Latest / Inside Commerce: Ecommerce Strategy, CX and Technology Podcast / EP267: Re-engineering Pangaia’s Shopify Storefront to Increase Operational Efficiency, with Commerce UI CTO Michal Klim
Transcript
- 0:02You're listening to the Insidecommerce podcast, hosted
- 0:05by experienced consultants James Gerd and Paul Rogers.
- 0:08Insidecommerce is your fast track to digital knowledge
- 0:10feature in interviews with leading e-commerce retailers,
- 0:13industry thought leads and the C-Suite of key tech companies to
- 0:16explore the latest in digital strategy, customer experience
- 0:18and tech. Discover what the latest trends
- 0:21really mean and get practical advice based on experience to
- 0:24help fine tune your ECOM strategy and tech stack.
- 0:27Follow us on LinkedIn and reach out to James and Paul if you'd
- 0:30like to discuss an e-commerce project.
- 0:32We hope you enjoy the latest episode.
- 0:35Today's episode is. Sponsored by Centre and grabbing
- 0:37more about them later. Let me set the scene before I
- 0:39introduce you to our wonderful guest today.
- 0:42So we talk all about technology and re platform and platform
- 0:45selection, but we're coming this from a different angle today
- 0:48because sometimes the platform isn't the problem that the
- 0:51business need to solve. A platform can be fit for
- 0:54purpose but it just hasn't been implemented or configured
- 0:56correctly to deliver against the operational needs of the
- 0:59business processes that the business wants.
- 1:01All the businesses processes have evolved and the platform
- 1:03hasn't caught up technically. And RE platform involves a lot
- 1:06of effort and complexity and brings risk to the business, not
- 1:09to mention a big chunk of CapEx cost.
- 1:12Sometimes RE engineering what you've currently got is an
- 1:15actually better solution for the business, whether that's short,
- 1:19middle or long term. So our guest today is highly
- 1:23experienced, the CTO of a leading development agency
- 1:27called Commerce UI and was speaking to him recently about a
- 1:32project that he's done for a leading brands Pangaea about re
- 1:36engineering on shop. And I thought it'd be very
- 1:37interesting for us to talk about how you approach re engineering
- 1:40and why sometimes this can be the right approach.
- 1:42So warm welcome to Michael Klim. How are you, Michael?
- 1:46I'm fine. Thank you.
- 1:47Thank you for having me here. It's a it's a total blast to be
- 1:50here and discuss this this project.
- 1:53Yeah, I really appreciate you taking the time.
- 1:55I always enjoy reading what you know, what you share on
- 1:58LinkedIn. I recommend anyone to follow
- 1:59Michael got a lot of knowledge in the E com development space.
- 2:04And before we get into the questions, Michael, do you want
- 2:07to give people a bit of a flavour for who you are and and
- 2:09and what commerce UI is? Yeah, sure thing.
- 2:12So yeah, as you mentioned, I'm I'm Michael Kleem, I'm Co
- 2:16founder and city at Commerce UI. My main responsibility is to
- 2:20oversight of our technological efforts and I guess starting
- 2:25quite recently and trying to become more of this technical
- 2:28public voice of the company when it comes to to to Commerce UI.
- 2:33Commerce UI is is award-winning boutique agency.
- 2:37We are team of a team based in Poland.
- 2:40We do design and technology and we work a lot with fashion,
- 2:46beauty, premium, premium brands, mid market, mid market plus
- 2:49size. Amazing.
- 2:52Thanks very much. So let's get into this because I
- 2:55think this is a really interesting topic because I
- 2:56think a lot of times when people see problems in tech and if
- 3:00people aren't involved in the decision making around how that
- 3:03platform was initially like engineered that you know how it
- 3:05was configured, that for how the front end was built, they can
- 3:08think platforms at fault. But technical process debt build
- 3:12up is just standard. It happens in most businesses.
- 3:16Can you talk people through like how does like process technical
- 3:19process debt may often make people think the issue's the
- 3:21platform rather than it is? It's actually about the
- 3:24configuration base. And you talked about thoughtful
- 3:26maintenance, which I think is a really nice way of putting it.
- 3:29Yeah, yeah, yeah, definitely. So I guess it always starts with
- 3:34this kind of like a very common story for for most of the tech
- 3:38projects like you starting with with a clean state, you define
- 3:43your requirements, your tech stack, your timeless, everything
- 3:47goes right. And then you start developing.
- 3:50And when you near the launch, when you near the end, like at
- 3:55this point, most likely you already have a some depth
- 4:00accumulated, like it's either because of business reasons,
- 4:04maybe timelines, timelines changed, maybe some business
- 4:08requirement changed along the way.
- 4:10The point is that you most likely already have some, some
- 4:14technical depth there. So then you jump into the
- 4:19retainer mode with some agency, maybe another button, maybe your
- 4:22internal team, and you go into this maintenance and maintenance
- 4:27it's, it's hard thoughtful maintenance, it's even harder.
- 4:32So this thoughtful maintenance is when you, when you do
- 4:38maintenance, when you fix things, you introduce new
- 4:41features, bigger and smaller, you optimize existing flows, you
- 4:44introduce new business flows. But at the same time you fix
- 4:50your current approach, You optimize your current approach
- 4:53and you consolidate your, your solutions taking into account
- 4:59all of the developments in, in your e-commerce platform in, in
- 5:04all of the first parties that, that are there and basically all
- 5:09of the new things. So that that's the thoughtful
- 5:12maintenance and that's the ideal situation.
- 5:14But the reality is that it's rarely happened happening and
- 5:22most often developing new features, shipping new things
- 5:25will be prioritized. And that's completely fine.
- 5:28That's completely fine. We, we are all in this
- 5:32e-commerce space and this is very, very dynamic, very
- 5:35competitive space. So that's totally
- 5:38understandable. But at this at at some point or
- 5:42eventually, the test will arise to, you know, to the point of of
- 5:47of no, of no return not to be too dramatic there.
- 5:52Greben create digital flagship stores for iconic brands of
- 5:55today and tomorrow. We are the number one Centra
- 5:58agency in the world with more than 50 Centra implementations
- 6:01to date. We are also e-commerce design
- 6:04specialists, platform and tech partner agnostic.
- 6:07Try our new e-commerce design subscription.
- 6:10Visit greben.com. Yeah, that makes a lot of sense.
- 6:13And actually it's really interesting because you can you
- 6:17can build up technical debt because at the point in time you
- 6:21made a decision, maybe the platform didn't quite do
- 6:24something how you wanted to. So you you've created something
- 6:27custom or you've bought in a third party, but then the
- 6:30platform road map progresses. It makes that customisation
- 6:35redundant and there's efficiency in reverting back to like native
- 6:39functionality rather than having all this custom code built.
- 6:41And so do you see that a lot when you're when you're working
- 6:44with clients where actually it's historical decision might be
- 6:47right, but actually now 12 months, 24 months later that
- 6:50decision can be improved? Yes, that's that's a very often,
- 6:56often scenario to be honest. And it's not always an easy,
- 7:02easy, easy scenario either for, you know, technical partner or
- 7:07the business, because there is now you have a decision to to to
- 7:11make like you've already invested into building something
- 7:14custom. Most likely you invested a lot.
- 7:18This is like a custom solution. It may work or may work to, to
- 7:23some extent. It may have some, some issues.
- 7:27And now your platform introduced this natively.
- 7:30So you have to try to assess best with your with your
- 7:37technological partner. You have to assess if it's
- 7:40actually worth investing into jump into this native solution
- 7:44the platform offers you. Most likely the maintenance cost
- 7:47in the long run will be will be will be lower.
- 7:51But it's not obvious obvious economical choice.
- 7:56Yeah. And I think that's really good
- 7:58advice because just because it's native doesn't mean it's the
- 8:00right thing for you. I'm going through a few
- 8:02conversations with with customers at the moment on going
- 8:05native subscriptions in Shop 5 versus sticking with a third
- 8:08party. And the the native saves your
- 8:10licence fee. And it, it's simplicity because
- 8:12obviously features get rolled out and they become available,
- 8:16but there are compromises to make and that the effort of
- 8:19migration is significant and brings risk.
- 8:22So yeah, it's, it's great, isn't it?
- 8:24We, we all face these nuanced decisions in technology.
- 8:27There's no, there's very rarely a perfect decision.
- 8:29But let's, let's put that lens now on the, we talked about
- 8:34Pangaia as one of the recent projects you've worked on.
- 8:37And it's, it's not, you know, it, it's common in, in
- 8:40development where you start working with the company.
- 8:44They've had multiple agency partners before who've made lots
- 8:46of decisions, you know, for a lot of time for very good
- 8:48reasons. But those decisions might not
- 8:51necessarily be what's right now. And your focus is on changing
- 8:54the technology backbone to tackle technical debt.
- 8:57So can you talk us through how did you approach that because
- 9:00that feels quite complex. And what were some of the key
- 9:03like issues, causes of that debt that you found Centre is the
- 9:07only. Commerce platform built
- 9:08exclusively. For fashion brands, over 300
- 9:10fashion brands run on centre. Including icons like Paul Smith,
- 9:12Jeff Ave. and Nudie Jeans. Centre solves business logic
- 9:15challenges for fashion brands. I know of the platform and they
- 9:18are on a mission to help you scale up rapidly globally across
- 9:20all online and offline channels. Take your fashion brand global.
- 9:23Faster, further, stronger today. Yeah.
- 9:27So let let me start from the end.
- 9:30So whenever you, you accumulate the technical depth and you are
- 9:36past this point of of no return, you are already most likely in
- 9:40the place that there is like a lot of frustration around the
- 9:44project in, in basically every department.
- 9:47Like what from 1 developing new features takes longer time,
- 9:54sometimes unreasonably long time, at least from the
- 9:57perspective of non-technical people there, there's more bags
- 10:05or bags are more frequent, causing another frustrations
- 10:09for, for the users for the business operations are
- 10:14frustrated because most likely they have to do like a lot of
- 10:18worker runs, even for a simple task and simple tasks just take
- 10:24longer time. So, and most likely the site is
- 10:28not performing like this. This is, this is rather a
- 10:34combination of, of, of different, different faults or
- 10:39different, different difficulties in the store.
- 10:42So at this point, at this point, there's a lot of frustration
- 10:47around the project and depends on the case, of course.
- 10:50But having this discussion with the clients can be either, you
- 10:56know, like very complex, like they've already paid for some
- 10:59solution. They've it's already there.
- 11:03There is a lot of, you know, fear that the situation will
- 11:06happen again. So on.
- 11:07So it can be very complex pursuit client that OK, now we
- 11:10have to now we have to do the revamp or it can be like very,
- 11:16very easy. Like the frustration may be so
- 11:19big that every, everyone feels that OK, some something bold has
- 11:25to be done in order to change the situation.
- 11:29In that case, I guess our, our task or people like us is to
- 11:37actually assess the, the situation and, and understand
- 11:45what actually has to be changed. Like do we have to change the,
- 11:48the, the e-commerce platform? Like maybe it's, it's actually
- 11:51fit, fit for the task and we just have to do a revamp of a
- 11:56particle particular part of the of the technical technological
- 12:02solution. So yeah, sometimes, sometimes it
- 12:05can be hard to persuade clients. OK, we need something bigger,
- 12:10bigger in terms of of of of the project.
- 12:13But sometimes you have to persuade client key to actually
- 12:16hold on and and stick to the, the, the existing solution.
- 12:20So I guess in Shopify there is this kind of 1/3 option that
- 12:27resonates with with the business very well because either you
- 12:34there is like this kind of a in between option.
- 12:37You don't have to deal with your technological stack entirely.
- 12:41You don't have to just maintain and fix the of the things.
- 12:45You still can retain your core functionality, your core
- 12:51e-commerce platform, your core third parties integration and
- 12:55rebuild this kind of storefront front end integration layer,
- 12:59however you call it. So the way you actually connect
- 13:03all of those technological, all of those third parties, those
- 13:08integrations and and this part that this actually that the user
- 13:14interacts with. So there's this kind of a in
- 13:17between the work that's very much it's, it's very usually
- 13:22very, very reasonable way to go with.
- 13:26And I know you said that you, you started with a clean slate.
- 13:28So rather than, hey, what have we got?
- 13:31Let's try and, and, and tweak. There's start from start from
- 13:34scratch, rethink it through. Can you, you talk us how, how
- 13:38did you, how did you come up with the solution?
- 13:40What solution? How did you refine the site set
- 13:42up? Because for a lot of people they
- 13:44they were just saying, OK, so we just changed the theme versus
- 13:48actually changing maybe like you know, business logic around that
- 13:52as well. So how did you ensure you build
- 13:54you build back better? Yeah.
- 13:57So I, I guess there is a couple of layers there.
- 14:00Like for one, we have to assess, we have to assess technological
- 14:06stack from from an integrations point of view.
- 14:09Like are those integration actually makes need the
- 14:13business, business expectation? If so, let's retain them.
- 14:17If not, let's try, try to find another solution.
- 14:21Then there is the code base itself, like how, what's the
- 14:27technological that like what, what are the solutions?
- 14:30What's the, what's the actual state of the, of the code base
- 14:35and in terms of, you know, like best engineering practices and
- 14:38so on. Then there is all of the
- 14:43operations of the operations and how operations work, basically.
- 14:49And what, what pain points are there like in, in, in Shopify?
- 14:57The there is a common story that people actually trying to find a
- 15:00walk around and non-technical people can do a lot.
- 15:04I can, they can add custom CSS, they can do, maybe they can add
- 15:08a different app for one, you know, one single case they
- 15:14couldn't achieve before. So they installed the app, they
- 15:17achieved this how somehow, but this app is ruining the, the
- 15:21site performance. So we, the, the way the
- 15:25operation works also impacts the, the technology technology
- 15:29there. So we have to assess all of that
- 15:33and pick and choose what stays and what's not and when, when
- 15:42you pass, when you're, when you pass, this point of no return
- 15:46when it comes to technological death.
- 15:49Now you have to convince clients that OK, we have to do the front
- 15:53bar from from the ground up basically.
- 15:57And how much of this was about the theme itself?
- 16:01Was it this about building a custom theme?
- 16:04Was this a mix of theme class, you know?
- 16:07Yeah. Bespoke custom apps and
- 16:10elements. Yes.
- 16:12So I guess a lot of difficulties were about about the team about
- 16:18like a lot of complexities, a lot of technological depth were
- 16:22was hidden inside the inside the team.
- 16:25And there are a couple of reasons for that, I guess
- 16:29definitely. And guy was is a very dynamic
- 16:32brand. So there's this kind of a
- 16:35priority for, you know, ever changing innovation there.
- 16:39So a lot of features were introduced without, you know,
- 16:43cleaning the old ones. So and they were built on top of
- 16:47of old solutions. At the same time, the code base
- 16:52went from from partner to partner.
- 16:55And every time the code base, the code base is changed, the
- 17:02the ownership of the code base changed.
- 17:04The technological that will arise like even if the next
- 17:09agency, next partner is is approaching the the challenge in
- 17:14a in a thoughtful way, there is no way that they will follow all
- 17:19of the conventions. There is no way that they will
- 17:21be aware about about about 3 decisions being made or every
- 17:27convention there is in the project.
- 17:29So they will start doing the things in their own way that
- 17:34just natural and that will also accumulate the technological
- 17:38debt. So we, we had to deal with that
- 17:41and there was, there's, there was so much history inside the
- 17:45code base that we just had to redundant from a perspective on
- 17:50a, on a team. But there was also the point I
- 17:55mentioned about the apps, like the whole app list had to be
- 18:00assessed once again. And we had to pick, pick and
- 18:03choose the best, best solutions at stake.
- 18:06We have to remove some, some of them, some of those apps were
- 18:10just installed because at the time Shopify didn't, didn't
- 18:15support some cases. And now you could do it natively
- 18:20without an app. So there's no, there's no reason
- 18:24to keep it around actually. So another, another layer of it
- 18:29was the the editorial or the operation side of things.
- 18:34So there was, there was a lot of custom solutions, a lot of
- 18:45custom code added inside the team editor or added inside the
- 18:51online editor. And it, it, it accumulated to,
- 18:55to, to a state where it was so hard to reason about it.
- 18:59Like even for a people that actually work, work with it on a
- 19:05daily basis, but also for us, for technical people for
- 19:08injuries, it was so hard to reason about it.
- 19:12Nobody knew why it's there and what's what it's actually doing
- 19:17what's what are the consequences of removing it, that it's just
- 19:21made sense to you to start from from a clean state.
- 19:24And now knowing what are the expectations, knowing that their
- 19:28clients better and what is used and what's not, we could do much
- 19:34better decisions. Yeah, tonight I've seen it on a
- 19:38few projects where where people have added a lot of complexity
- 19:42into a theme and not just on Shopify, Shopify, big commerce,
- 19:46you know, pick, pick your econ platform that's using a theme
- 19:48versus like a custom front end. And it's to give the, the
- 19:52content teams the the NTH degree of flexibility for I think we,
- 19:57we told those for like, you know, we want to be able to, to
- 19:59put our movers tech to any part in the image.
- 20:02We want to be able to change change the layout around.
- 20:06And you have, you have like 20 or 30 different permutations of
- 20:10a single layout and all of those require code in order to make
- 20:14those changes. And then in in practice, the
- 20:17content team maybe uses two or three of them.
- 20:19And what might have been a great aspiration in discovery if we
- 20:23want the ultimate flexibility isn't really practical and the
- 20:26trade off hasn't been understood.
- 20:29Do you see that quite a bit? Then there's just the the need
- 20:31to be or the ability to simplify and say that do you really need
- 20:35this, this level of flexibility? I, I, I honestly think it's
- 20:39crucial like you can simplify the code, you can simplify your,
- 20:44your technological approach. But the biggest factor that
- 20:48drives the technology depth, that drives the complexity of
- 20:51the projects is those this unnecessary complex of business
- 20:55logic. Like as you mentioned, you, you
- 20:59people sometimes may have five different header templates or
- 21:04maybe very complex, complex overlapping merchandising logic
- 21:10that nobody, nobody can reason about or nobody is aware how
- 21:15it's actually what, what we showed as a first item in the
- 21:20collection. So I guess this is a very good
- 21:25first step to, to, to, to tackle this kind of this kind of
- 21:31immigration, this kind of a technological debt is to
- 21:34actually assess what's needed to be there, what is actually what
- 21:39we are actually using and what's actually driving the, the
- 21:43conversions, driving the size and get rid of everything else.
- 21:49And don't get me wrong, it's not always about creating very
- 21:53simple, very minimal solutions like I guess having five
- 21:58different header templates, it's fine.
- 21:59But to some extent we have to be aware that it, it comes with
- 22:04additional cost, it comes with additional codes, it comes with
- 22:08bigger code bases. Bigger code bases means bigger
- 22:11box and and more expensive maintenance basically.
- 22:16So we have to be aware about the cost and if it's not a business
- 22:22critical thing for you, if it's not a value that or if it's not
- 22:28something that generates your unique value, it's always good
- 22:32to to to to step back and think if it's if it has to be actually
- 22:37that complex. The thing is that what we found
- 22:41out also that even if it's created in a in a good way, like
- 22:46it's, it's working, it's not buggy.
- 22:48It's it's, it's created properly.
- 22:51From an engineering point of view, operations people are in
- 22:55general, people are not using complex things.
- 22:59If it's something, if you have to read a ton of documentation
- 23:03before you use it, you most likely will will try to to avoid
- 23:09it as much as possible. Yes, Yeah, I think that's The
- 23:15thing is, is, is a feature might sound great, but is it adding
- 23:19the value that in terms of the effort that's required to to
- 23:22develop and maintain it. Another area that I really want
- 23:25to talk about is automation. And this is a hot area for most
- 23:29e-commerce teams that how can we, how can we simplify
- 23:32operations by automating tasks and how can we use automation
- 23:37tools to, to solve some of the business challenges?
- 23:41And you, you do a lot of work in Shopify with Flow and you've,
- 23:44you've talked about some really cool use cases on LinkedIn.
- 23:47And I know with the Panguide project, you've got a couple of
- 23:50really good examples of where you've, you've used flow
- 23:53alongside things like meta objects to help you around like
- 23:57product data, content management.
- 24:00Do you want to talk? Do you want to talk through a
- 24:01couple of examples of how Flow Builder can help solve some of
- 24:05the business challenges? Yeah, sure.
- 24:08So I guess I guess flow and meta objects I are underestimated in
- 24:16in in Shopify work. So for us there is Shopify
- 24:22before flow get attention before meta object were introduced and
- 24:26there is Shopify after after those.
- 24:29So before flow and before Meta objects, whenever Shopify didn't
- 24:37prepare its solution for your challenge, you're pretty much
- 24:42forced to create a custom solution like and this custom
- 24:46solution will have to live outside of the shopified.
- 24:49So it could involve, you know, creating a custom server,
- 24:52creating a custom database API and things like that.
- 24:57In general, like a completely separate technical solution
- 25:01which has to which have to be also maintained and creating
- 25:06such solutions dries up the cost, dries up the the
- 25:10complexity of the whole project. It requires more, more
- 25:13maintenance. So when the flow got more
- 25:18attention, when meta objects were introduced, we actually got
- 25:24low level. We as an engineers, we got low
- 25:27level tools to solve those unique challenges.
- 25:30And every brand has some kind of a unique challenge.
- 25:33It can be either on a front furnishing side of things like a
- 25:38unique selling flow, or it could be on an operations level.
- 25:43Like you may use enormous spreadsheet of, of the products
- 25:48that you got from a central team from a central team.
- 25:53And you just, you just want to upload this, this CSV or maybe
- 25:58automate this, this spreadsheet to, to add the products and is
- 26:02up the the the operation side of things.
- 26:05So every brand has this kind some sort of unique challenges.
- 26:09And with flow with meta objects, we can actually accommodate them
- 26:13without going out from from Shopify ecosystem that drives
- 26:17the the whole complexity down, that drives the the cost down
- 26:21and time to to to the production down.
- 26:24So additionally, so additionally, flow can make
- 26:33merchants like so much easier in terms of, you know, I've already
- 26:37mentioned importing content, but you know, automated 314
- 26:41synchronising data between the synchronisation flow can also
- 26:47make some some apps obsolete. Like you can introduce a very
- 26:51simple automated flow that is doing something that, that you,
- 26:57you had to install app for. So and you, you had to pay for
- 27:01it. So, and now you, you, you can do
- 27:03it full flow. And recently flow got much more
- 27:07attention. So there's much more things you
- 27:10can do do with it. When it comes to Pangaya,
- 27:14there's this, we have this kind of a cool example.
- 27:16So Pangaya products have so many colours that like that's their,
- 27:23their thing. Every product has, you know,
- 27:25more than 20, it may be even 30 different cause and what Pangaya
- 27:30is doing, they follow this kind of standard approach of having
- 27:35separate colours as a separate product instance of variants.
- 27:40So each colour permutation is a separate product inside the
- 27:45Shopify and all of those are combined on a single PDP and it
- 27:50fails for the user as a single product.
- 27:52So now you have to now you have to combine those products or
- 27:57give the merchant option to combine them.
- 28:00And this is this is pretty standard approach in Shopify.
- 28:05People are doing this for four years and Shopify noticed that
- 28:10and actually prepared an app for it, it called called Combined
- 28:14Listings and it's it, it's doing the its job.
- 28:17It's, it's pretty good, but it's has its own limitation.
- 28:20For one, you have to do it manually and for two, it has
- 28:25limitation of around 20 products, if I remember
- 28:28correctly. So less than less than Bangaya
- 28:33news. So what we've done in this case,
- 28:35we used Shopify flow. We we react for when product is
- 28:41created or updated, we run Shopify Flow that runs some code
- 28:47that actually looks for a different product in the same
- 28:51naming convention and connect those through the meta objects.
- 28:56So it's creating a new meta object that's holds those
- 29:00connections between the product that's later on can be used on
- 29:03the front end and that's done with just Shopify flow and just
- 29:08meta objects in a very automatic way.
- 29:10So for one, there is no longer like for 1:00 we, we just were
- 29:17able to do it and that wouldn't be possible with, with combined
- 29:21listing up. So for one, we just could do it.
- 29:26And for two, it's just simpler because it's, it's, it's
- 29:29followed convention, it's done in an automated way.
- 29:33So this is I, I guess this is like a quality of life change
- 29:38there with the flow. Yeah, and I think this is, this
- 29:42is one of the great ways a lot of EE com technology providers
- 29:45have evolved in the last few years.
- 29:47Is, is flow based stuff is is not new, but the scale at which
- 29:52it can enable businesses to, to to handle automations.
- 29:56I mean, I've used it, I don't use it in nearly sophisticated
- 29:58way as you guys see, but I've used it with clients for doing
- 30:01simple things. I wanting to auto tag customers
- 30:04with an account enabled flag when they create an account
- 30:06versus when they've done a guest order, for example, so that that
- 30:09can be passed through to CRM so that people can understand and
- 30:13model people who are account holders versus those who aren't.
- 30:15And, and look at how they then on board people to register
- 30:18accounts for incentives. So there's all these simple
- 30:20things you can do as well as the more complex stuff that Michael
- 30:23his team does, which is way beyond my technical
- 30:25understanding. But the, the other question
- 30:28relates to that, which is an interesting one is because you
- 30:30said that the, so they've got the men and women's.
- 30:32So you'll, you'll have like a black hoodie and it will be the
- 30:36same item with different sizes that the, the, the men and women
- 30:40shop across. And therefore the product, the
- 30:44men's and the women's size version of the size has the same
- 30:48skew, which obviously has a stock challenge.
- 30:51So I'd love to hear how you've handled that.
- 30:52So that they're pulling from a single stock file, but they're
- 30:55essentially two different products in Shopify handling
- 30:58that so that you don't end up in that horrible stock issue.
- 31:02Yeah, so. Essentially they they have two
- 31:05different products in Shopify that is actually one product.
- 31:08So we can we can call those displays.
- 31:10So so we have one product, but we want to display in in it in a
- 31:15different manner. We want to have a different
- 31:18photos in terms for for minimum 1.
- 31:20So and I guess the this concept is borrowed the concept of
- 31:26display. I was first aware about it from
- 31:30Central. So there are two, two separate
- 31:34products, but they, they, they're single SKU and they have
- 31:38two different inventory levels. So whenever you buy a man
- 31:42version of the of the product, what should happen the, the, the
- 31:46stock level of the woman version should also go down.
- 31:50So those have to be synchronized and at some point of course
- 31:56there is an ERP there. So, so the order goes into the
- 32:00repeat, the repeat comes back with the updated inventory
- 32:02levels. But, and that's a very common
- 32:06scenario, it happens over time. It can take, you know, half an
- 32:10hour, it can take one hour. That's just too long in, in this
- 32:15time, the stuck could be, could be bought out.
- 32:18And you know, that gets to a very poor, poor user experience.
- 32:22So for that, we also used Shopify Flow.
- 32:26Whenever 1 inventory level changes, we actually we we're
- 32:32looking for different version of the product in a similar fashion
- 32:36to the previous example, OK. And then we update the inventory
- 32:41level of this product that we found.
- 32:43So we essentially we're synchronizing those inventory
- 32:47level in between. And I guess it's it's achieved
- 32:51in a quite cool way because Angaia site is an online store
- 32:56coupon 0. So it's a liquid best site, but
- 33:00we found a way to actually utilize hydrogen for this
- 33:04project. And and in this exactly in this
- 33:07exact case. So Shopify flow has also has its
- 33:13own limits it you can do a lot, but you cannot do everything.
- 33:18And in this case you you don't have this kind of a ready made
- 33:22solution for it. So once again, we need an API,
- 33:26we need a server to to do custom things.
- 33:29And historically we would have to create this custom server
- 33:34outside of the Shopify hosted somewhere else.
- 33:36It once again, it's at complexity, it acts costs.
- 33:41But with hydrogen, hydrogen can be also can also served as a
- 33:46server as an API layer. So basically we've spinned up
- 33:51very lightweight hydrogen project project next to this
- 33:56online store 2.0 we're using flow, we're calling out the,
- 34:01the, the hydrogen hydrogen is doing this all this custom stuff
- 34:07and, and there we have it basically.
- 34:10So we've built very custom integration, very custom
- 34:13solution without going outside of the Shopify like everything
- 34:17is baked in. It's very easy to reason about
- 34:20the the architecture. It's not scattered and we not,
- 34:24we are not adding additional cost though.
- 34:28Yeah, I really, I really like that example.
- 34:30And I guess the, the the use cases for flow are like they're
- 34:36exponential though, because it keeps adding data points that
- 34:39you can connect to. So I'm really intrigued to see
- 34:41where this goes. Another part of the project I
- 34:43think is interesting is the the choice of like specialist third
- 34:46party tools that you've built into the stack that aren't
- 34:50Shopify's. So Claview is 1 product
- 34:54discovery engine. A lot of people will know it.
- 34:57And with Claver, a lot of people typically go down the route of,
- 35:00of doing the standard JavaScript implementation.
- 35:03And that's not, that's not saying that's a bad thing at
- 35:06all, because it works very, very well.
- 35:08But obviously JS can have an impact in terms of like page
- 35:11speeds. You went for the API integration
- 35:14method. I'd love you to talk us through
- 35:16why that approach and what were the benefits to to you and
- 35:19Pangar of going down the API route.
- 35:21Yeah. So essentially there are two
- 35:23reasons. And whenever we had an option,
- 35:27whenever we see that the integration has have a, you
- 35:30know, like a mature API, mature customs allow for, you know, for
- 35:37building custom solutions in a mature way, it means that it's
- 35:41easy. It's, it's, it's easy to to
- 35:44reason about. We are actually thinking if
- 35:46that's a bit rude. The the reason for it is because
- 35:50the those out-of-the-box integration as you mentioned,
- 35:53they are working very good in, in terms of Kavel.
- 35:58It's, I would say it's, it's, it's making wonders actually.
- 36:02But like with every general solution, they have to account
- 36:07for many different cases and they have to think in a general
- 36:12way. And essentially that means that
- 36:13they have to load much more code then most likely you need for
- 36:18your single case. And that will always impact the
- 36:22the site performance. So loading a lot of of of JS
- 36:26code will impact your performance.
- 36:28And that's the very first step that you, you have to do when
- 36:32you optimise performance is to minimise the, the impact of this
- 36:36client, client JavaScript code. So whenever we have a chance to,
- 36:41to actually do it in a custom way and build custom, we already
- 36:46have those, you know, components that, and it's more about how
- 36:51the data is provided to, to this, to this view, whenever we
- 36:56have a chance to do that and it doesn't drive the cost very
- 37:00much, we, we try to follow this road.
- 37:03So that's what we've done in here.
- 37:05We we saved a bit of a performance there.
- 37:10Another reason for it was that that in the Pangaea case they
- 37:16have this kind of unique, I would say maybe unique or
- 37:20interesting approach to how they do certain sections, how they do
- 37:24recommendations and search. Sometimes they want to show
- 37:28manual results, sometimes they want to they want to show a
- 37:34keyboard results and maybe Shopify driven results.
- 37:37So they have essentially those three ways of showing different
- 37:40recommendations that are different results.
- 37:44If we would go with this customs with out-of-the-box solution,
- 37:48essentially that would mean that they have to have three
- 37:50different sections and they have to know when to use one.
- 37:55Now they could have a just one of it tailored for them and they
- 38:02just choose if they want to use a Clivo or or a custom
- 38:05solutions. And that's that's that's just
- 38:09creating, you know, like a tailored editorial experience
- 38:12for them. And it's so essential part of
- 38:15the site that optimizing this, this low level operations flow
- 38:20was, was a marking gender. A nice nice that makes a lot of
- 38:27sense and the another question that comes up a lot and some of
- 38:32the Shopify project that we're going to know is classic versus
- 38:35new accounts. Obviously new accounts is where
- 38:38Shopify is is headed. There are there are some key
- 38:43differences. I know with new customer
- 38:47accounts, it's using the one time password.
- 38:50Classic used to have the ability to like standard e-mail login.
- 38:52And one of my clients who got an older demographic, they went for
- 38:55Classic because of that, because the one time password they
- 38:58thought would actually put their audience off.
- 38:59But there are certain certain things that aren't as compatible
- 39:03with classic as they are new. So you got, again, you get all
- 39:07these nuances of technical choices and you move with the,
- 39:11on this project with Panguy, you move from Classic to new.
- 39:14What was the business driver for that and what what's the
- 39:16advantage of new to Panguy? Yeah.
- 39:21So the whole idea for this project and like the, the, the
- 39:24main, you know, notion or the main main idea that we follow
- 39:28throughout this whole development was to, to simplify
- 39:31and to get closer to, to, to native Shopify.
- 39:36So that was the, the, our entry points to, to, to this, this
- 39:41discussion, this decision. So I thought I agree with your
- 39:45case about the new customer accounts.
- 39:47They are not of use. They're very, there are a couple
- 39:51of cases where there doesn't make sense, especially that
- 39:56account extension is not yet there.
- 39:57So either you take it as it as this or or you don't.
- 40:01So it's kind of a like a this, this transition phase we're
- 40:06facing right now. So definitely not obvious
- 40:09choice, but looking at the problem from a simplification
- 40:14point of view, this takes out so much complexity out of from out
- 40:19of the out from the project that.
- 40:23It's very much worth to to consider it, even if it's very
- 40:27at this point, it's very basic, but the account in general is
- 40:31this kind is this part of application that I like very
- 40:35much that Shopify decided to to build their own solution or
- 40:40out-of-the-box solution for it, for one.
- 40:44And I always like to compare it more to to check out.
- 40:47So you have to worry about the security.
- 40:49That's always that's always one thing.
- 40:54Another thing that account is this kind of is this part of
- 40:58application that people like to to be standard like you, you
- 41:02want to change your address, you want to change your profile,
- 41:06maybe look at your order, maybe spin up the return process and
- 41:10so on. So like this is very like
- 41:15sometimes there, there may be more to it, like loyalties,
- 41:19maybe some custom flows, maybe you would like to do more in
- 41:24terms of self management. But in a nutshell, people doing
- 41:27the same stuff on the account and they expect it to just work.
- 41:31It's it's more pragmatic approach to to the to the to
- 41:36this part. So that's taking, it's taking
- 41:42out a lot of complexity out of the project.
- 41:44It's really doing, doing it's job very, very well.
- 41:50In terms of Pangaea, as I mentioned, it wasn't an easy
- 41:55decision. We had to make some compromises
- 41:57that that we, so there is certain functionalities in terms
- 42:03of loyalties and so on that has to be gated and cannot be added
- 42:07to the account at this point because because account
- 42:11extensions not there. So we had to build this custom
- 42:14gated pages next to the account and think how to show that user
- 42:19in a in an understandable manner.
- 42:23But we after our assessment, we get to the conclusion that it's
- 42:28still still worth it, worth it. And whenever account extensions
- 42:32will be rolled out, we can we can migrate those gated pages
- 42:38step by step. Yeah, that makes a lot of sense.
- 42:43And this is the reason why I love doing these episodes
- 42:44because I've worked with so many different brands and you find in
- 42:49each and there's a commonality in terms of how maybe the
- 42:52projects approach from a core architecture point of view, but
- 42:55the front ends build and the way the themes developed and whether
- 42:59you go for, you know, classical new or you use a particular
- 43:02third party app changes depending on business needs and
- 43:06effort and time. So I find these conversations
- 43:09fast. I could ask you like hundreds
- 43:10more questions, but I won't. I appreciate you have a business
- 43:13to run as well. So Michael, really appreciate
- 43:15you taking the time to chat to us today.
- 43:18If anyone wants to, to reach out to Michael and and prove him on
- 43:21any more of these kind of, I guess, engineering decisions,
- 43:24you can contact him through LinkedIn.
- 43:26He's all there's also a link on our landing page on the website.
- 43:30Also, if you want to, if you've got a potential RE architecture
- 43:35needed on on shop for you considering these these sort of
- 43:38complex decisions, he's always happy to share advice and chat
- 43:42to you. Thanks for joining us, Michael.
- 43:45Thank you very much. It was a pleasure.
- 43:46Thank you for having me. Bye.
- 43:48Bye. And thanks everyone for
- 43:49listening, hope you found it useful.
- 43:50Do shout out with any questions and keep your ears peeled for
- 43:53next week's episode. You're listening to the Inside
- 43:57Commerce podcast, hosted by experienced consultants James
- 44:00Gerd and Paul Rogers. Inside Commerce is your fast
- 44:03tracked additional knowledge featuring interviews with
- 44:05leading e-commerce retailers, industry thought leads and the
- 44:08C-Suite of key tech companies to explore the latest digital
- 44:11strategy, customer experience and tech.
- 44:13Discover what the latest trends really mean and get practical
- 44:16advice based on experience to help fine tune your e-commerce
- 44:19strategy and tech stack. Follow us on LinkedIn and reach
- 44:23out to James and Paul if you'd like to discuss an e-commerce
- 44:25project. We hope you enjoy the latest
- 44:27episode.