiencing at the 75th percentile, the 95th percentile so that you really understand the whole range of user experiences.
And then kind of just to wrap up, aside from measuring, you need to think about optimizing.
What are you doing to make this experience optimal?
So I took a few principles from Don Norman, who's one of the principals and emeritus at Nielsen Norman Group.
Also wrote some books you might be familiar with, The Design of Everyday Things and a lot of other design books.
And he's very interested.
He's a very human-centric design forward thinker.
So these are some principles that he has come up with that I've kind of applied to web performance and user experience.
Eliminate confusion wherever possible.
Of course.
If people know that they're on the right path, they're heading in the right direction, they're going to be more willing to wait and they're going to feel better about their wait times.
Make the wait appropriate to the result.
If I'm on a travel site and I'm looking for good flights or vacation options and the search results page takes 30 seconds to render, I'm okay with that.
I feel like, okay, there's a huge database and you're getting me the best deals.
I'm fine with that.
That wait feels appropriate to what I'm getting.
If I'm on a restaurant site and I've taken ages just to kind of find out where they've hidden the menu or their address, because why would anybody need either of those things, and then I finally find that link and I click on it and then I have to wait for a PDF to download because that seems to be what so many restaurant sites seem to do, that's not appropriate to the results.
When people get to the page that they want, try to meet or not just meet but exceed their expectations and then end really strong.
As we talked about with the colonoscopy effect, a strong finish can make up for problems at the beginning of an experience.
So you want people to leave your page, leave your site feeling really good about what they've experienced.
And yes, if you do all these things, you will probably improve some measurable business KPI for your business.
That's great.
But you'll have done something that's kind of more important than that, in my opinion.
You'll create an experience that's not just tolerable, it's delightful.
You know, why shouldn't web experiences be delightful?
And ultimately, coming back to this idea that time is life, time is all we have.
Treat yours as the precious thing that it is.
Remember that for other people, time is a precious thing and just respect that with them and for them and take that into all the things that you do, I hope.
So thank you very much.
This is where you can find me online.
You can also find me, I'll be here for the rest of the day, I'll be around on the breaks.
And here is my list of citations.
I hope you can read them.
But I will be posting my slides online if you want to find them.
Thank you.
Thank you.
Thanks.
Bye.
