Darryl: Hey there, welcome to My Bloody Website, the show where we talk all things online, for small and medium business owners or executives who still refer to "their bloody website." I'm Darryl King your cohost ...
Edmund: And I'm Edmund Pelgen, the other cohost.
Darryl: Hi Edmund, how are you today?
Edmund: Good Darryl, yourself?
Darryl: Good. Episode Eight. If we keep this up, we might actually get to ...
Edmund: A full season, and then NBC will sign us for the second year, and we will be totally on Netflix.
Darryl: Yeah, okay ... back to reality. We might just talk about our topic for the day, which is speed and why it matters. Everyone wants to go faster, Edmund; well why should they?
Edmund: Look man, I'm going to be asking the questions today, because you're the master of speed, right? And when we talk about speed, we're talking about website speed ultimately, right? And I'll preface this by saying Google loves speed, but I'm going to, where it matters for our customers, is website speed. Darryl, tell us your experience about all of the websites that you come across, and your speed experience and why it matters.
Darryl: My speed experience? [inaudible 00:01:12]. Okay, why does it matter? Yes, we've talked about, you know, Google likes it, and Google likes it for lots of reasons. But we as people that run websites, business owners, marketing people, salespeople ... we should love speed too, because what it's really about is the performance for the people we're trying to serve, the customers or users of what we do. So if something is slow, if it takes forever to load, if it doesn't work right, there's a really, really, really high correlation with people just giving up and going to a competitor or doing something else. And our lives are all at pace; everything we do these days is at pace. And ...
Edmund: Is there data to back that up? I mean, is there data to say if my website loads slowly?
Darryl: Oh totally. There's a whole lot of statistics about speed performance and how it affects conversions and [inaudible 00:02:05], all of those things. So, speed that we're talking about, is speed to get a response, speed for the webpages to load, speed for any processes and transactions to occur. And, that's all. You know, we're not going to talk about the detail behind it and other sorts of things and what tools might run faster or slower and how you make them run faster, necessarily, but that's what we mean by speed.
And it matters because, it matters to the end user. If they get a poor experience ... Think about it like this: if you stream Netflix or movies through iTunes or whatever it is you do, you know the whole buffering, like "Ah, it's taking forever" ... It breaks up the experience, alright? And then it falls out, and you've got to go back, and you've got to reload it. No one watches a movie if that happens. They give up; they walk away. And that's the same experience.
Edmund: So how does the average business owner, someone like me, someone else, how do they determine whether their website is performing at speed, is fast or slow? Like, how can they make an objective assessment of it?
Darryl: Well there's a subjective thing they can do first off, and that they should do regularly. Most people don't look at their own website with any degree of regularity, okay? They don't go there. And they go, "Yeah well it looked good last time I looked." So, a great thing to do is to behave like the person you're trying to service. Get your phone: wherever you're sitting, in the lounge, you're sitting out at the pub, the beach; try loading your website. What does it look like in the phone? Does it load fast? What is like going 3G and 4G? Don't just do it on wifi in the office: "Oh yeah it looks really good to me!" You know, like ...
Darryl: Do it like someone using it. What does it look like on one bar, two bars? You know, like have a look at it. So those are the things you can do, and that's where cranky boss messages through to the ... "I saw it and it looks like rubbish, aaah!"
Yeah so, okay, so that's what I think it's experienced like. Now, sometimes it's not. Sometimes it's just a bad connections on that day, whatever. But, that should be a red flag to say, well maybe is there a method to do a more robust test that can be more consistent?
So then there are tools. You know, Google has its page speed test that they offer; people can go and find that. But I really using a site called GTmetrix.com.
Edmund: How do I spell that? G-T-M-E-T ...
Darryl: GT, metrix. G-T-M-E-T-R-I-X. We'll put that in the show notes.
Edmund: You'll put it in the show notes.
Darryl: And GTmetrix allows you to go there, it's really simple; it's like a search box on the front. Put in your website URL, the full URL. So there's several ways to test it as well. Put in, you know, it's going to be "https" because you listened to Episode Two, so we're at Eight now; if you haven't got https by now, you need to come around, you deserve to be spanked because you should've got that.
But you've gone and put in "https://," the full URL, and test that. Now, I'll talk in a minute about why you should test it other ways as well, but test that. You hit the button and let it do its thing. It spins, blah blah blah, comes back with a report, and it does a page speed score using the Google method, and it also has a YSlow method and it tests some other things as well.
Edmund: Do I need to care about what those are?
Darryl: Not in massive detail until you get the red flag. So you know like what happened that day when it went orange and red for you and you came around and you were crying and I had to hug you and ...
Darryl: It's not the end of the world, Ed. But it gives you a rating. And so, it uses colors and numbers. So like for example I'm looking at a site, page speed score A, 98%; YSlow B, 87%. So they're different. So that's what I like about this tool because I could go to one and go like, "Aah, it's perfect." But in another environment it's not quite. So this gives you a bit of a multiple snapshot of how it's performing.
Edmund: And I think it's important to mention too that whilst they're different tools, you might go away and do things to speed it up, it's a great way of comparing relative performance like, "I did it yesterday, and today it's this." And it's a great tool from that perspective, yeah?
Darryl: Yeah, and you could create a free account, log in, and save it, and regularly do it. Now there's a couple of things you have to think about. Where does speed matter from? So, if I'm a dot-com, so one of my sites has a global audience; it's a dot-com. So speed is going to be different by the tyranny of distance. If I'm in Germany, and it's hosted in Los Angeles, there's no escaping the fact that there's a lot of pipe between those two environments. So there are natural delays. So I need to test it and get an understand of it.
So what GTmetrix also allows you to do, is you can change a setting, and we won't talk about that here, as in how to do it; you can choose to be from Vancouver or somewhere in the U.S., from Australia or in Asia or Europe. So you want to look at it for the largest segment of your audience.
Edmund: Right, so it allows you to simulate the behavior of your target customer basically.
Darryl: Yeah. If you're an Australian customer you're going to choose Sydney, right? And you want to see what it performs like from there. Now, that's important because you're trying to fix user experience. The secondary part of that is, good speed for users is great from a Google perspective; they like speed. But, don't do it just for them, do it because it makes logical sense. Have a better experience to work from.
So, make sure you're testing for the right place, because you might have B and C, and then you change it, you know you're in Canada, you change it to Vancouver, and suddenly you're As because you've done most things right, I mean it wouldn't be that extreme, but you've done most things right and the distance has an impact.
Edmund: So this tool gives me this A and B, or what rating out of A, B, C, D, F; what other data does it give me that me as the average person can understand?
Darryl: Well, probably not a lot! But there's a couple of things, so it will give you a fully loaded time. So people, when you think about it, so when the first byte's made, that's when my connection, the first byte's sent and responded to, to the destination, where the website's being [crosstalk 00:08:23] right?
Darryl: But then from that point on, there's all sorts of things. In one of the episodes way back we talked about packets and all that. It's a whole lot of data, got to be transmitted. And then the page is rendered, fully rendered: here is the whole page, everything loaded. Even the hidden analytics code in the back that came from [inaudible 00:08:37]; it's all loaded. So for example on this site, 1.9 seconds, fully loaded time, we've got a green arrow to say that that's an acceptable range. I mean, hey, if you can get under a second, brilliant; there are ways to keep tweaking it. This particular site's got a lot of heavy media on it, so 1.9 seconds to load big media are considered to be pretty good.
It'll tell you page size; that's one that everyone can understand. We've all tried to upload a 10-MB file to Gmail at some point, and "Oh no, we're going to stick it it Drive because it's too big," and it's taking forever to upload it, whatever it might be. So, if the page size is three, four, six, eight, 10 MB, you're never going to get it super fast. It's too big.
Edmund: So on that topic there, I know we're going to talk about, you know, what data the tool tells us and then how we can go and really super speed up the site. But as a business owner, what can I do, or what can I not do, that will slow my website down? As in, when I'm writing content and don't ... What are the things that you see that people always do wrong, that ends up slowing their sites down?
Darryl: The biggest one, is everyone wants a carousel, a hero carousel on their website. And statistically, there's a lot of numbers and studies done that say that, you know, you get five slides going through, that the retention rate and the click-through rate once you get to slide two, that just falls off a cliff. So like one, you know, 78% of people see it and might interact with it, and then it gets like 30%, whatever, it just gets down to almost zero. And most carousels, "Ah we want full width, we want this" ...
Edmund: Images ...
Darryl: You know, each one is 400 kilobytes ... You know, because it's got to be crisp and all the rest of it. So we have these massive carousels, and there's one to two MB of images that people might not even see or interact with.
So, when you give direction to the people that are building your site, you design team or whatever, think about it. And, you know, we'll talk about mobile first and mobile design later on, in some other episode, but think about the experience on a phone. If I am getting 3G, 4G connection with a couple of bars and it's not great, trying to load 2 MB just to load the top section of a site, that's a decision that you made; you can stop that. You could tell the designer, "Let's keep it fast; I want good looking. I want to use imagery." But if you then stick seven, 10 other images on your homepage, and embed a video that's been pulled, and all those things, there is a cost of loading to them.
Other things people do, because we go and get SEOs, and people that help with conversion optimization, well I've got this AB testing tool, and I've got this analytics tool, and I've got all these things running the background of my homepage that I forgot I'm even using. And they're all adding to the code behind it.
But, you know, writing a bit of content will not make your page really big. If your site's just all text, unless you've got 20,000 words on one page, it's not going to be the cause of it. The cause is going to be graphics, software, and construction. Graphics is the killer, yeah.
Edmund: Okay. So I'm uploading pictures and stuff; why can't I just take a picture of my product and stuff, and upload that picture to my computer and then upload that to the website? What do I do to make that optimize my photos so that they're nice and small without getting all pixelated?
Darryl: So, well you've got to understand that it's not just about cropping something; you've got to scale things, and the formats that you choose. But when you take a photo on a camera, the native way that these things are going to spit them out, they're yay big, right?
Darryl: And they've big files. So they need to be run through image software. There's lots of different ones out there; you can get tutorials on it. A lot of people will upload the RAW one, and expect the content management system's going to do all of the compression for them, and everything, and it won't. You can remediate that; there are plugins that will do it, but if you do it right upfront ... Ask your designer what are the native sizes; make sure the image is cropped to that size. Don't do this big and expect the page to ... [crosstalk 00:12:36]
Darryl: Because that's a mistake people make. It's displayed this big, but the RAW file upload is this big, so that's what's actually being delivered to the browser. That file, but it's just being displayed that way. So I've wasted bandwidth. I've slowed it all down for something that's not actually visible.
Edmund: Right. So basically what you're saying is, optimize my images correctly before I upload them, and if I can't do it, get someone to help me to do it ...
Darryl: Get someone to do it. And then there are techniques to compress them better, but the basic ... Have a look at the options. So, some files will be a better file size as a PNG than a JPEG. Others will be better as a JPEG than a PNG, so make sure that you get the lowest possible quality done, and then whoever's doing that needs to look at them; what will they look like at that size? "Ah well yes, it's great; I can do a 10% JPEG compression but it looks like crap." So don't do that. You've got to find that balance, but try and keep your files under 100k. Like, get them down there.
Edmund: That's images, and I know that's a really important one. But the other thing I wanted to ask you about, because I know we have some personal experience with this, because a lot of people use WordPress, and for us non-developers out there, these plugins for tools are really awesome ways of adding extra functionality to our site and adding some really cool features. Now I know you gave me a bit of shtick about plugins; now, what's the deal?
Darryl: Okay. So you're a serial plugin offender.
Edmund: I hope that I'm not [crosstalk 00:13:59].
Darryl: [crosstalk 00:14:00] get it, right? At the end of the day, using a developer to help you work on your site costs money. And, you know, last week we talked about a lawyer, right? When you're a guest on ... And people don't do things because, there's a monetary cost involved. So I get why I might want to install a plugin to do something on my own website myself, and that really is the beauty of the world we live in, that you can do stuff yourself, and you should have some capability to manage your website. It's not about, just have a web developer do all the work for you, take some action and do it.
The problem there is, plugins are plugins, and they're not all the same. You know?
Darryl: There's apples and oranges and rotten pears and all the rest of it. And you have to be careful about what you install. Now, there is overhead in a system like WordPress for plugins. You know, like when a page loads, some plugins run on everything, every page, every page load; it's invoking little parts of it to see if it's needed. You know, good plugins are written really cleverly; bad plugins aren't. There's the security implications of all these plugins. So the recommendation I have is, run with the fewest number of plugins you can do.
But people go, "Well, I'm not really using it." But you've got it active. It's loading. You're doing things. Even to the point though, even if you deactivate it, if you haven't used it in a month, get it off the site; then you don't have to update it, you don't have to do all these other things that just add bloat to your site.
The fundamental rule I say is, if you want a faster site, you will have less plugins. I've never been able to optimize a site fully down to sub-one-second, running 50 or 60 plugins that are doing, "Well, this one does the little bit here, and this one does that there."
And I think that's the other thing: a good web developer will help you choose a plugin that does the bulk of your needs in an area. So I don't have two or three things. And I've heard instances of two sets of security plugins installed, three backup plugins installed. I'm not looking at anyone in particular, but, "Yeah they're still installed; that one wasn't doing so good I've put this one on." And they're actually running in parallel, which is affecting all sorts of things, and that really shouldn't be happening.
So, find the plugin that does the job you need. Make sure it's a verifiable good plugin. And then get rid of all the rest. Try and run a leaner site, because it's about the user, not all the little widgets.
Edmund: Okay. So those are the things that I can do as the site owner. Now, I want to just step back a little bit, back to GTmetrix, where you said I can go into GTmetrix, put in my site, and it will generate this report; it'll give me a score using Yahoo tool and the page speed tool; and it'll tell me how many seconds it takes to load and how big the files are. Let's just say I am not going to make my site fast. I'm not going to do the work; what do I do with this information? Who do I see, what's the next steps? How do I go about making my site faster at that point?
Darryl: And the reality is you probably don't. You know, it says optimize images. So there are some things you could do. So one of the first things is you want caching. Most people have perhaps some familiarity with caching, because it happens in your browser. So that's when it remembers things, makes things load faster. When you get a problem, sometimes you get, "Well clear the cache on your phone or your browser, whatever it might be."
So at its server level we can cache things and at the browser level we can cache things. Now if we help the way we send information from our website to the browser, we can improve how the browser caches the website. There are plugins, and we talked about not having too many, but caching plugins that help create cache files, help improve the way that browsers cache things, and some of them even do image optimization. So I've put up the right size; I've put up what I thought was a pretty good size; then they have the capability to then even crush them down and make them really really perform better.
But there are other things that I wouldn't do, and I might not even be comfortable doing that. So you're going to go to your developer ...
Edmund: Right, so the web developer ...
Darryl: The web developer. And say, here's my report. Look at this; screenshot it, print it out, whatever you want to do. There's a lot of stuff there that looks not so flash. You know, like ... look at this one here, 98%, but there are some image dimensions on it that aren't specified; that's where, that's the image size. Now, it's still 99 out of 100, but that would be a little tweak. Am I going to get to that level? On this site, probably not at this point. But there are things that a web developer can do really quickly, including using a great caching plugin that will take care of a whole bunch of these things in one hit. Because they realize that the site owners really are not going to go down. They might have actually bought a theme, and they might be running a site. So they don't actually have an on-call web developer. So in that instance, a plugin like, and I'm thinking WP Fastest Cache is my one [crosstalk 00:19:23] at the moment; I really like what that does.
Edmund: Yeah ...
Darryl: You install it, and already, the next time you're on GTmetrix, you've gone up X%, because it's just taken care of a few things.
Edmund: But even that plugin, for some people it's going to be too much, right? I mean, do they go to a local, do they trust an online website where they can load a job, to say, "Hey, please speed up my site"? I mean, what is your recommendation, say, where do they find people to help them do this at an affordable price?
Darryl: Well look, you can use, there's some WP services out there that offer, you know, X much per month. You can go to Fiverr or Upwork, you could do stuff. You can go to a local web dev if ... I mean, you might have had someone help you do bits of it. If you've got an IT person that is web orientated that has a fair clue about stuff but they're just not that front-end design type of person, but they helped set it up, they can probably help do that for you and tick the boxes. There are great tutorials online about how to use these things as well. Like installing the plugin itself. So it's not necessarily dead easy. You can work your way through it. And a lot of these plugins have light version and premium version and all the rest of it.
Darryl: Some of the things it wouldn't do. I think like everyone, you should have some advisor, some helper, that helps you get the best out of your website, and ...
Edmund: You're right, and like you said, most people could install that WP Fastest Cache, and tick the boxes and let it run and optimize images and see what it does. I would like to suggest that as long as there's incremental improvement the next time you run GTmetrix, it loads faster, the file sizes are smaller, you're getting better over time. So that leads me to my next question, is, when do you know, good enough is good enough, right? I mean, at what point do you accept that this is the result that I've got to live with, or I should just keep pushing to get it faster and faster. How fast is good enough?
Edmund: So two seconds. Two seconds is pretty good; that's a nice, fast load time for ...
Darryl: And it's fully loaded. So the user's probably got a large chunk of what they're seeing within a second or so. So if you can't get under two seconds, you will generally find that the first byte time is probably a real issue. The response of your hosting is probably significant in that. And that's when, you know, people often bag shared hosting. I don't have the same problem with it. If I can get 0.7 seconds off a good site in a shared environment and my users are happy with the site performance, done deal. But if you are on a clogged network, if your servers that are underperforming, you know they oversell, that's likely to be the change. And, you know, they recently moved a few people around, some hosting environments, to places that have deliberate focus on their hosting environments for speed.
And they make significant differences. I know there was a site I worked with you on, that was four, five, six seconds, and most of it was in the response time. You know, we'd optimized a fair chunk of it; it's now in that two-second range, because there's still some plugins that, you know, we're working on, but that's a significant jump. Having that, from, that was purely the hosting environment that made that difference. And everything else stayed the same.
Edmund: That's it.
Darryl: Big change. So, hey if everything's running fine, you can get under two seconds, I'd think for a period of time you'd probably say that job's done, now focus on other things. But if you're not, the last thing I'd say on that is, but don't rest on your laurels, because all the stuff changes. You know, the environment that you're on upgrades. The things you're embedding change. The phones, the browsers are all changing weekly, month ... You need to check this as I said as sort of a monthly habit. Alright, so something hasn't gone wrong, and it's a piece of code broken on my site. Are we leaching out some third party thing and I found a site recently where they had a third party widget installed on their site for reviews and stuff like that?
Edmund: Mm-hmm (affirmative)-
Edmund: Okay. Look, we've covered a good lot of material. One thing I did just want to ask your final bit of advice on is that my SEO down the road, he's told me that I need to move my site to an amp and roll out a CDN. What's he talking about and should I even consider those things for my small business?
Darryl: And this goes down to, there are some limits to what you can do well. So if you're global, and you want local performance, so we talked in the beginning about testing locally. So the problem is, let's just say I'm hosting in Australia but I have U.K. customers. And no matter what I do here, I'm at 1.4 seconds. But if I test from the U.K., it's three seconds, or 2.8 or whatever. And I want that better; I want a great experience. So the native way that you would do that is to add a CDN. So CDN stands for "content delivery network," and there are free CDNs, and once you go over thresholds you pay.
What they do, is they effectively cache hosted versions of your site around the world on different server farms, and they intercept that request. So if a U.K. person is trying to go to your domain, they deliver from a U.K.-based server or European-based server, that copy of your site. Now, the CDN, all the technical stuff is there; they update regularly and keep your content and you can flash them and send out fresh content. They manage all of that. Very smart, lots of things you can do. So you will get that implemented with the help of a technical person, web hosting, whatever.
So that's one thing. Some hosting companies have plug-and-play buttons to do it. But there are a few things that can go wrong with that; you need to think about it. But you would use that on a more global scale. If you're being hosted in Melbourne, and you have Melbourne customers, you probably don't need a CDN.
Edmund: You don't need it. Alright, that's good.
Darryl: AMP stands for "accelerated mobile pages." And this was brought in by Google, effectively. And the way they looked at it is a little bit like the way they implemented https; they just kind of said, people just aren't making change fast enough. You know, there's not enough momentum to force a business owner to invest in a new website for mobile, or for this or for that.
And they are processing gazillion, squillion levels of data every day, trying to deliver. And what they are very good at is distributing information very quickly globally to people. So what AMP in a very top level sense is, is they have their own format of your pages that they basically host and run for you that are fed off your site, that are highly structured to run well on mobile. So they will run really fast.
Now, they improve desktop performance and everything as well, but they're designed that any user anywhere in the world, in theory, on a mobile's going to get a very fast response, because it works well for them. You go to Google, and you want some data; you click, you get it, you come back, you do whatever. That's improving usability. So at the top level, AMP is about that.
Now, implementing it is a whole different level, and probably another show in its own right. You are probably going to need help to implement it. Yes there are AMP plugins, but if you read the top level guys that do it, even they sometimes have to be very careful about how they implement it.
Edmund: I look at it from the perspective of, if Google is talking about site load speed and everyone else is talking about site load speed, then it's an issue. And that's why this whole episode, "Speed," is so important.
So if I was to summarize what we talked about today, site load speed is really important, not just from Google's impact, because yes they like to rank fast ones, but it has a big impact on our customers' user experience on the site. And ultimately that will translate into more business, right? They won't leave; they'll hang around; they'll hopefully buy from us.
You can use a free tool like GTmetrix just to make a quick assessment of the site. You could probably do some fixes yourself, but the reality is you've got data and information with this tool that you can give to a web developer who can help improve things.
Anything else Darryl? Anything else you think we should throw in there, any resources or guides or takeaways you can think of?
Darryl: No, I think by the time we've linked to a few of these, I mean I think the thing there is, we've talked about the prioritization of it, but this is about empowering you to ask the right questions of your provider. So you can go to your provider and say "I want you to fix this for me." Now, if a provider can't fix it, and you know, a month in, nothing's changed, well then that helps you. It's like with the other advice we're giving you; if you're not getting the right advise, this should empower you to understand these things more. They matter to your online performance; they should matter to you as the business owner. Get them fixed, move on to the next thing that we'll be talking about.
But no, I think you've summed it up really well. That's good enough. We're going to link to GTmetrix's page speed in the show notes, so we'll do all that.
Edmund: Excellent. We'll put all those-
Darryl: Make your site faster. And check it. The key thing here is, check your site yourself from different locations without the benefit of your high speed internet in your office. Do it out on the road; do it when you're at the beach. Do it without wifi; do it on 4G, and just check, how does my site work on my mobile phone?
Edmund: Yeah. And if you hate it then your customers are going to absolutely hate it.
Darryl: Fix it!
Edmund: Yeah. I think that is an awesome roundup for today's episode. If you enjoyed the content that you're hearing today, once again, please leave us a review on iTunes. If you want to be notified when the next episode's out, subscribe on our website ...
Darryl: Review on Facebook, you want to do that.
Edmund: On Facebook, yeah, we'll take reviews from anywhere we can. That's awesome. Thanks so much. I'll look forward to chatting to you next week when we continue this discussion about My Bloody Website. It's goodbye from me ...
Darryl: And it's goodbye from him.