 Hey, gang, sorry about the technical difficulties. I had a laptop failure last night as I was getting ready to leave, and had to switch over to an ancient old laptop, but it's amazing how you can make the old hardware still work. I'm Paul Thompson. I'm actually getting a real kick out of being at WordCamp Hamilton today, because I grew up in Hamilton. I've been 20 years out west, living in the mountains near Banff National Park, and I've just been back for less than a year. So this is like coming back to my old stomping grounds. This is great fun. I've been on the web since 1996. I've been doing website optimization since 2003. I actually lived through that fiasco that Joey Coleman was talking about this morning with Moveable Type. That was my introduction to CMSs and content management. My primary focus, I'm an SEO and a technical web marketer, so I don't build a lot of sites. I generally build a couple per year just to keep my hand in and for special projects, but my focus is building websites into powerful marketing platforms. So it's the optimization layers that go on top, and then making them usable and functional for their visitors. I've worked on everything from local small business websites up to a major technical audit for emarketer.com. So all of the things we're gonna be talking about today apply and will be useful to you, whether it's your own website or whether you're responsible for major client websites. It's gonna be as actionable as possible. I want this to be the kind of thing where you can go home tonight and actually do a couple of these things on your website. There are a couple of really tech-heavy slides. Don't worry about them. You'll get a link to the slides at the end. I've included them in there because there is some step-by-step stuff for configuring Google Analytics that you're gonna wanna do, and I don't wanna make you walk through the whole process as I describe it, but it'll be on a slide for you to look at and have resources for later, okay? So let's dive right in. If you talk to any SEO, any web marketing optimization specialist, even a lot of developers, every time you ask what should I do to improve my website, a good SEO or marketer will give you the same answer. And it drives you nuts, right? It's like, all I want is a straight answer for God's sake. And I'm not getting one. There is a really good reason. We're really not trying to be prex. The web has developed to the point now where each website has a different audience. It has a different purpose. It has a different set of needs and requirements. And to apply a standard set of needs and requirements to everybody's website means it's gonna fail on 90% of them. We have to do that personalized approach that says, let me fix the problem that my website has in a way that my website users will benefit from rather than trying to say, we'll just do it the way the average person wants it and we'll all be good. And that's a real challenge. And that's what we're gonna talk about today is how to dig out from the data, how to let the data tell us what it is about our particular website that's not performing well and then let us monitor when we make changes to see if we fix the problem. That's the challenge that we have in a lot of cases with, as was being talked about in the A-B testing session this morning, you can go to the web and look up the best practices for how banners should be configured or how a slider should be configured or how you shouldn't use a slider which is quite a little preference. But those are best practices. Those are, you know, everybody thinks such and such. Do your users think that? You better know because if you start making average person changes and you have a distinctive website as we all try to have, you're not serving your user, you're just guessing. Okay, so we're gonna make the data, give us the answers so we can stop making those kinds of guesses. So some examples of what we're gonna ask our data today. How do our different website visitors behave? And we're gonna get into that in a sec, but think of all of the different kinds of visitors and the different characteristics they're gonna have to your website. If you're really lucky, you're running an internet. One of those websites that's only for the people who work inside the company and you know that everybody uses it exactly the same way because they all work in the same company. Most of us don't have that luxury world we're working on a website that's country wide or international. How fast are our website pages? There's a real problem here that I'm gonna describe to you that we're gonna solve because right now I'm betting 80% of us don't even know or we're being this led into thinking there's performance that we don't have. Which pages are doing their job well and which ones aren't? Because if I'm gonna spend time, I mean, how many of us have experienced that saying I'm gonna take Sunday afternoon, it's raining, it's crappy day, don't have anything else to do, I'm gonna spend the afternoon working on my website, improving my website. And then you leaf through your WordPress dashboard and then you open up Google Analytics and you look at those first three reports and your eyes start to bleed and you close Google Analytics and then you go out for a coffee, right? We want to make sure that we're spending time on the things that are gonna have an impact rather than wasting time on things that might look pretty or might make a few changes but don't actually have impact on the end results. So we're gonna get our data to tell us where to put that focus. What new content are our visitors begging us to create? There's actually a tool inside Google Analytics and inside your own website that will do that for you. We're gonna cover that. Where and is our site navigation not working? That's really hard to know that unless you do a bunch of major broad scale tests where you put a bunch of users in front of your video or in front of your laptop and videotape them trying to navigate your website and it's called user testing and there's a whole process for that but it's expensive, it's time consuming and it's after the fact. So we're gonna figure out which parts of our navigation and which parts of our menus and our site structure aren't working. Which pages are really close to scoring us a whole lot of new traffic but they're not quite making the grade yet? Let's work on those. I don't wanna work on the stuff that's failing so badly that it's gonna take me six months to ever get it to rank. I wanna work on the stuff that I can work on on Sunday afternoon and by next Wednesday I'm starting to see some impact, okay? And the last one is just that aspect of opening Google Analytics and then your head exploding. Why can't such powerful software tell me what I need to know instead of me having to beg for it every time I wanna go and look something up? So we're gonna teach you analytics, how to do that. The first rule of analytics, do not use average data. It will lie to you. It will lie to you every time you look at it and in ways that you won't understand until it's too late and you've screwed something up. Okay, those two graphs have the same average. Imagine this is the temperature of the place you're planning to go for vacation. Okay, you look at the travel brochure, it's like, oh great, the average temperature is 20 degrees. Yeah, there's 20 degrees in somewhere in North Florida. There's an average of 20 degrees in Siberia where you either freeze your tiny heiny off or you bake and die. Which of those two places do you wanna go to? Well, without knowing how those graphs were arrived at I'd be really reluctant to spend thousands of dollars on airfare to pick. So we can't use average data. How many people use the bounce rate of their website as a reference point? Don't be shy, okay. We're gonna talk about how to get real value out of that number because by default, it's the average of your entire website and it tells you nothing, okay? So we're gonna overcome that. And the way we overcome it is we have to start segmenting the data. We have to look at the data as it applies to a known quantity or a known group of visitors. And we can make the segmenting by all kinds of variables for those. We can compare how mobile users experience the website compared to a desktop user. How somebody from paid traffic experiences our website differently from somebody who came from organic. A bounce visitor compared to a non-bounce visitor. What's unique about the guy who doesn't bounce that makes him stick around so we can do more of that to the guy who does bounce because we want him to start sticking around. Even geographic. You've got a web store that sells products in Hamilton and in Toronto and in Ottawa. Well, the behavior of a buyer from Ottawa is gonna be very different from somebody who lives in the town that the website is located in and has its own storefront in Hamilton. I wanna know those differences. Those are the kinds of segments we can ask analytics for. And it's built in. This doesn't take a whole lot of extra work. The first thing you're gonna need to do is choose the segment from the list. That's actually a button that you'll find clickable at the top of the screen. Sorry, this projector is very bright so you may have trouble seeing some of the less dark text. But all you do is click that button and this whole list of pre-built segments drops down for you. These are all built into Google Analytics. You just put a check mark on them and they will add that segment to your reports. And it's everything from sessions that had conversions, tablet and desktop traffic versus mobile. So remember, Google Analytics calls mobile and phone the same thing. Anything bigger than a phone, it calls desktop or tablet. So if you wanna compare phone to everything else, you have to do phone versus desktop and tablet. If you wanna see only desktop, there is a segment for that. Once you select one of those, it's going to give you two lines in each of your reports. One line for the average for the website and one line for the segment you just picked. If you really wanna make it work, don't use average website data compared to your segment. Compare two separate segments. Compare mobile to desktop. Don't compare mobile to average. Because average already includes mobile averaged into it. It's gonna be inaccurate information. But these are all built in thanks to Google. They've also done a really nice job of building a gallery of custom-made versions of segments that other people in the marketing world have said, this is a segment I find really useful and I find it so useful, I'm going to customize it and I'm going to make it available to share. And it's literally a button you click it'll open your Google Analytics account and say, do you want to have this segment in your account? You say yes, please. And all it's giving you is the structure. It doesn't give you anybody else's data. It'll just let you have that segment and then you can either use it as is or customize it. But there's this huge library of all of these neat ways of looking at data. So really take advantage of that. The key to remember is if you create too many segments you're gonna have data that, you know, a report that has three, four, five, six lines of data for every piece of information and it gets really hard to read. So when you're starting out, compare two segments. Compare mobile to desktop. If you want a really eye-opening comparison compare social to organic. Because you're going to be astounded how little value your social traffic is bringing you. Shouldn't say that after the last report. But it gives you a starting point. It lets you know where you have work to do. Some people, by converse, they've got this massive engagement from their social traffic. It's like, cool, go work on your SEO. So you had a quick question? This is the main screen when you go to Google Analytics. Yep, this is the screen if you're in any report. Right at the top of any screen you're going to see like normally all you see is all users. And most people think that's all Analytics can show you. But if you just click on that magic button that says choose segment, it's going to give you this dropdown and you can pick another one. And it's just a matter of putting a check market. And then to remove it, if you click on that choose segment again, one of the options is to remove it and it'll just make it go away. The other real beauty of segments, you can apply them to older data. You can't do that with Google Analytics filters. They only take account of data from the day you created the filter forward. This way you can actually go back and filter old data. It's hugely valuable for example. How many have an analytics spam filter in their Google Analytics account? Yeah, I got news for you. All the problem that we have on WordPress with bots attacking our website, they also attack Google Analytics and they will give you all kinds of useless data in there. You can create a segment that will filter out all of that bot traffic. If you want to know more about that tweet at me afterwards and I'll point you to a good resource for fixing that. But that's an example where it's great to fix that from now on going forward. But if I have to report to a client from my last year over year and a whole bunch of that is spam data that I don't know what's what, it's kind of nice to create a segment and nuke all of that data in my report so that I'm using real numbers. Next problem. My favorite issue to deal with and it's gonna get right up in everybody's face is starting in July. We really don't have, the typical ways of measuring website speed really don't give us a good handle on what's happening with speed on our website. I got bad news for you. Speed is not much of an SEO ranking factor. It got a big deal made of it but if you actually read the specifications it applies only to a tiny number of websites and you have to really, really, really suck at speed in order for it to penalize you. John Mueller from Google himself said like if we're talking in the range of 25 to 30 seconds you're going to have a ranking problem compared to the guy who's five or six seconds. Well, geez, if we're up at 35 or 30 seconds we got a lot more problems than our SEO. But the problem is, I mean, for example if we use something like PageSpeed Insights, Google's own tool unless they're a huge website with a massive amount of traffic your speed is unavailable. They're doing that as an aggregate from their toolbar information and you just don't get to know. You got optimization, medium, score 74, WTF. There's the ultimate. This is a Pingdom score with a score of A and a load time of 60 seconds. Okay, so my point is those numerical or letter scores are pointless. These tools are great for pointing out to you technical elements that you may need to take into account. They have not kept up with modern technology. None of these scoring tools take into account you're using HTTP2 on your server. Well, if you're using HTTP2 you shouldn't be combining all of your JavaScript files into one file. But why slow in PageSpeed? Both tell you to do that regardless. Well, no, don't undo all the work I did by getting a server that has HTTP2 for God's sake, just leave me alone, ignore it. So you have to know which of those points to ignore and which ones to keep. This is Gemetrics, which is a favorite tool. If I showed that to my client my first reaction would have to be, oh my God, we have to fix your website. Your speed is killing your website. Except when we do an actual speed measurement it's under two seconds. Am I going to invest that client's resources in my time in improving on two seconds? Not likely, I'm going to bet I can find bigger fish to fry. If you really want to get useful let's go to page load timings. This is available in GTmetrics and it's telling me, this is the one that Google is just saying, they really care about. Time to contentful purse paint. That's when the page starts to become visible and usable to the visitor. Your banner is showing, your headline is showing, your text is starting to show, there will still be stuff at the bottom of your page that's loading, your footer may not have loaded, you may have ads in the sidebar that won't have loaded but the visitor can start reading, can start using, can start making sense of what's there on the website. That's the cue we all use for whether we're happy with a website. We don't care how long it takes for that crap in the footer to load. We want to know how long we start getting feedback that says the website is useful to us. If that contentful purse paint is over two seconds people get itchy, they get trigger happy, they hit the back button, they go looking somewhere else. Okay, conversely, if you've got a load time of eight or 10 seconds but your time to first paint is one second, go find bigger fish to fry, that's not a problem for your visitor, they're getting feedback that says stick around, there's value to be found here. Okay, you have to be logged into GTmetrics with one of their free accounts to see those timings. That tab only shows up if you're logged in. So actually, you just go to GTmetrics, and ask for an account, you'll set up a free account, and then any time you do a test, there will be a tab for the timings underneath the test results. Sorry, the question was when you're logging into, or when you're going to the site, how do you use the site? It's literally just got a box at the top where you type in your URL for it to test, but if you log in first, it'll give you more customization options and more data. Webpagetest.org is still my favorite tool for this kind of one-off testing. It's got the most number of options, but it means that you have to take more trouble to configure them. This is my favorite standard setup. One of the beauties we have, they have a testing server in Toronto. So if we're trying to test for websites that serve clients in our region, you're gonna get a more accurate result here. The closest that GTmetrics can test from is Vancouver. Okay, you can select different speeds, connection speeds. A lot of those basic tools are testing straight off the internet backbone in a data center. That's a speed and a latency that none of us in the real world will ever see. Pingdom is famous for giving results of 768 milliseconds. And in the real world, it's four and a half seconds. So that's an advantage here. Again, this is under test settings, and you can customize them. The other thing that I love about Webpagetest is you can test the first view and the repeat view. That means you're testing your browser caching. Is it configured properly? How big an impact is that having on the speed? Okay, the other huge thing, you can do this a little bit with GTmetrics, but it's easier here. Starting in July, Google is switching to the mobile first index. All of their assessments of your website are gonna be made on a mobile view of your website using a mobile web crawler, including speed. Now, we all should be doing this anyway, because again, speed isn't about satisfying the rankings. It's about satisfying our users. And if you look at your analytics and you use that mobile segment, you're gonna find out that for the average website now, there are between 30 and 50% users on mobile. And depending on the site, new sites and event sites, 70 to 80% on mobile. So we need to be making that the first focus of the testing that we're doing. Fairly straightforward here in that we can create settings. There's actually a default. This is a default that's available from the dropdown. It's for LTE. It's kind of a typical user in the city, the speed they'll get on a phone. And in particular, you'll notice it's the 70 milliseconds, it's the latency that's much higher on a phone. But that's gonna give you a closer one-at-a-time idea of how fast your website is for a phone user. The big problem is we're still missing a whole range of variables that have major effects on the user's experience of our website. Where are they? Are they in Toronto or are they in Vancouver? How fast is their own network? These are estimations of a network speed, but it's not the user's actual network speed. Are most of your visitors logging in from a company that has 2,000 people sharing a single gigabit internet connection? Or is it a family of five where three people are on the TV watching Netflix and somebody else is on stream and then the last person is trying to look at your website? None of these one-at-a-time processes take that into account. So we're gonna use this secret little element built into Google Analytics. It will tell us. It has that tracking built in. It will tell us the actual experience of actual users on actual pages of our website. The other big problem with the one-off test is you can only test one page at a time, which is really difficult. And if all you're testing is your homepage, you have no idea whether your product pages are too slow to convert, okay? We're gonna get around those weaknesses of the other testing tools, the location, their network speed. How busy is your server? If everybody's visiting your website in the middle of the day and that's when your server is heavily loaded, they're gonna get a much slower experience because your server's slowing down. Or maybe it's not and you're okay. Real difference from mobile and usable. You're actually gonna see the variations in speed for mobile users, okay? It's called RUM, real user monitoring. We're gonna see the real user results. Big one at the bottom, you're gonna see the effects of your caching. Everybody says, oh, I put caching on my website so it's faster. Up to 40% of your visitors aren't getting a cached page. It's expired before they get to it. This will show you the real experience of that. You knew there had to be a gotcha for something so nice and interesting. Only 1% by default of the page views are tracked. But all it takes is a customization of the tracking code in order to get the other 99% tracked. To a maximum of 10,000 page views tracked per day, that's the maximum Google will give us. So if your site is more than 10,000, you'll want to set the tracking code to something a little bit lower. So this is the tracking code that you customize. The part that's outlined in purple there, this is the script. And all you're doing is adding site speed sample rate 100 at the end. If your site's over 10,000 visits per page, scale it back proportionally because you don't want it to measure all the pages in the morning and none of the page visits in the afternoon. That just wouldn't be a good assessment. If you're using WordPress, Google Analytics dashboard for WordPress has this functionality as a one-click check mark. There are other analytics plugins that have it as well. You're gonna see, I'm gonna mention this a number of times. It's the broadest applicable plugin that I've come across and I've been using it for years. It's just been bought by the same guys who own Monster Insights, but they've sworn up and down that they are not going to make it a paid product. So all I can do is go by their promises at this point. So you're gonna see mentioned of that a couple of times. This can also be added very easily if you're using Google Tag Manager. It's a variable that you can automatically apply to your tracking code inside Google Tag Manager if you're using that. But this is an example of what you're gonna see when you ask, using the section, it's under the left-hand sidebar. So you're gonna click behavior, site speed, and then overview. But look at the difference between the speed for mobile users compared to desktop. You're getting real numbers. This is what your visitors are actually experiencing. 4.7 compared to 2.5 seconds. Now that's an average. Remember what I said about averages? Because that's averaging in all of the different mobile browsers and all of your different visitors and all of your different pages. What's the next thing we gotta do? What was the first screen? Segment, right? We gotta now start dividing that out and looking at the speed for individual pages and assessing what impact that is having, right? So here's an example. In this case, this is what's happening by segmenting by browser. I'm still dying to find out who the Blackberry user is on this website. But we can also get automatic reports for speed by country and by page. So when you click on page, it's gonna give us the actual speeds experienced on the different pages of the website. Okay, now again, this is average for all users on mobile. Right, I'm using a segment here that is mobile users. I could also do a comparison of desktop and mobile for these. But what I wanna look at here is we've got a page. This is a Flores website. We've got a page that's for everyday flowers and we've got a page that's for birthday flowers. I know from this website, those are functionally similar pages. They're designed to do the same thing. One's designed to sell flowers to people for general occasions and one's designed to sell flowers to people for birthdays. So they serve pretty much the same purpose. I'm trying to get a context that matches between two pages. But look at the difference in bounce rate. There's 2.2 seconds load time and 44% bounce rate. There's 5.14 seconds load time and there's a 77% bounce rate. There could be other issues with that page that are making people not stick around on the page. But when I see a doubling in the speed and a doubling in the bounce rate, the effective bounce rate, that's where I'm gonna go and look first. I'm not gonna waste my time looking at the speed of all of my website. I'm gonna find those outliers and I'm gonna focus on those first. My bet is, and it came through, somebody had loaded a massive big product image on that page. And the minute you start thinking, like I see all the heads going, right? Like logically, when your analytics notifies you of that problem, your brain says, oh, I know what that was. Martha uploaded one of those damn cell phone images again. How is that possible? Sorry, I shouldn't have said Martha because half the time it's Mark too, okay? The next thing we wanna do, remember we're talking about focusing our time on the pages that are causing problems. So I wanna go over a couple of the terms really carefully. There is a very specific definition for bounce rate. An Avinash Kaushik's definition, he's the Google analytics guru from Google themselves. His version is, I came, I puked, I laughed. But the actual technical definition is, I arrived on the page and triggered the analytics tracking code once and I did nothing else on the page to interact with analytics again before I left. Again, you're gonna see why that's important in a minute. So in general, they got to this page and they didn't like anything else about the page so they left my website before they viewed any more pages. Landing page is simply the first page that they saw on your website. Have a look at that report because you're gonna be surprised to find out that less than half the people on your website start on your homepage, which is why just testing the speed of your homepage isn't gonna be enough, okay? And exit page just means it's the page they were on as the last page before they left the website. Okay, so what's a good bounce rate? Anybody? The one thing I will tell you is if your bounce rate is less than 20% for a consumer website, you have broken analytics tracking, okay? You do, it's a truism, it's an instantaneous diagnostic. It usually means you have the tracking firing twice. So go and fix it because all of your data from page to view level is inaccurate and shouldn't be trusted, okay? So just use that as a line in the sand. If it's under about 20%, you really need to check your accuracy. The reason I say it depends is it depends on the purpose of the page. If your contact us page has a high bounce rate, that's probably good news. It means somebody wanted to know where the hell you were. They Googled you. They found the phone number for Freddie's catering and they landed on your contact pages. Oh, there's their phone number click. Well, that's a bounce. That's a highly successful bounce, but it's a bounce. So you don't wanna consider that bad. It simply is what it is. It's a number, oh, I got a 95% bounce rate on my contact page. Well, that's kind of its purpose. Come, get the information, please. If it's your home page and you're bouncing at 90%, you got a problem for the most part. Now, maybe your home page has your phone number and a big call to action saying the easiest way to get a hold of this this week is to call us, okay? Take those contacts into account. Would you think that your blog home page would have a high or a low relative bounce rate? How many would say hi? How many would say low? For the most part, it's gonna be high because it's gonna be people who are gonna come see what they expected to see. They don't know you yet, they're likely to leave. It's also one of your biggest opportunities. If you can get that home page bounce rate down, it means that you're getting them to come to your blog post because of the blog and then they're going on to be attracted to other things. Okay, so this is the purpose of bounce rate is analyze whether it's good or bad in that circumstance and then improve it. So if my, let's pick product and category page, just as an example. So if you've got a product category page, my blue running shoes, all the different brands of blue running shoes, if the bounce rate on that page is really high or really low, it doesn't matter. Whatever the number is is a starting point and you want to drop it from there. If it's 30%, great, congratulations. Get it down to 25. If it's 80%, you got some work to do, get it down to 50. It doesn't matter where it's at, it's what you're gonna do about that and it's the fact that you can track it, you can make some adjustments, you can make some changes, you can make some optimizations and then check it again in two weeks and see whether it's bounce rate has improved or not. Okay, a lot of analytics is about trends, it's not about point-in-time data because you can't compare your point-in-time data to the next website to know whether it's good or not because of that very first slide, it depends part. What if you have a call-to-action question somewhere that relates to, say, a membership page? That's not gonna bounce. As long as they're going to another page, that's not gonna register as a bounce. A bounce is only, they came to only one page and they left without doing anything else that had an interaction to analytics. Yeah. Can you just be quickly about any time when they didn't get a single... Say again? Like a single page, I think, just a single page. Yeah, if you have a single-page website, the question is, how do you assess bounce rate on a single-page website? That's one of the reasons why single-page websites are so hard to optimize because you don't know. So interestingly, we're gonna get to the next thing, which is, sorry, this is just an example of segments, of the way you can divide out the types of visitors before you start assessing bounce rate, okay? So obviously your organic visitors are probably gonna have a higher bounce rate. They're the ones who just kind of discover you on their own from the search results. If you've got a high bounce rate for your paid ads, you're getting the wrong traffic. You're paying money for people to use up your server resources and not earn you any money. That's a real signal. Again, what is too high? Doesn't matter, get it lower. Use it as a starting point and improve it, okay? Same thing with mobile versus desktop. When you do that look and you see the difference in bounce rate, you'll know that your mobile site needs better optimization, and it's almost always speed-related. I am gonna get an answer to your question, but it's more easily explained on a slide that's coming up. So this is an example of the way I'm looking at some of these. I'm picking out outliers. There's a term in Aikido called soft eyes. It's the idea of not deciding ahead of time what you're gonna find when you look at the page. Look at it and watch for things that don't seem to line up. They seem to be anomalies or outliers or exceptions. So in this case, I've got a situation here. This is a directory site from down in California. They have a blog post for the running paths about the running paths in La Jolla. Well, if I'm looking at some of this information, I'm getting 117, I'm getting an 88% bounce rate. Sorry, I'm getting a 94% bounce rate. 93, 94, sorry, this is all users compared to organic search. They have segmented out. So I've separated the average of the website from the organic search visitors. I've got a bounce rate of 94% on that thing, but it's getting me a fair number of visitors. That's where I'm gonna spend my time trying to put something on that page that encourages people to look into the rest of the site. Is it another page with a map? Is it another page that says if you are a jogger, you'd probably like to see our bike trail map as well. I'm gonna start using my understanding of my visitors and my website to give them somewhere else to go and make it more obvious on the page how to get there. There'll be a call to action, there'll be something that helps them understand that there's somewhere to go next. Because right now, I'm giving them one piece of information and they're running away. Conversely, if you look at the, this is the events page, you can see that it's got a fairly low bounce rate because people are seeing the overall listing saying, oh, there's an event I'd like more information about. What's the date for the dog run? Yep. Is it time on page? It can, it's really misleading because our people, the question was, can time on page make a difference? It's really hard to know what people are doing on the page while they're sitting there. Is it sitting open in a tab somewhere or are they actively reading the page? So we're gonna get around that. We need more detail. This is the combination of these two questions. A bounce page doesn't tell us enough about what happened during the viewers looking at that page. Did they spend 10 minutes reading every word and hanging wrapped attention on everything we said? Or did they read the first headline and go, boop, we don't know. Event tracking in analytics is the solution. It gives the analytics another interaction to track while they're on that bounce page. Again, a pitch for Google Analytics dashboard for WordPress is a bunch of check marks. You can track the number of clicks on external links on every file they download, on the phone number. Because if they're tapping that click to call on their phone, that contact view of the contact page that bounces, that's not a negative result for me. If I know that they clicked on the phone number during that bounce, that becomes the most valuable visit I've had that day rather than, oh crap, I've got a bounce rate on my contact page. Okay, you can track, are people viewing the videos? If they're not, get it off the page and replace it with something else. Scroll tracking. You can actually track how far down the page somebody is reading. 25, 50, 75, 100% down the page. That's way more useful than time on page because they could be leaving it open in a tab for 20 minutes and then closing it without ever having read it. Key on a one-page website because every visit is gonna be a bounce. You have to instrument the events so that you can understand what's going on. Can you mark down your question? Remember, we'll catch it at the end. I'm just, I wanna make sure I can get through everything that we wanna cover. What do people want us to write about? It's built-in-day analytics. It's called internal site search. Your own website, people are going to that little search box at the top and typing in what they're trying to find. And they're using their own words. They're telling you what words they use for that search. Then you have to decide, are they in that search box because I don't have that information on my website because I can't find it on my website even though I do have it? Well, one says I need to fix my navigation. The other says, hey, I got some new content ideas to write from. Okay, or I do have it and it's easy to find. It's obviously not answering the question that they have, I wanna know that. Quick easy steps, two buttons to clicking your Google Analytics account. You enter what is called a search query, the query parameter in WordPress, it's just a letter S and then you click save. And now inside your Google Analytics reports there is going to be a section for exactly what the behavior was of every visitor who typed something into that search box on your website. They're gonna tell you what words they used. They're gonna tell you whether they viewed another page after they entered the search, which means they found something they were looking for. They're gonna tell you how much longer they stay on the website after the search that they made, you can now assess the quality of your search tool. And I'm sorry WordPress, but the default tool in WordPress sucks rocks. Get a good search tool in there if your website depends on search. A lot of people use search as a navigation. They don't even bother looking through your blog posts. They just type in how do I and their subject. So this is gonna tell you very quickly that your search tool is not doing its job. Sorry, the example here is I've got two terms I know for a fact do not exist anywhere on that website. Guess what I'm assigning my content writer next week. And you're gonna assess it, right? You're not gonna write content just because there was one search in six months. What about our site architecture? Is it working? Is our navigation working? Which navigation links actually get used? Do not use UTM tracking for internal links on your website. You will ruin your data. That's what event tracking is for. But even easier, turn on advanced link attribution. Setting in Google Analytics, click the two buttons. Again, you'll get the slides after so you can see the instructions. This is all it's doing. It's adding this particular, sorry. In addition, you're gonna add this to the tracking code on your website or click the button in Google Analytics dashboard for WordPress, you see why I'm starting to recommend it. And now what you're gonna get when you go to your website, you're gonna install this Chrome extension and you're gonna see percentage numbers for every single piece of your navigation. It's gonna tell you what percentage of visitors clicked on that navigation, what percentage of all the clicks on the page went to that navigation. When I start to see things that have no clicks on them on the entire navigation over a six month period, I start really reconsidering whether that should stay on the main navigation. Should it drop into a footer? Should it be in tiny little text at the top or should I remove it all together? Because it's using up valuable space and SEO juice that it's not earning. The great thing is it even gives that data to you for the dropdowns. Okay, it's now only visible if you're using this plugin. You used to be able to see it inside Google Analytics, it's not there anymore. So grab that Chrome plugin and tuck it in there and you'll be able to see those results. And use more than your homepage. Look at what happens on your product pages. Look at what happens on your blog pages because they will be different. That's the whole point to our segmentation question. Which pages could I be getting more traffic out of in the first place? This is that idea that they're almost there but they're not quite earning their keep. They're not quite getting the attention that they deserve. So conversely to that jogging routes page, that was getting traffic. It was just hitting and bouncing. This is stuff that's not quite there. We wanna focus on the low hanging fruit. The easy ones to get there. Google Search Console is where we need to go for this. How many have Google Search Console set up for the website? For those of you that don't, the first thing you need to do when you get home tonight is set it up. That's how Google's gonna warn you that they've detected a problem with your website. The email address in your Google Search account is what they will notify you with. So if for no other reason do it for that, but you're gonna see what we can do here. This is the section on the left-hand side under search traffic for search analytics. We're gonna actually get Google Search Console to give us a list of all of the pages on our website that rank between position 11 and position 20. Which counts as page two. Everybody's heard the old SEO joke. Where's the best place to bury a dead body? Page two of the search results because nobody ever looks there. Well, our goal is buried on page two. Because if I'm gonna work to improve the ranking of a website to get it more traffic, do I wanna start with a page that doesn't have any rankings at all? Or do I wanna start with a page that's just that close? Well, sorry, I'm a lazy son of a gun. I want these stuff. You're gonna select clicks, impressions and position, and then specify that you wanna know about it by pages. Now, by Google Search Console normally, it's gonna give you the queries. It's gonna tell you the keywords. It's also the only place to get that data accurately, but we actually wanna know about the pages. And this is an example of what it's gonna give you. So I've asked to sort by that far right hand column by positions. And I found a couple of examples of stuff that it's at position 15. I'm getting 32 clicks out of 3,600 impressions. I'm getting 49 clicks out of 900 impressions. There's a lot of activity there and I'm hardly earning any of it because I'm on the second page of position 15. Those are the pages I'm gonna go and look at. How can I improve the page title? How can I improve the page description? How can I up the content on the page? Maybe the damn page just needs to be re-updated. It's that jogging trails page and it hasn't had new maps added for two years. Those are the pages I'm gonna start with and I'm gonna work my way right down those opportunities. Always work off the comparison of the number of impressions and the value that's there. There's no point in optimizing a page that you're at position 15 for that gets three impressions at a time. The new Search Console beta gives you 16 months of data but it doesn't let us sort this way easily. So I'm gonna say stick with the base version of analytics for now or Search Console. We're picking the stuff that's living in the middle of page two and optimizing it to drive it to page one. Why do we have to keep diving into analytics every time we wanna know what stuff is happening and what the important stuff is? Well, we don't. Let's get analytics to do the work for us. First example is intelligence alerts. Very top right-hand corner of your Google Analytics you're gonna find a button for intelligence and it's gonna give you this kind of a report. It's gonna use artificial intelligence to pull out things that it thinks are different about your website this week or this month and tell you about them. That's level one. You can actually type in plain English questions. How many visitors did I get from Ontario compared to Canada last month? And it will interpret those words and give you the results in your report. So that's the first level of intelligence alerts. The second thing we can do is create a custom alert and I do this for my page speed. For all of my critical pages, any critical conversion page, if it's a sign up page, it's the sign up for my newsletter page, I wanna know whether the speed of that page slows down substantially or not. I can simply tell Google Analytics, monitor page speed, page load speed is what we're looking at and tell me if it goes down by more than 25% or 30% or 50%, whatever your leeway for parameters are. So I'm gonna get a notification the day after my page slows down rather than a month later when sales have dropped off by 30% for that product. Now you can do it aggregate over a whole section of your website or you can do it for individual pages. But this is analytics warning me that it's detected a problem that it's gonna take me weeks to notice on my own unless I'm digging in every day and checking the speed on every page. These are just some examples of the kinds of things that I've included. If I get a jump in referrals and I'm expected to jump in referrals, I wanna know where that traffic's coming from because I probably wanna get a hold of that website and start a conversation, get into their comment section, have a conversation with the website owner about getting them some products to test. You know, whatever it happens to be, use the context of your website to understand what the value of that connection would be. But that's telling me somebody, a whole lot more people than usual have linked to my website. You can do the same thing with social. Hey, I'm getting a whole lot more attention from Facebook. Well, go and find out who mentioned you. It's gonna take some spelunking, but you can do it. Page speed slowing. Any of the normal metrics that you're looking at, you can assign an alert for. So Google will warn you in proactive fashion instead of you having to go and look for it. This is an example of the page speed load. This is a wine website. New Zealand is having problems. Turns out we had to put a CDN in place that would better serve New Zealand. The CDN we were using didn't have a node in New Zealand. And unfortunately, New Zealand and Australia are really hard to serve high-speed websites too if your website isn't located in one of those two countries. So we literally switched CDN because our CDN didn't have a node in New Zealand. It was being served from Australia in the speed when nobody's buying wine when they gotta wait 35 seconds on the order page. This is just an example of the setup page. So I want the condition to, the alert, one page load time increases. It increases by more than 25% of the same day the previous week. It's really dangerous to compare it to the day before because there will always be fluctuations and there will be just all kinds of issues and things. So it's almost always better to compare it to a period the week before rather than the day before. Because otherwise, you'll just get a string of the same warning over and over again. These are just an example of some of the other useful alert possibilities that we talked about. Social traffic, referral, purchases of a particular product. Hey, if everybody's buying blue running shoes today, I'm gonna put blue running shoes on my homepage as my primary advertisement, right? That's proactive, that's analytics, giving you data that you then use to change your behavior on your own website to better appeal to your visitors. Any time you pull up a report in Google Analytics, any changes you make to customize it, let's say you're looking at a view and you decide, okay, I'm gonna put a custom segment of mobile and I'm gonna put a date range of the last two weeks compared to the two weeks before that and I'm going to pull up a report that is of a particular page, a particular product. Once you've done all that configuration, if you ever wanna see that page again, you have to repeat all those configurations. Way up in the top corner there is a save as a shortcut. It puts a little shortcut in the left-hand side that saves all of those customizations that remembers the segment. It remembers that you compared two weeks to the last two weeks. All of that stuff will be saved. So when you go and look at that exact same report the next time, just click on the shortcut. You don't have to dig all the way down into the interface and reset all those checkboxes and remember how you set it. Just create a shortcut for it. If you wanna put a bunch of shortcuts on one page, oh, sorry, this is the little power users tip, use the hell out of that calendar in the top right-hand corner. Remember we talked about segmenting? You can also do comparisons. You can compare the last two weeks to either the two weeks before that or the two weeks at the same time the year before. That's context. That's telling you what's happening compared to before. Okay, really decide on your seasonality because it's risky to compare to the previous two weeks if the previous two weeks included Christmas or Black Friday or whatever your situation happened to be and use annotations. Make life easier for yourself and all of the rest of the people who use your website. You click the little drop-down arrow here. You type in your annotation, sorry, you click the drop-down arrow to get this view. Right-hand side, you click create a new annotation and you can make a note that will live in your analytics forever. Today, I changed the configuration of my homepage. Oh, look, my bounce, because what's gonna happen is two weeks from now, you're gonna go, oh, look, my bounce rate on my homepage has dropped by 20% or somebody else is gonna say, oh, great, our homepage bounce rate dropped by 20%. And then when they look up and analyze, it's like, oh, yeah, Paul, yeah, that was the day change to lay out of the homepage. Okay, so this is a way of making notations that everybody who's looking at your data can see and remember. Get mentioned in a newspaper or on the radio, make an annotation, because you're gonna see a spike and six months from now, when you're trying to do that report, you wanna remember why it spiked instead of like I've had to do spend two hours trying to remember what on earth happened on June 3rd that caused it to spike. You can collect a number of those reports and make a dashboard out of them. And a dashboard simply collects a bunch of data that never exists in its own report altogether. You're creating a report that pulls data from separate reports into one place. Nice enough to create a gallery for us. There's some really good dashboards already pre-built. Great blogger dashboards, great product sale dashboards. Have a look at those. Again, it'll live in the left-hand column of your analytics. You pull up the dashboard and it'll update all of the information. Let's use an example. So this is a blogger's dashboard for that directory site that I work with. And they, the writers don't need to see what's going on the rest of the website. But this is showing them things like where's the majority of their traffic coming from? Which posts are getting the majority? Which social networks are bringing them the majority of their traffic? What towns are reading the traffic? So they can look at all of that in one report really quickly. They glance at it on Monday mornings when they're planning their content schedule for the week. Again, it's providing the data to us instead of us having to go and look at 15 different reports and remember and combine them into a spreadsheet and drive ourselves a batch. So this is just a reminder of what we've worked out. We're gonna segment the heck out of everything. We're gonna monitor the real page speed. That only works once the site's been running for a while. The one-off page speed checkers are great if you're in the middle of trying to optimize a page and you wanna compare it from five minutes to go to five minutes later. But when the site's been running for six months, that stuff's gold. Correct the underperforming pages. Find those pages that are having high bounce rates and usability problems and fix those problems. Get the content up there. Visitors are typing into your search bar. They're already asking for it. You know they want it, give it to them. Adjust the architecture and navigation so that people are finding what they need and you know they're finding what they need. Work on those successful pages. Get the ones that are that close. Get the software to give you the data so you don't have to spend all kinds of time digging it out. We covered lots of questions in between. I'm around for the rest of the afternoon. I'm happy to talk about this silicaus come home, but I don't wanna leave you too long not being able to get to the other end. So. So there's the link there. It's also on Twitter if you want a clickable version. Your presentation is a problem, really, in depth. It's tough, because this is so step by step that it has to be lots of data in the presentation rather than you know, coop pictures of cats. But the main purpose is so that you've got that, you know, those instruction slides and things. You can actually follow them in this copy of the presentation so that you can work with them from there. But like I said, I'm around. I'm happy to talk about it.