Latest / Inside Commerce: Ecommerce Strategy, CX and Technology Podcast / EP220: The Page Speed & Performance Conundrum: Balancing CX with Speed Gains for Ecommerce Sites
Transcript
- 0:02You're listening to the Repath and podcast sponsored by
- 0:04Patchworks, hosted by me James Gettin, Paul Rogers.
- 0:07We bring you interviews as industry thought leaders, see
- 0:10sweeps of tech companies, leading retail brands and our
- 0:13aim is to help you make better decisions through business and
- 0:15discover new ways of doing things and new technology.
- 0:17We also return to you tell other people about podcast, share our
- 0:20posts on social media and if you need to show, give us a rating
- 0:23wherever you're listening. So thanks to let's me set up our
- 0:25latest episode. So cracking topics, a page speed
- 0:29and performance conundrum balancing customer experience
- 0:31with speed gain. So this was inspired by an
- 0:34article from our guest Liam will introduce himself in a minute on
- 0:37LinkedIn in which he featured not only his own thought
- 0:40leadership and expertise on this but notable luminaries for our
- 0:43industry such as Yoda, render beta and our previous podcast
- 0:46guest and respected consultant Harry Hoffman all specialist in
- 0:49different areas of of site optimization performance and and
- 0:52the key maxim is this right, The less you put on the site the
- 0:54faster it is. But any commerce, we're not just
- 0:56trying to make pages fast, we also want good brand experience,
- 1:01good customer experience. And actually those latter 2 are
- 1:04sometimes more important than actually pure speed itself.
- 1:08And also the key maximum here today is that measured speed in
- 1:11terms of what we can look at in metrics and tools is different
- 1:13to perceived speed of a user. And we we've got to balance that
- 1:17and not just obsess about speed. However, brands also need to be
- 1:21aware that that crap forward pages impact customer
- 1:23experience, provide bad brand experience as well.
- 1:26And we've got to make sure we balance the technical with that
- 1:28brand. And I know Paul and Liam will
- 1:30talk more about it. There's loads of data out there.
- 1:33Deloitte reports shows that the brand the travel ban too.
- 1:37We saw a 31% decrease in bounce rates, 78% reduction load times
- 1:41with better tech optimization, eBay measured or .1 second
- 1:46improvement in low time resulting in a .5 increase in
- 1:49add to car. Pick your start out there.
- 1:51There's lots of them to show that this actually makes sense.
- 1:54So with that as our set up and Liam, do you want to give people
- 1:58an introduction to you, people who've not come across you
- 2:00before? Yeah, perfect.
- 2:02So thanks for having me on. That's all.
- 2:05So yeah, I'm Liam, I'm Technical Director at Milan, so I've
- 2:08worked alongside Paul for four years.
- 2:13Previously I ran like digital, the digital side of Shopify Plus
- 2:18agency and then further back, yeah, just development
- 2:23essentially for a few different agencies.
- 2:25So yeah, my backgrounds like computer science development,
- 2:29that sort of technical solution side.
- 2:32Excellent. So let's get into this.
- 2:34And we're going to, for those listening, we are going to be
- 2:36balancing the technical piece and also with that balancing
- 2:38with the customer experience piece.
- 2:40And so it's not just going to be purely on metrics, don't worry.
- 2:42But let's start with the commercial bit.
- 2:44What commercial impact does good performance have?
- 2:46Why, for those who aren't so experiencing this part of it,
- 2:49why do e-commerce teams need to take this seriously?
- 2:54Yeah. I mean I think you've just
- 2:55touched on some of it as well like this tons of data out
- 2:58there. There are studies between the
- 3:01conversion rate increase with the improvement of page speed
- 3:07metrics. I think yeah I'll take the exact
- 3:09measurements with with a pinch of salt, but essentially we
- 3:14there is the correlation between those and we've seen the
- 3:16correlation. So I think from a commercial
- 3:20point of view especially sort of you know higher volume or
- 3:25something the IT makes perfect sense to try and at least to a
- 3:30certain degree have this page, the page be to a good point.
- 3:34I think like it. Yeah, the worst is the more
- 3:39impactful is there's like a threshold.
- 3:42But yeah from a conversion rate point of view it makes perfect
- 3:46sense. And I think the, so you
- 3:50mentioned like Yotta, so we've looked just prior to this.
- 3:53So you've mentioned like Yotta one of those tools that maybe
- 3:57we'll come onto through this, but the the way that they track,
- 4:00so they track monetary value against the improvements that
- 4:04they've made in the bin page speed by only optimising.
- 4:08So they're optimising optimisations will be rolled out
- 4:11for like 90% of users. So they can measure the 10% of
- 4:15users against the 90% and see put monetary value on the
- 4:19conversion rate between them. So yeah, we can you can measure
- 4:23that like there is an impact on like a commercial impact between
- 4:29the two. So yeah, I mean that that's
- 4:34generally the reason behind it all.
- 4:35And then Paul, we were talking Paul beforehand on guest
- 4:41commercial impact for brands. So we we work with a lot of
- 4:44premium and luxury fashion brands where actually the value
- 4:47is in the reflection of the brand, I guess in in come on.
- 4:52To that one, you're jumping ahead.
- 4:57Commercial wise, yeah, so. One question, I had one
- 5:02interesting sub question or I think it's interesting.
- 5:07So back in the day like maybe five years ago, there were so
- 5:11many studies around the impact of performance and revenue and
- 5:16there was an incremental positive or negative impact as a
- 5:20result of seconds in load time and back it.
- 5:24But back in the day, the average performance was much worse.
- 5:29And I think almost every platform and definitely the SAS
- 5:32platforms have a much better starting point.
- 5:34So I guess one thing that I was going to ask you is what's your
- 5:37view on like do you think you need the same level?
- 5:41Because we've always talked to clients about discipline from
- 5:44like a content perspective, from a release perspective, from a
- 5:46test perspective. But like, do you think it's as
- 5:49important nowadays as it was years ago when the starting
- 5:53point, even in Shopify world, was pretty terrible?
- 5:56Yeah, I think it's definitely got, yeah, it's gone the right
- 6:00way. And I think the, you know when
- 6:02you're talking about the difference between a second, you
- 6:06know between bad performance and good performance is yeah the
- 6:10margins a lot smaller than five years ago.
- 6:13However, I think the more and more we're using like heavier
- 6:17assets. So like page where is still
- 6:21going up, the amount of assets we're putting on the page, the
- 6:23amount of video that's like being streamed or loaded on the
- 6:26page, all that sort of thing keeps that going up.
- 6:28And I think the only thing to add to that is like in the so
- 6:32from our void view, it's improved.
- 6:35So like now we're all on 5G, we're all on like super fast
- 6:40Wi-Fi. But a lot of the brands are like
- 6:42we're working with serving in other territories in the world
- 6:46where actually the Internet connection is not that strong
- 6:49and they're still potentially having poorer page loads.
- 6:54But but yeah, overall I definitely agree that the the
- 7:01level of importance maybe chopped to it.
- 7:04Yeah, yeah, I think that makes sense.
- 7:06I think it is. There's been a lot of obviously
- 7:08all of the platforms have really doubled down on I think the
- 7:11average like front end toolkit has improved a lot.
- 7:16Next question, so you were about to talk about this and this is a
- 7:19bit that I find really interesting.
- 7:22So I think I'm usually, and James, I'm sure you've got the
- 7:26same experience with different agencies and even different
- 7:31developers. You tend to get different levels
- 7:34of of obsession and prioritisation around
- 7:37performance. So we've definitely worked with
- 7:39some agencies in the past that were so focused on performance,
- 7:43it would every time you bring up a third party, it was like a
- 7:46massive debate to try and get it implemented even if that third
- 7:50party offered significant value somewhere else.
- 7:54And I think you end up in some kind of balance that when it
- 7:57comes to like the overall kind of experience.
- 8:02And I guess like you know you talk about like video and you
- 8:05know really high res imagery and then even the third party
- 8:10everything else. Like where's the like, what do
- 8:12you think on kind of where the balance is?
- 8:15Yeah. I think I'm probably further
- 8:17down that scale than you in terms of like I would
- 8:23historically be on that side of like adding all these tools are
- 8:27going to you know increase different things and like let's
- 8:29not just drop widgets in all over the place, that's to like
- 8:32API integrations for different fittings.
- 8:35And I still think to a degree that's the right approach.
- 8:39But yeah, because of what we previously said, I think just
- 8:44the significance of it has probably reduced.
- 8:48But and then I think really it's context of what the, you know,
- 8:54what who the merchant is, what the priorities that the merchant
- 8:57are and just being pragmatic about it all.
- 9:00I don't think there's like A1 size fits all and that for any
- 9:03brand. Let's let's make sure the
- 9:07performance is this extreme. But yeah, I do think, I mean the
- 9:12whole this whole conversation is going to be about the sort of
- 9:14compromise and balancing act between that's still being
- 9:17significant but also especially in the Commonwealth like the
- 9:21brand and experience side, you know, balances.
- 9:26That, so that takes me on to a sub point.
- 9:30Yeah, I personally think like more of a priority.
- 9:36Again we work with quite premium brands kind of build this like
- 9:39brand experience and introduce like you know branded content
- 9:43and they invest a lot of money on art direction and everything
- 9:46else. I personally think the rendering
- 9:49part is as important, if not more important than load times.
- 9:54And I think then like the perceived performance and you
- 9:57know, everything, feeling very consistent and premium, They're
- 10:01like you look at like an on running.
- 10:02I put them in like a really high bracket for this kind of stuff.
- 10:06How do people like achieve that and what's your view on it?
- 10:11Yeah. I mean and we'll probably get
- 10:14more into you know different without getting to drive more
- 10:17into different metrics and things.
- 10:18But like a lot of metrics or measurements for page load will
- 10:24be like the full end to end. So from the initial point of the
- 10:27page starting to load until when it's like technically fully
- 10:30complete where from a user point of view actually the last 50% of
- 10:34that you don't see, you don't notice.
- 10:36It's just behind the scenes things are still loading.
- 10:40So a lot of, a lot of the importance is like prioritizing,
- 10:46prioritizing assets been loaded onto a page, the order of things
- 10:50loading and then again we'll get into like Google Co web vitals
- 10:56and things. But some of the metrics that
- 10:59they bring in are more focused on the users interaction or the
- 11:02users like perceived experience on the site.
- 11:07So like CLS where you essentially don't want to, it's
- 11:13like pushing for things not to be shifted around on page, let's
- 11:16say prioritizing the order of those assets been loaded.
- 11:20Patchworks the world's most modern integration platform.
- 11:23We can connect any key e-commerce system from
- 11:25storefronts to marketplaces and ER, PS to fulfilment solutions.
- 11:28This allows retailers and partners alike to simplify their
- 11:31tech stack integration and do away with point to point vendor
- 11:33specific apps. We can automate and streamline
- 11:35the flow of data across your entire e-commerce business.
- 11:38Find out more at We Are patchworks.com.
- 11:40I I think this the the point you are making about about
- 11:44obsessively focus on page speed and not thinking about that that
- 11:48compromise. But it's so important because
- 11:50I've I've got this I've worked with where they have a they have
- 11:53a development team that's built our platform of an open source
- 11:58and they have a have a. We won't use external plug
- 12:02insurance not not 100% rigid but very much we want to build
- 12:07because we can control it better.
- 12:08But actually a good example is wanting to use like a a a AI
- 12:15search tool, but build out a search capability from scratch
- 12:18versus using the market leader, which there are times when using
- 12:22a third party. The capability it delivers and
- 12:25the scale and speed at which that capability is going to
- 12:27increase over time means building it yourself as bonkers,
- 12:31the cost and effort and maintenance and therefore you
- 12:33have to have that little trade off.
- 12:34So I think it's really important that people think about what the
- 12:37end goal is. Speed.
- 12:38Speed needs to be factored in to then you work with that third
- 12:41party and the integration like you're saying APIs, avoid doing
- 12:45lots of JavaScript grid stuff if you can't, yeah, it's
- 12:48interesting to see those dilemmas.
- 12:50Let's let's move on to SEO. So there are several SEO studies
- 12:55that have shown small correlation between rank and
- 12:57page speed. So not massive like, OK, get
- 13:00your page speed sorted. Your rank's going to fly up
- 13:02because we know that it's far more complicated in terms of how
- 13:05such changes like Google rank. But I saw a study from Google a
- 13:08delight stating every 1.1 seconds your page speed improved
- 13:11can increase conversion rates by 8%.
- 13:13I've never seen that personally. Now I'm caveat, I don't work
- 13:17across all different types of sizes, all different scales.
- 13:20So I'd love to know from the from from the basis as you work
- 13:24with work you've done, what impact have you seen between SEO
- 13:27and conversion from a speed point of view?
- 13:29Have you seen it? Have you seen it?
- 13:31Have. Impact, Yeah.
- 13:34So from my side, I think those two, I probably got two opinions
- 13:39on those two parts. On the conversion rate increase,
- 13:44like I've definitely seen and still see the correlation and
- 13:52that effect. So I think even before this, I
- 13:55mean not .1 seconds to improve conversion rates by 8%, yeah,
- 14:00pretty extreme. But I've definitely seen say
- 14:04like yachter or whatever when I've checked is like for like so
- 14:10users that are optimized against users are not sort of like 10 to
- 14:1315% conversion rate increase on the one that I've just glanced
- 14:17at. And we'd see that commonly.
- 14:19Like yeah, when you're measuring those sorts of things, I think
- 14:24SEO Paul you might have like your backgrounds probably more
- 14:27SEO folks than mine. But from a yes, the SEO point of
- 14:30view I think like Google's strategy from years ago was
- 14:35always to bring that in as like a major ranking factor.
- 14:38And then it felt like there was a point where actually so many
- 14:42sites were feeling their Co web vials that it never properly
- 14:45became a ranking factor and more of like a sort of warning or
- 14:49threat that at some point this will have impact you.
- 14:51And I've never, from an SEO point of view, I've never seen
- 14:56like a huge like noticeable impact like correlation between
- 15:03page speed and and actual organic rankings.
- 15:06And the one I've seen is like, yeah, I guess failing Coed
- 15:10vitals. It's more of like you're
- 15:12penalized potentially in your rankings rather than getting
- 15:15rewarded for having good page speed.
- 15:17But yeah, Paul, I don't know if you've seen anything more
- 15:19concrete, but. I do think this did have a very
- 15:24negative opinion on sorry negative impact on development.
- 15:30So this whole thing because 1-2 major impacts one, every SEO
- 15:35consultant will send really generic performance
- 15:38recommendation and documentation to developers when it's like no
- 15:42way relevant to the tech stack and like everyone's got this
- 15:45like crazy opinion around performance.
- 15:48And two every time rankings drop or you know visibility drops or
- 15:52organic traffic drops, SE OS will a lot of the time will
- 15:55blame performance and it's like this blanket kind of almost like
- 15:58excuse which is very annoying. We've, I guess we've monitored a
- 16:02lot of our client performance so we've got like a random tool
- 16:05which is very similar to one that we used to pay for.
- 16:08I forgot what it's called the basically just take the core web
- 16:12vitals metrics by day per client.
- 16:16And I've personally since we've been doing that and since we had
- 16:19the other tool I've forgotten what it's called.
- 16:21I've never seen like a noticeable impact.
- 16:23I've never really seen anyone prove the impact of like
- 16:28improvement or reduction in performance to organic visit.
- 16:32I've never seen like a close like correlation and to me I.
- 16:37Think the SEO one's higher to measure as well than compared
- 16:41with it and set up some tests. But we've SEO like there's so
- 16:44much, there's so many variables. Exactly.
- 16:48Just give somebody an out to say it's just slow.
- 16:51But I actually think the as a ranking factor, it's pretty
- 16:56small. And yeah, I've not seen like,
- 17:02yeah, I've not seen major correlation either way.
- 17:05The only good thing I do think that, oh, sorry, Jane.
- 17:10I was just going to say the only good thing, I think Core Web
- 17:12Vile did make a lot of more. They basically made best
- 17:17practices a lot more prominent and important.
- 17:20And that was I think that was one of the things that judged a
- 17:23lot of the platforms as well because they don't want friction
- 17:25with clients as a result of like their base theming and
- 17:29everything failing core environment and development
- 17:32teams the same. So I do think it's helped a lot.
- 17:35Go on, Jed. Yeah, no, I I I 100% agree.
- 17:38I think, I think they're, they're a really solid set of
- 17:41principles, but they're not absolutes because you can
- 17:45actually have an effective page that might not hit green on all
- 17:49of those core web vitals metrics, which we talk about in
- 17:51a minute. I've seen sites actually from an
- 17:53overall speed that instant rendering the critical path
- 17:57rendering is really good. But maybe they've got a few
- 18:00issues in terms of like largest content paid sample, which is
- 18:03obviously not ideal when you can do stuff but it's not it's not
- 18:06like the site has failed just because that's not under the 2.5
- 18:09second threshold. So I think you need to you need
- 18:12to use them as guidelines rather than absolute.
- 18:15And yeah, I'm I wanted to answer that question because there are
- 18:18a lot of incredibly smart SCOS and technical SEOS out there
- 18:21that we all know who do brilliant work on helping
- 18:24businesses make better decisions architecturally and who
- 18:27understand that balance and nuance.
- 18:29But I've I'm still yet to hear that definitive.
- 18:32It's really fundamentally easy to measure because you're right,
- 18:34Lynn. There are so many different
- 18:36factors that can can influence rank.
- 18:39Isolating it to speed is very difficult.
- 18:41How you control that. I've yet seen methodology where
- 18:44maybe there is one. I just haven't read enough about
- 18:46it. So there you go challenge the
- 18:48technical I see is maybe there's stuff we we need to know about
- 18:50that we're not aware of. Let's let's go tackle the the
- 18:54elephant in the room. I love this one because we know
- 18:56the industry's been banging on about headless composed all in
- 19:01many respects for sensible reasons for the last few years.
- 19:04But I've seen enough information people say.
- 19:07I've heard people being told as inclined being told you need to
- 19:11go headless to get a fast website, which is just not true.
- 19:14And you still be slow on headless if you don't get the
- 19:17right setup or you use the wrong third parties or etcetera,
- 19:20etcetera. What role do you, But let's look
- 19:23at the positives, What role do you think headless can play in
- 19:26boosting performance? What have you seen?
- 19:30Yeah, I think so as well, just to go back and plug the the blog
- 19:35that you sort of referenced at the beginning that that all came
- 19:37out of basically the fact of you can achieve or like 99% of
- 19:44brands can achieve a really strong performance on Shopify,
- 19:47on Shopify with a sort of theme build natively with the right
- 19:52work put into it. So that was sort of the premise
- 19:55of the blog really because this question comes up quite a lot.
- 20:02And then from a headless point of view, my answer is sort of
- 20:07always the same that is. So I'm like a fan of headless in
- 20:11the right circumstances. I think there's a lot of
- 20:12benefits. I don't think the reason to go
- 20:16headless is never purely just to get a faster site speed.
- 20:21I think it, you know, it has to be for the other concrete
- 20:24reasons and that's, you know, that's a factor or that's a
- 20:28benefit of it. But it's not the core reason to
- 20:30go headless. It's sort of an easy way out to
- 20:34say, you know, we we can't achieve a good experience so
- 20:41we'll do this massive change of tack to build headless and still
- 20:47as you said, there's no guarantee.
- 20:48So I think, yeah, my answer is that a headless opens more
- 20:55opportunity to get a better page speed potentially than a liquid
- 21:00theme, but definitely don't guarantee anything.
- 21:02And actually I've I've seen or like we've probably all seen
- 21:06headless builds that end up being slower than what could
- 21:09probably have been achieved on a liquid theme.
- 21:12I've definitely seen headless builds that have ended up a lot
- 21:15faster than could be achieved as well, just to just to balance
- 21:18that out. But yeah, there's definitely no
- 21:20guarantees. Like still needs to be done
- 21:22right. Still need people who know what
- 21:24they're doing from the performance side, yeah.
- 21:30So next question, we kind of touched on this already, but in
- 21:34terms of ecom development teams and then also people managing
- 21:38content and kind of you know, trading the site and whatever
- 21:41else like beyond content and you know selecting the wrong third
- 21:46parties or third parties not building their widgets and you
- 21:50know Java and writing the JavaScript well in an optimised
- 21:54way, like where else do people go wrong with performance?
- 21:58I think commonly so, especially when it's noticeably slow.
- 22:04So I just to go back to content as well.
- 22:07I think that it is probably 50% of you know and all that and the
- 22:12sort of first steps will be reviewing that like imagery,
- 22:16videos, using the right file types, having imagery at the
- 22:21right dimensions and things rather than you know keep 3000
- 22:26by 3000 raw photography images, having a process like the match
- 22:33and having a process in place to actually get those assets, you
- 22:37know, right. Optimize for web I will.
- 22:39And to keep into that process. So I think that is probably 50%
- 22:45of like the first steps And then yeah after that it tends to be
- 22:50like so lazy loading of assets and that side for the rendering,
- 22:57the use of the overuse of Javascripts or bringing in you
- 23:01know JavaScript frameworks or inheriting JavaScript libraries
- 23:06that end up bloating the code massively.
- 23:11Like all of those things are sort of the main contributions.
- 23:14I'd say like there's it's probably not nothing
- 23:19groundbreaking. It's like they're the they are
- 23:21the big factor still that that that impacts these things and
- 23:27then like to a smaller level so like you can get.
- 23:31I've seen examples where it's like innocent developer mistakes
- 23:36at some point through the liquid like can lead to like massively
- 23:42underperforming sites in the Shopify team.
- 23:45Just like looping badly through a set of products and things
- 23:48like that, where actually it can go from a from a front end point
- 23:51of view it goes under the radar because it all looks like it's
- 23:54perfect. But actually, behind the scenes
- 23:55it's just like inefficient coding or inefficient use of
- 23:59loops and doing the wrong things can lead to like drastically
- 24:04longer page loads or sort of just ensuring those things are
- 24:08not happening, yeah. And so one other thing that I
- 24:14know we've commented on in the past is kind of like loading
- 24:19particularly third party JavaScript, but also just, you
- 24:22know, scripts in general in the wrong order so you'll get, you
- 24:25know, critical parts of the page loading later on in the journey.
- 24:29So actually like building the full page experience and loading
- 24:32all of the content takes longer than it needs to and then
- 24:36anything else could feasibly load after.
- 24:38It's like, what's your view on that?
- 24:40Yeah, I think that and this one goes back to like the perceived
- 24:42performance part. So the you know the full end to
- 24:45end page load is still the same but from a dev point of view
- 24:51using like lazy loading then you can defer or like async the
- 24:55right JavaScript and and basically making sure those
- 25:00files and assets are loaded in a certain order where from the
- 25:03user point of view it feels faster.
- 25:05The you know the critical path of of the load and the critical
- 25:11parts of the page that the user are going to see first.
- 25:15A loaded verse that feels instantaneous.
- 25:18I think actually like aside from that list of the sort of bigger
- 25:23things of optimising and lazy loading things that makes a
- 25:27bigger difference from a user point of view.
- 25:29So like metrics wise you probably won't see differences,
- 25:33but from a user experience point of view that's probably the bit
- 25:36that has like biggest impact putting some consideration and
- 25:39time into the the order of loading these things.
- 25:44That makes sense and I guess which and I guess I should know
- 25:49this, but I actually don't but in totality of reviewing
- 25:53performance and impacts and ongoing benchmarking, but but
- 25:59aside from the tool that I mentioned earlier like what
- 26:00tools do you use to assess this and the different areas and
- 26:04impacts? Yeah, you should know this, but
- 26:06it's bad from you. Yes I always start with like I
- 26:12always start with Search Console basically so that so from search
- 26:16console you'll see like the full list of index pages and the
- 26:22state of the whole site really rather than just running the
- 26:25home page and a product page through any of these tools.
- 26:29It's like the full state of performance across the site and
- 26:35start to see like how many of the pages you know from a Co web
- 26:39vitals point of view passing fill in like they hygiene across
- 26:43the site. And then I mean there's a
- 26:47massive amount of tools that have a big a big amount of
- 26:49crossover between them. So it does come to like personal
- 26:52preference at other times, but like web page, dev team metrics,
- 26:57Google, Lighthouse are all options that do the same sort of
- 27:01thing which is like lab testing, homepage, product page, PLP and
- 27:09getting the metrics from those. And then yeah, I think these are
- 27:12the tools I mentioned in that blog again that are like really
- 27:15user metrics. So from that point of view, I
- 27:17think if everything's indexed in Search Console well like a
- 27:21metrics are like reasonable, It's drop into like real user
- 27:25metrics and trying to improve like what your users are
- 27:28actually experiencing. So yeah, as a range of tools,
- 27:32that's probably the the order of running through things.
- 27:37If you're just a related question, so if somebody's
- 27:40starting out and they've not got a deep level expertise and they
- 27:43don't want to spend loads of money with with somebody else
- 27:45paying to do this, what what? What's the number one thing they
- 27:48should do? Because I've always loved saying
- 27:50to peoplethatwebpagetest.org is great for just some indicative
- 27:54stuff going into. If you look in your search
- 27:56console, Google and then go to Google's Page Speed Insights are
- 28:00there. Are there any others that are
- 28:01low cost free that someone can self-serve on without having to
- 28:04be really technical? Well, I think those as a in that
- 28:09context you've just said like you want to get to a point where
- 28:12everything's at least like healthy.
- 28:14Like I know that there's nothing major going on.
- 28:16Those I think are a good starting point.
- 28:19So Search Console making sure everything's passing with my web
- 28:23page test and Lighthouse. I think what I also and what we
- 28:27like we do with our clients is tracking those.
- 28:30So like we'll set something up that tracks those metrics daily
- 28:35and just sits behind the scenes and we get the sort of a daily
- 28:39insight on those metrics. So the reason for that is not to
- 28:44look at it dearly and actually point as soon as there's a drop
- 28:47or anything, it's more like just to glamorise it slightly, it's
- 28:52more like a sort of stock market viewer.
- 28:54This is the level we're at and the the actual metric or the
- 28:58actual value day-to-day doesn't really matter.
- 29:01It's like the trend over the week, over the month.
- 29:04Am I trending in the right direction or trending downwards?
- 29:07There's always going to be fluctuation, always going to be
- 29:09like a spike because Cloudflare has gone down or a spike
- 29:13because, you know, we've got high traffic that day.
- 29:15And it's not about like pointing at that one thing and sort of
- 29:18panicking about it. It's just getting that like
- 29:20measured view, be able to see what you know, what trends
- 29:24you're going in. So I think that, yeah, they're
- 29:27for a yeah, I think that would be the starting point for me.
- 29:33Yeah I like the fact that you're that you're you're doing those
- 29:35reports because I I would advise that too.
- 29:36And and also yeah, the not the knee jet reaction because you
- 29:40can track, you could have a chart which trends the speed,
- 29:43the average social speed and then you could have conversion
- 29:45rate and the conversion rate could suddenly go up or down and
- 29:47then panic because it might you might see the speed change but
- 29:50then you realise that somebody's just per 30% off.
- 29:54Yeah promotion code or that's why you commercially despite or
- 29:57as you say something else has happened, has dropped it.
- 30:00So, yeah, so. I think there's no value in.
- 30:04I think there's no value, and I've had it like historically as
- 30:07a developer where a merchant or whoever will be fastly running
- 30:12through running page through all of these tools.
- 30:15And if one of them gives a red, they're all going to be slightly
- 30:19different. If one of them gives a red, like
- 30:21send that across because it's not superb on them all and then
- 30:26doing that every day and then suddenly there is a bit of a
- 30:28fluctuation. It's like, yeah, it needs to be
- 30:32more of a sort of longer term like trending view.
- 30:37Yeah, exactly. I think that's really good
- 30:38advice. If you suddenly see it drop down
- 30:40and stay down and and that is persistent, then you know
- 30:43something's fundamentally shifted rather than a short term
- 30:46tactical promotional site. Actually, yeah, let's talk about
- 30:50fonts. An area that I've looked into
- 30:52but I'm not going to profess expertise in and the do's and
- 30:55don'ts because there are there are certain like maxims out
- 30:58there about OK you need to preload fonts, use web friendly
- 31:00fonts and actually there's a far wider range of fonts that can be
- 31:04used. But fonts can in in impact some
- 31:07of the core web vitals like cumulative layout shift because
- 31:10if you have a a full back font whilst your brand on load you
- 31:14can suddenly get that flicker an adjustment.
- 31:16So what, what? What, from your point of view,
- 31:19do you, when you're advising clients?
- 31:20Like what? What's good look like for font
- 31:22management and what can it do if you get it wrong?
- 31:26I think fonts is one that's like and historically as well has
- 31:30always gone under the Red Arrows.
- 31:31Actually it one of after the some of the bits we've talked
- 31:35about it's one of the more impactful things sometimes if
- 31:38done badly I'd sort of I guess nobody really realises it but
- 31:44firstly even in font and this comes back to the brand, the
- 31:48brand types even selection of font and that font's hosted.
- 31:51So you know if you're using a third party or like Typekit used
- 31:55to be the one that's that's huge page load addition to get the
- 32:01one specific font that you're using.
- 32:03So actually how the fonts are hosted, if it's web hosted or if
- 32:07you've got the font files locally loading on the site is
- 32:12the ideal. And then even things like.
- 32:16So by default you might be loaded if a font can come in,
- 32:19like font family might have 12 different weights and styles
- 32:23that all get loaded by default. And it actually does again does
- 32:26make a big impact where just loading the specific fonts that
- 32:29you're using across the style sheet, it's the best thing to
- 32:33do. And then yeah, like the font
- 32:36swap is just one small CSS point that can help fight that bit
- 32:43that you've just talked about which is one point loading then
- 32:46flick into another one. So yeah, mix of basically going
- 32:50through all these steps. But yeah, fonts is 1 to look at.
- 32:55Thanks. And my, my, my next, actually my
- 32:58final question for today is around pre fetch and pre render
- 33:00because this has been used massively in things like I guess
- 33:04content sites, travel, etcetera. There's certain industries where
- 33:07they've really doubled down on using these.
- 33:11From a pure e-commerce point of view and looking at the sort of
- 33:13brands you work with, how much, how extensively is pre fetch and
- 33:17pre render being used and is it? Is it a critical thing that the
- 33:21e-commerce teams should be having discussions about, or is
- 33:24this a it's useful to look at once you've got the basics and
- 33:27the fundamentals sort? Yeah, I think probably that.
- 33:30So I think we've you know covered a lots of other little
- 33:33bits and and things like defer and async are the sort of
- 33:37stepping stone towards say like free rendering.
- 33:42I think that almost gets to the point where you know when we're
- 33:47talking about headless it's it's more custom and sort of a bigger
- 33:52amount of effort for. Yeah, it's more going away from
- 33:59like the native behaviour of the Shopify theme which I think is
- 34:03probably what between me and Paul we tend to advise.
- 34:07It's like we want to be, we want to be best in class and get all
- 34:10the best practices of a lot of this bachelor stuff in.
- 34:13But without moving away from sort of native Shopify
- 34:18functionality and behaviour and more towards like that goes a
- 34:22step towards headless or you know JavaScript frameworks being
- 34:26used rather than liquid where actually it's yeah potentially
- 34:32over complicating and moving away from what's Shopify is sort
- 34:35of handling well. And and the the point that we've
- 34:39talked about internally actually is obviously and this is going
- 34:42off topic a bit, but Shopify is sort of moving I think towards
- 34:47like no code, no code set up sort of trying to get more areas
- 34:52consistent. So like check out extensions and
- 34:54also that sort of things like to get everybody on a consistent
- 34:57code base and actually things like pre rendering, pre
- 35:01rendering when when Shopify have more consistency across all the
- 35:05stores, it's probably that sort of thing that they'll bring in
- 35:08and handle and handle natively as well.
- 35:13So yeah, my advice typically would be to sort of stick
- 35:18towards the more native Shopify way of doing things in that
- 35:21context to know what you think. Paul's.
- 35:26Yeah, I think we have a similar view and I think this will that
- 35:29stuff will get more important over time.
- 35:32Yeah. And I think Liam said for a long
- 35:34time that the move towards sections was moved towards low
- 35:38code. And I feel like the next step
- 35:41for the smaller brand will be more no code.
- 35:46So yeah, And then at that point have we managed to platform
- 35:49level, application level. So it'd be interesting to see
- 35:52where that ends up and that could be an episode in itself.
- 35:55To be first, last question, which metrics do you think
- 36:01internal teams should be looking at and why?
- 36:06Yeah, I mean the ones I and you could pick out from a range of
- 36:10metrics, I would stick to Google's called web vials.
- 36:14Like they've got massive teams of people that have picked those
- 36:18metrics because it's like a broad range of perceived
- 36:24performance. It's like a, it sort of covers a
- 36:28bit of everything in the grand scheme of page loads.
- 36:32So I think really the Coed vitals, especially from like a
- 36:38merchant's point of view have not been super experience, but I
- 36:41want to make sure it's a good point.
- 36:43I think tracking those plus like time to 1st buy or time to
- 36:47interact to are probably the two that I think are most like
- 36:49noticeable from a user and beneficial to improve.
- 36:55So they would be my yeah to to try and shortlist it.
- 36:58I would stick around those metrics.
- 37:02Cool. I've really enjoyed this
- 37:05episode. I really like the ultimate.
- 37:06I recommend anyone go and have a look at.
- 37:08We'll put a link on the podcast landing page.
- 37:10Have a read through because it's not just Liam's perspective and
- 37:13knowledge and expertise and that but you've got other people in
- 37:15there as well who are experts in this field and sharing their
- 37:18insights and what they see working.
- 37:20So lots of value in in looking at and do do reach out if you
- 37:22want to probe further and get get more technical.
- 37:25Liam loves that, definitely loves getting this technical
- 37:28weed. So do reach out and ask him
- 37:29questions. And thanks for listening
- 37:32everybody, We hope you enjoyed it.
- 37:34Keep the ear open for our next episode we drop on every week,
- 37:36and don't forget to give us that rating if you haven't already.
- 37:38Cheers everyone until next time. For more information on this
- 37:49topic, head over to Replatform dot FM for our audio podcasts to
- 37:53discuss a project, or if you'd like to chat about any of the
- 37:56topics covered in this episode in more detail, please reach out
- 37:58to myself, James Gurd, and my Co host Paul Rogers via LinkedIn
- 38:01and Twitter. Thanks again for listening and
- 38:03keep your ears peeled for the next episode.