How to Build A Perfect Website

Wow, what a treat to be back at Turing. Amazing. Okay, we're going to go fast because we've got 100 or so slides to get through in 28 and a half minutes. Um, let's talk about how to make a perfect website. It's an appealing idea, isn't it? We certainly all spend a lot of time as individuals, as marketers, as developers and engineers and founders, being frustrated with websites which aren't perfect. We have entire industries dedicated to optimizing these things. Hands up anyone who has finished optimizing. Hands up anyone who has a perfect website. Interesting. But what would that even look like? Let's take a minute to conceptualize that. How would you go about producing that? How would you go about maintaining that? I think I have an answer. I think I know how to build a perfect website. And that feels like a big deal because our websites are usually one of the main ways that we reach, engage, convert, and retain our audiences. That's why many of us are here, whether they're B2B, B2C e-commerce, whether they're lead gen, whether they're brochureware, whether they're just blogs, how well our websites perform can make or break our businesses. Now, performance is an interesting word. It's not one that we've historically heard a lot of in the context of websites. Sure, we talk about the performance of our campaigns and our channels and our visits and our conversion rates, but we don't talk about performance of the thing itself. But what would that mean without context? Well, I like both of these definitions because they fit together so nicely. Performance is the capabilities of the machine and how well that machine works. And it may come as no surprise to you that many of the websites that we build and use are not particularly performant. Now, performance is an increasingly important part of online conversion and revenue generation, etc., etc., etc., but it's a broad, vague, and nuanced concept. I want to get much more specific and look at what we mean. I wanna start with speed. One of the most tangible parts of performance, one of the most easily measurable and most impactful parts of performance is speed. And there's a mountain of research which shows this to be true. In 2014, Wired Magazine found that almost half of people expect a site to load in 2 seconds. In 2016, Google found that delays of over 3 seconds lead to 50% abandonment. In 2020, it goes on, a 0.1% improvement in speed can generate up to 10% impact on conversion rate. There are hundreds and hundreds of these studies which paint a really explicit painting of how important speed is to your users. It's because you're competing for their subconscious attention. A half-second delay might mean me glancing at my phone, or looking out the window, or changing tabs, or feeling doubts about my purchase intent. You're not just competing against comparable businesses, you're competing against Netflix. You're competing against me feeling a bit hungry. You're competing against me wondering what the notifications are on my phone in the moment— at the moment that I've forgotten to turn off. Every millisecond that you leave me waiting for your website to finish loading means that these other needs risk gaining more focus. So given all of that, you would think that everybody would be on top of this, that we would all be optimizing the hell out of our websites and that we'd be running straight for the bank, laughing all the way. Let's look at some data and see what's really happening. This is part of a study from Pingdom in 2018, who are a performance monitoring platform who analyzed the top 100 e-commerce sites in the world based on Alexa's dataset. And they found and categorized their performance into broad grades. These are an amalgamation of a whole bunch of metrics and objective user experience evaluations and so on and so forth. Most of the sites score a C or a D. That means that they are relatively slow. They are not performant. Now, if this was your child's school report, a C or a D might be a bit disappointing. For businesses who are competing to grow or to win, this is either gross incompetence or criminal negligence. These sites have direct competitors, and those milliseconds are lost ground in a war of attrition on attention. This is shocking. The study goes on to break down the best, best and worst performing sites. Here we can see the lowest 10, the slowest 10 in their set. You will recognize some of these brands like Under Armour, Gap, Best Buy, and so forth. Okay, pause. This is from 2018, so it's a little bit outdated, but that's fine. We can test these sites now and see how they're doing and maybe see some change over time. Let's start by looking at Gap, a brand which has been trading for over 50 years with 3,500 stores globally and has had a monumental impact on popular fashion and culture. These guys should have their act together. They score an E. Ouch. In fact, when I pretend to be a mobile phone in my Chrome browser on a poor 3G connection, it takes 13 seconds for the product banner to load on the homepage. Users expect a site to load in 2 seconds. 50% of users abandon your site in 3 seconds. That is not good. They are harming their engagement. They're harming their reach. They're harming their growth. They're harming their SEO. They're harming their conversion rates. They're harming their revenue. Interesting, right? Let's look at Best Buy. Best Buy is the biggest retail brand in the US by an order of magnitude, $47 billion in USD revenue in 2021, countless awards for their operational efficiency. They score an F. This tool does not go lower than an F. Interestingly, their structure isn't too bad, but their server is so slow, it takes 30 seconds for their homepage to load on my mobile phone. Ouch. Maybe these are just coincidences. One more to make sure. Let's go through a few more interesting ones. Everybody's favorite US sportswear brand, Under Armour, who their whole brand is about being physically performant. You would hope they would think the same way about their website, except they score a D. Interestingly, these guys have done some clever stuff. It's not too bad, but the sheer weight of their analytics kneecaps their sites. They're measuring 1,000 data points on everyone who loads their pages, but it takes so long for those pixels to fire that those people have all left by the time they try to measure them. That's not performance, that's self-sabotage. These are all bad. They have been bad since 2018. Many of them have been bad since 2012 when Pingdom started doing this research. Okay, I will accept that if these sites have been bad for a decade, they're probably not representative, and we should look at some other examples of the broader ecosystem. We have a whole generation of disruptive startups who are reinventing industries and changing the world. You would think that they would do much better. Packed Coffee, disrupting coffee, C. Impossible Foods, disrupting meats, D. Purple, disrupting mattresses, F. You get the idea. There is a trend here. Fair enough, maybe the disruptive startups are too busy disrupting to be optimizing and performing. Maybe they've confused move fast and break things with have a broken website. It's not ideal. Maybe the big players and the established giants are doing better. Coca-Cola, who have a global monopoly on hydration. Forbes, one of the most influential publications in the world. Dell, one of the most influential technology corporations. You see where this is going. Nvidia leading the AI and machine learning revolution. Vodafone, one of the most significant telco companies in the world. Vodafone, who published a case study on Google's very own web.dev website earlier this year showing an A/B test they ran to speed up their site generated a fortune in cash. They still score an F. Most websites are bad. This is not the case of a few bad apples. And fair enough, I've cherry-picked some embarrassing examples to make a point, but this is broadly true across the web. Some sites do score well. Disney, PayPal, Netflix, Deliveroo have invested time and resources in making their sites fast. But they're still not perfect. There are still gaps and opportunities when you look at their accessibility or their usability or how they handle internationalization or privacy or any other number of areas. Most websites are bad. And it's always the same types of things. Their images and video are huge and not lazy loaded and not optimized properly, or their servers and hosting are outdated or not configured properly, or their JavaScript and CSS is a bloated nightmare and they're not doing basic optimization, or they're just getting general standards wrong. They're missing meta tags, they're confusing their caching directives, they don't have their HTTP headers set up properly, and 100 other things. It's the same things over and over and over again. And okay, it's easy to understand why, because this is arcane. This is backend dev, frontend dev, databases, developers, servers, PHP, pixels, people, and processes— most complicated of all. Solving this requires cross-functional magic, often against a backdrop of technical debt and a lack of governance and processes. So it's hard for people to quantify and manage and ask questions like, how well are we doing? And what is the impact of that? And what's our direction of travel? And how do we benchmark against competitors? And if we invest X, what does that return. So understandably, many organizations haven't invested enough in performance. There is a solution. This is the good news. We have tools, metrics, and framing to allow us to prioritize, quantify, and impact. In May 2020, Google launched a Core Web Vitals initiative, which is a set of metrics designed to definitively and universally measure how fast a web page is. 3 metrics: Largest Contentful Paint, First, input delay and cumulative layout shift. I will not bore you with the details of how these work right now, but the basic premise is collectively, these are a good measure of if a page is fast. Dwell for a moment instead on the naming convention. Core Web Vitals. Not important things you should think about, but critical fundamental requirements. I think that understanding and excelling on these metrics is gonna be absolutely critical if you want to thrive and compete in the coming years. And handily, there are all sorts of tools to measure these. There's all sorts of different places you can go and get this data, whether that's in your browser, in your development pipeline, in tools you want to test from your desktop and your phone. There is readily available access to these metrics in a myriad of different formats. There's links at the bottom that you can go through later. Also, in addition to the lab data that you get at the point of testing, these tools collect field data over time. Any users visiting your website running a Chromium browser have their experience pooled into a big global universal database which Google collates for all the websites that those users visit. So if your website gets enough traffic, You can go and get real user data and analyze the distribution of experiences. You can use that to identify problems and opportunities and so on. All of that real user data gets compiled into a big database which you can go and look at in dashboards online in Google's Data Studio tool. Your site is in there, your competitors' sites are in there, and you can view those metrics over time. There is potentially gold to be mined in analyzing and understanding how these metrics are changing for your competitors. In short, Core Web Vitals solves the this is hard to quantify problem. Now we have explicit, universal, and comparable metrics against competitors which are tied back to user experience and by extension to revenue. And there's also motivation to improve these scores. Alongside all the research which just says if you make things faster, you make more money, Google also explicitly rewards speed. It's an organic ranking factor and influences paid search, obviously, as well as all of your other channels. And you get compounding impacts from ranking well here. Higher visibility means more traffic, which means cheaper ads, less paid investment, more growth, more investment back into performance, more money. Everybody wins. The good news is you don't just get scored and then left out to dry. Google provides a whole bunch of recommendations and metrics and infrastructure and support to help you do this. There's an incredibly rich ecosystem of documentation like this. Um, here is an example of a Lighthouse test giving me Core Web Vitals metrics in my Chrome browser, and then literally a blog post with Google moving the mouse saying, click here to see more information. This specific image is problematic. Here is why. Here is what you should do about it, which links you to a whole bunch of examples which says maybe just copy paste this HTML instead of whatever nonsense your developer did. And if you want to understand the reasoning behind that and why and get a bit of an education into some of the thinking that produces that code, here's the general approach and the underlying principles. And of course, it's not just as simple as saying put this code on a page. You also need all the workflows and processes. They solve that as well. There's a huge amount of documentation which says, okay, here's how you develop continual improvement processes, here's how you integrate your workflows, here's how you do performance budgeting. All of this is a solved problem. Performance is a solved problem. Now, you can maybe argue that one of the reasons that so many sites are still slow is that this is a little bit new and it's a little bit challenging, but all of these problems have been solved for quite some time. The tools and the documentation and various flavors of these kinds of testing processes have been around for years, yet most sites aren't performant. So there's something else going on. There's an underlying problem here that we need to understand that prevents organizations from making their websites not awful, aside from not having a well-lit path to doing so. And I think the answer is one of two things. Either various people throughout the organizations don't know what they don't know, which is forgivable, or even knowing about the importance of this and the availability of clearly well-lit paths, They can't or won't or haven't invested the resources and time and priority to make this happen, which means that as you step back and look at some of the brands and websites which are scoring poorly and where they're evidently struggling, you start to realize something quite profound. You realize that Core Web Vitals don't just measure speed. And this is the key thought I want you to take away today. Indirectly, these metrics also measure a whole bunch of other things, some really interesting things. Like the technical maturity of the team running the site, like their resource and resourcing and budget allocation, like their core competency and capabilities. Maybe there's evidence that they don't have enough dev resources to make their sites fast. That tells you a lot about their strategy and their investment that you might be able to capitalize on. It tells you a ton about their priorities and politics. It's easy to spot when they're prioritizing the wrong things and the wrong departments and the wrong teams, or when the internal SEO team or agency doesn't have enough leverage or executive buy-in. There's a huge amount that you can infer from these metrics, and over time you can see how all of this changes. This is profoundly impactful competitor insight that you can use to steer your strategies. Why on earth is any of this important? Let me zoom out for a moment. And many of the things that businesses compete on— we're on a trajectory where the thing that they make or do or sell is increasingly commodified, and to some degree we're all in a race to the bottom. That means that we're competing on 2 things fundamentally. One is discovery, and i.e., the cost of user acquisition, and the other is retention, i.e., brand preference, flywheels, etc. Those 2 areas. And your competitive differentiator across those 2 things is essentially how good are they versus competitor offerings from the perspective of your users. How well are they differentiated? How well are they presented, etc.? And that's not just speed, but more broadly, it's the overarching user experience. And broader still, we can tie all this neatly into one word: to describe this quality of experience. It's quality. You're competing on quality. And Core Web Vitals metrics therefore become quite a good user— quite a good proxy for understanding the quality of a website, at least the user-centricity of it and how well the organization is committed to being user-centric. And yes, the quality of a website doesn't necessarily correlate perfectly with the quality of the product or the service or the thing that they offer, but I damn well bet it correlates closely. You could even maybe argue that Core Web Vitals scores might be a good indicator for whether a brand deserves to be featured in a consideration set. And that gives us a really neat way to frame all of these challenges into a simple concept, and that is that the quality of your website determines how well it performs. I'll give you— this does not sound profound, but this isn't the way that we've historically thought about websites at all, because we haven't had the metrics, the tools, the processes, or the mental models. And even with those things that we now have, there's still a problem because nobody in your organization is likely responsible for quality. You probably don't have a quality department. And even now that we have the measure— have the metrics and the ways of measuring speed and other areas, quality is still quite challenging and amorphous. But that's the next step because this ecosystem is still changing and evolving. Web Vitals isn't just about speed. There is a broader ecosystem. There are new metrics on their way. Um, in the next few months, we will have a responsiveness metric, which will be the 4th, which we'll add to the existing set, which will measure how well your website responds as you scroll and interact and touch it, not just at the point of loading but after load. How responsive is your website? Probably not very, because we've not had the tools or the motivation to make it so. But soon that won't be the case. Also next year, we get smoothness. Who is optimizing their website for how smooth it is. But there's money on the table there. More importantly, these metrics don't exist in isolation. We're not just measuring speed. Web Vitals also measure SEO, accessibility, general best practices, and other areas. You can see here equivalent measurements and metrics and standardized ways of performing and scoring accessibility, best practices, and other areas, which you can do not only on your site but also on your competitors. And these things all sit in turn alongside areas like responsiveness, like security, like safety. So they're also measuring how well you cater to the millions of users who rely on assistive technology or who are on old mobile phones, etc., etc. And those metrics themselves also evolve and change over time. They're continually iterated and refined and altered and improved based on changing standards and technology and user expectations, which means even if you score well today, you might do nothing at all and score badly tomorrow. It's a bit scary. The brands which get this right and invest in improving their quality now will leapfrog ahead. They will do better in discovery because Google will reward, reward them. Their flywheels will turn faster. Their costs and revenues per customer and acquisition will improve. They will unlock more resources to improve quality, etc. Their velocity will increase compared to their competitors. And that's really interesting when you consider that you can also measure all of this not only for you, but also against the people you're up against. So if you're scoring poorly on these metrics, which I would implore you to test, that's probably a strong lagging indicator that you have deeper organizational resourcing or prioritization challenges. And those are going to hurt you deeply and increasingly deeply over the coming years. Speed can be an enormous competitive advantage, not just inherently, but also because improving your speed forces you to also improve your content, your media, your marketing, your development, your people, your processes, your politics. All of those problems need solving in order to make a fast website. Speed correlates incredibly well with quality. And quality determines how well you perform. It's no coincidence that the word quality shows up over 400 times in the guidelines that Google provides to its employees to evaluate how good a website is. Google is evaluating the quality of our websites and demanding increasingly high standards, and it is madness that we are not doing the same thing. So, what do we do? They say that you shouldn't aim for perfection. They say that done is better than perfect, that you should strive for progress rather than perfection. I disagree. I think that those kinds of sentiments are born out of industries that don't have ways to measure and define perfection, and they are the kinds of sentiments that cripple businesses and continue to mean that they are slow. We have the tools, the metrics, the performance, the processes, the incentives. We have the toolkit. To be in second place is to be the first to lose. In a winner-takes-all ecosystem where algorithms decide which brands and websites we get served without a process whereby the user chooses in the middle. To be second is to not be present at all. You have to build a perfect website. And that is possible. In fact, it's easy. You don't believe me? This is a perfect website. I will buy a drink for anyone who can find anything objectively imperfect about it. Tweet me. Okay, I'm cheating. This is my personal site, and whilst it has a big hero image and content and analytics and so on, it's probably not representative of many of your websites. And I'm the only person working on it, so stuff doesn't break in the background and so on, so it doesn't compare well to your experiences and challenges. However, this one probably does. Yoast.com is complex. We have e-commerce, we have shopping cart cookies, we have internationalization, multiple currencies, fast-moving content, caching, security, teams of developers with different priorities and processes. You get the idea. This is a very complex ecosystem with lots of moving parts, and it doesn't score quite as well as mine, but it's close. And yes, we've observed improvements to bounce rates, to conversion rates, to revenue, etc., from having gone on a journey to improve this. Achieving high performance is possible if you have a full team of performance experts, or if you start from scratch, or if you have unlimited resources and timeframes, and if you can perfectly predict your future needs, and if nothing ever changes and no unexpected needs ever occur, and your competitors aren't making changes faster than you are. Ha. In the real world, making a perfect site is tricky. Even if you know what perfect looks like in your ecosystem, it turns out that unless you have an incredible amount of resources and an incredibly sophisticated team with far too much free time, you can't keep up. You can't plan well enough. You can't sidestep the accrual of technical debt. Even though it's paint-by-numbers, building good things, never mind perfect things freaking hard. You can't make a perfect website. You can't master enough or influence enough or monitor enough or fix enough to ever get beyond 80%. And even if you somehow get somewhere close to good, you can't maintain that because things change faster than you can keep up. But you shouldn't try to. In fact, I want to convince you that trying to optimize your website is probably a really bad way to invest your time, because good strategy means knowing which battles to fight and where to best spend finite resources. Trying to master performance optimization and security and accessibility and SEO and web standards and a dozen other disciplines that even I struggle to keep up with, and I spend a year working on all this sort of stuff, is not the hill that you or your business needs to die on. You should be focusing on product and service quality and reputation and brand and the effectiveness of your marketing storytelling. Your strategy shouldn't be hope. Your web development and engineering team shouldn't be trying to learn and master this sort of stuff either, because they don't know what they don't know. So many of our businesses just simply hope that our developers and experts— developers and engineers are expert enough to maintain and build and achieve fast sites, when all the evidence suggests that's not the case. Hope is not a good strategy. It doesn't mean it can't be done or that your website can't be perfect. It just means that you can't do it. Your organization can't, your developers can't, your editors and content people can't, but that's not the same as it can't be done. You just need some help. You do that by borrowing best practice foundations from others. You build on the work that other experts have done, you adapt it, you extend it to meet your own needs, and even still benefit as the original creators of that work continue to improve it. Imagine if your website automatically improved over time based on integrating the work of subject matter experts automatically, instead of a website which breaks all the freaking time. Wow. A website where your distance to perfect doesn't erode over time because it's always up to date with the latest standards. That's not madness. That's exactly what I'm proposing that you do. It's what I do with my website. It's what your competitors who eat your lunch are doing. The secret is to get other people to make your website perfect for you. I'm going to show you how you can leverage work done by other people in 6 minutes and 40 seconds. Um, the secret is that a perfect site has 4 foundational pillars. You can leverage the work of other people to fill those in, and then you don't have to worry about them at all. We use some analogies to visualize how they Together, you need an operating system, an app store, a publishing suite, and telco infrastructure. These are your 4 pillars. The combination of these are the secret to building and maintaining perfect websites. These— I'm going to give you the best-in-breed tool and solution for each of these sectors. These are all tied together in a really neat, nice way. They're specifically designed to do so. So you can choose your own solutions for these, but be aware that swapping out these specific pieces will weaken the whole structure. The first, you're going to love this, Ed Fry in the front, is WordPress. Pause. Some of you are undoubtedly on board the WordPress bandwagon and some of you are undoubtedly not. I want you to put aside all of your preconceptions for a moment of what WordPress is or isn't. It absolutely has its faults, its limitations, its problems and challenges as a content management ecosystem. It is not a perfect product. But as a universal architecture for best practice and for solving problems you didn't even know you had, like reliably loading an image without blowing up a user's mobile phone, It's unrivaled. Think of WordPress as a series of APIs that sit right at the foundation of your website, which mean you don't have to reinvent the wheel. Not as a publishing suite where you go to type blog posts. Yes, use it for that as well if you like, but as a series of APIs you use to build upon. It's a simplified interface on top of a ton of complexity which you shouldn't have to think about or build. And it's very trendy to say, but WordPress is not a good fit for my business because I have all these unique needs and tech stack. If you're a CTO with something to prove. And yes, headless websites are trendy, decoupled content management systems are trendy, Contentful is cool, Gatsby is cool, Jamstack is cool, and your tech people are convinced that they know best and can create awesome, incredible performance stuff. Yet all of the metrics and evidence we're seeing suggests that they can't. Think about that. I'm not bullying developers and engineers. I'm saying that we need to have realistic expectations when it comes to what you can achieve and maintain. And it's important to align those expectations with your strategy. Hoping that your developers can make your website fast through some fancy JavaScript stack is not a strategy. You need to get the basics right before you chase the shinies. You can have all the shinies on top of a WordPress architecture without needing to architect some kind of utopian technology vision which will never see the light of day and which will keep your dev teams tied up for the next 10 years unbuilding and rebuilding and changing. You can't allow your tech stack to constrain and cripple you and to increase your distance to perfect. Not if you want to win. The second ingredient is AMP. Some of you are undoubtedly AMP fans, some of you are not. Put aside all your preconceptions, please. Google did such a poor job of marketing and labeling and describing what AMP was that they've left a sour taste in everybody's mouth. And even as recently as this week, there's been a huge amount of controversy from the Texas state— not state— legal framework in Texas accusing Google and AMP of a whole bunch of nastiness. I would implore you to ignore that as well briefly. There's a huge amount of hot air and conflation and technical confusion in that. Hopefully, all that will go away. Ignore that for now. Go back to our scheduled program. AMP does a lot of things, but I'm most interested in 3 points. One is this is your library of website components. In the same way that Google and iOS's App Store allow you to go, I want my phone to be able to do X and Y, AMP allows you to say, I want a carousel, I want a banner, I want an image, I want a login form. All of those things are plug and play. And they're best practice, perfectly performant. You click, you drop them in, you don't have to build or maintain anything. As a universal component library, it is unrivaled in terms of similar things on the web. There are no similar things on the web. There is no framework like this for building and maintaining perfect websites. It's also a performance framework. It does that by setting constraints on the things you can and can't do and making sure the things that you do do are inherently lightning fast. The best way to make all of this happen without having to do anything more than click a few buttons is to install the AMP plugin for WordPress, which does some really cool stuff like this mess. This is the source code of my site. The bit I've highlighted is quite interesting. Vell module preload. I'm a JavaScript and CSS nerd. I have no idea what this does. Turns out it's some fancy precaching method that makes everything load faster. That's great. Turned up overnight automatically. What's really interesting is this plugin and the AMP standard is developed by a team of people who work at Google. Consider that. People from Google are working on my website for free to make it faster so that it ranks higher in Google. Huh. What's even cooler is this screenshot is out of date. A week later, this went away and was replaced by an entirely different cutting-edge standard. My website is improving over time because a whole bunch of Googlers are working on it for me whilst I sleep. Your website is getting worse over time. Very quickly, AMP does a whole bunch of other stuff that I won't get into because I'm running out of time, but the third component— yes, Yoast is the company that I work for. You are welcome to select your own SEO plugin of choice. But increasingly in the WordPress ecosystem, the SEO ecosystem with inside that is responsible for much more than just SEO. This is increasingly again a set of APIs and frameworks that allows you to communicate with Google, with Facebook, with third-party systems, to deploy schema, to integrate with things like the AMP plugin. Choose your own solution, but maybe look for one where the employees are good friends and collaborators with teams at Google and similar companies. Um, the reason this is important is because every time that your tech team spends more than 10 minutes doing any of this kind of stuff is time that could have been spent— spent time that could have spent doing other more important things. There is no need for anybody in 2021 to be developing XML sitemaps or worrying about canonical tags. These are solved problems. You just need an ecosystem that gives you the APIs and the REST infrastructure to produce that. The 4th tool in our pillar is Cloudflare, which I don't have time to talk about. But if you're not already using Cloudflare, then go and adopt this. This automatically makes your website more secure, faster, fancier, etc. One of the really cool things it does is essentially just sits in front of your website and speeds everything up automatically. And you do that again by clicking a button. You turn on the things and it automatically makes all of your images faster and your JavaScript faster. And then it allows you to automatically do a whole bunch of other stuff in WordPress as well, which is pretty cool. That is about it. That's slightly anticlimactic, but that's kind of the point because all of this is to solve problem. The only reason that brands aren't fast is because they don't know what they don't know, or they choose to be slow. Consider for a minute that even if you don't like WordPress and don't want to use AMP, The sites we looked at the beginning are your sites. These are like your sites, and the reason they are slow is a choice, either a conscious one or an unconscious one, that is out of ignorance or arrogance, which frankly, as a business and a strategy, is unforgivable. If they didn't know what they need to know, now they do. You have the tools, they exist. You cannot build and maintain a perfect website through hustle. You can only do it with the right tooling. You cannot go away and do this yourselves. You can't keep up with the speed. Please stop trying to make a perfect website. Stop accepting that it's okay for your site to be crappy. Get other people to make your website perfect for you. Use this stack. Run the audit tools on your own site. When you score poorly, be more concerned. Ask difficult questions. Make this painful. Go out and win. Thank you very much. Thank you.