Sander Mangel on the impact of microservices on eCommerce

Care to share?

In this episode of eCommerce Talks, we meet with Sander Mangel,  Chief of Technology at Vue Storefront and famous for being a long-time Magento Master.

Being in eCommerce for over 12 years, Sander has gained amazing experience as a developer, tech lead and now as an evangelist of modern technologies. Along with our host, Marcos Bravo C., they explain the concepts of microservices, headless and service-oriented architecture (SOA), and how they’re changing the environment of eCommerce. They also take a look at the leading eCommerce platforms like Magento, Shopware or Hybris, their adoption of the headless approach and other modern technologies, and why some of them were bought recently by software giants. 

As our guest predicts, the headless approach will grow in the foreseeable future, and so, together, we look for the answers on their usage and impact on eCommerce. How to securely use microservices and SOA to improve performance or rebuild old, monolithic architecture? How will microservices and SOA impact the security of IT architecture? And how to easily change applications to a  microservice-oriented architecture?

Watch the full episode or dive into our transcription of the talk. 

eCommerce Talks – video:

eCommerce Talks – podcast:

eCommerce Talks – transcription:

Marcos Bravo C.: Sander, thank you for joining me today. We’re very excited! I’ve known you for a little bit, but I’m sure that some people listening to us would like to know more about you. Can you tell us a little bit more? How did you end up in the eCommerce world? 

Sander Mangel: Right, so about 12 years ago I got hired as a developer at an agency. They were doing this thing called Magento, which I pretty much never heard of before. Basically they told me like: “You’re gonna be the lead on this department and get some books to start studying”. I had like 30 days to get into the whole Magento thing, three books, read them in the evenings, and then I started, like “hopefully I’m prepared”.

And I’ll admit it, the first couple of weeks were rough but slowly it started to grow on me. I did eCommerce before that, mostly bespoke solutions which means that everything is painful: taxes, discounts, products, you name it. That’s all you have to build yourself, right? And then you have the solution out there, that pretty much seems to have it all, and if not, there would be a module for it on the ecosystem. So that was interesting, and I started to get more and more into eCommerce since it’s a very dynamic way of working. Before that I did a lot of websites with a contact form and maybe some job openings, but it’s very static, right? And one of the things I love about eCommerce is the direct return on investment that this client is seeing. If you fuck up, they will not have their income.

So, that progress I did: a couple years of development, started to get more and more into lead roles, the dual consultancy at agencies, also freelance, ended up at a merchant for three and a half years. It was a very interesting experience, as you’re somewhat protected from the realities of business when you’re on the agency side, right? If something goes wrong or if you have to do a new feature, there is an eCommerce manager on the other side, on the client side, that scares off a lot of things and then, when you transition to the merchant side, you have to get down into the trenches, do the negotiations with the finance department or logistics, on how we handle packages, and that kind of stuff.

– So, literally you’ve been basically from development, all the way up to the salesman of the solution.

– Oh yeah. And that was interesting. It was a huge amount of learning and not on the technical side, but on the more soft skills side. How do you get people to buy into this kind of stuff? How do you explain X months of development on the feature that for them is like “I just want the button”? That was very interesting. After that, I went to the product development side. I got hired at Vue Storefront, which is a headless frontend for eCommerce, to help them push their product on the market. 

– Now, let me introduce you as a Magento Master.  Even when I got introduced to you, my first question was – what is a Magento Master? This guy is like a Jedi of the Magento world? How do you become a Magento Master? What does it actually mean?

– Yeah, it’s a term that when you’re not around the Magento space can come off as slightly odd. So, Magento in form of Sherrie Rhode, she’s a community manager,  Magento started this program about three or four years ago to recognize some of the more active community members which entails either doing blogs or conferences, or even just mentoring other developers. I’ve been organizing, together with someone else from the Netherlands, Meet Magento for a couple of years, and also an unconference, which is more technical conference, I also have done some online conferences. It’s kind of an escalated hobby, you could say. So, because of that I got elected a Magento Master, one of the 22-23 around the world. Yeah, so that’s very, very cool.

– That is very cool – that’s sounds very cool.

– It’s a bit of an honor I’ll say that.

– I’m gonna try to get in a little more details. You get to see, especially at the conferences, what’s happening with all ideas of microservices and headless pieces of software, which I’m still trying to put my head around. How is it changing the environment? People used to have just one solution, basically host it at their online store or in the computer, but it’s everything in one place. Now there’s tons, if not thousands, of different solutions that you can just literally connect to and I guess it saves you time. How do you see this happening?

– That’s a good question! I think headless has been around for ages. Basically, it means that you have an application, either online or offline, that doesn’t have a graphical interface. In this case it could be your average shop frontend. So, headless is becoming more and more popular, I think, mainly because as technology is getting more and more complex and the whole ecosystem is maturing, you see that it becomes more important to choose the best of breed when it comes to any sort of technology. For example, search is a very, very clear one, where you would have your normal search, which would run on mySQL, a pretty standard database type. A good example is that we all do the typos, right? I’m looking for ‘shirts’ but I type ‘shart’, or ‘shert’, or whatever, right? Now, on Google we’re used to it that it corrects us into whatever you wanna have, and on a lot of eCommerce stores they don’t. They’ll just tell you NO, we don’t have that. Which is …. I mean we’re talking about conversions, so it’s an important thing, right? 

Microservices Architecture for eCommerce eBook. Download for free >

– Yeah, you get frustrated if you don’t get what you want straight away.

– Exactly.

– Don’t go back to the store.

– So, we have these solutions out there. From Algolia or Clearvue to whatever is out there, that specializes just in search. BUT they don’t sell their software, they rent it. They didn’t want to give up their IP. So, they tell you: “Okay, install this plug-in or piece of JavaScript and it’s talking to our services”. Here you already see an introduction into service-oriented architecture, where you are talking to an API, an abstracted interface that code can consume. We do the same for payments. Nobody’s really handling credit cards on their own anymore, in the bigger stores maybe, but not the smaller stores. You give them off to a payment service provider and then make sure that you get the money. Why not do the same thing for your product catalog, or your checkout, or even your pricing, or stock, keeping? You wanna invest in the best of breed technology that’s out there. The decision of ‘making’ or’ buying’ is becoming more and more actually buying, and buying means a SaaS product. Everything is rented nowadays, right? You don’t really buy software anymore, and that’s a big change for the eCommerce platforms out there, especially the ones like Magento, or a bit more downmarket like WooCommerce, they traditionally have been a monolithic application made out of all these modules that you download from the internet, that might or might not be of high quality, that might or might not be actually leaking cyber information. I mean nobody knows. I think it’s a trend that will continue for at least the foreseeable future.

– You’ve been on both sides, you were developer and also on the frontside, literally, representing the brand as a salesman. How do you see the perception from the client-side? Are they willing to give away their monolithic idea of having everything under control, into just hooking up from smaller bits here and there? Is this something that people are into, od are they still not ready for it? 

– It depends a lot on the merchants. The more traditional ones want to own stuff, right, they are the ones that might even buy cars for the company instead of leasing them. They want to have this piece of software on their server, that they can go in and touch, they want to be able to move it around and almost feel it…

– It’s the ownership feeling.

– Right, and then you have a new generation of merchants that doesn’t really care about it anymore. They are used to renting stuff, they use Airbnb, they use bike sharing, they use whatever you have out there, and they’re also used to just renting a search engine. It’s perfect because you don’t have to take care of the maintenance, you don’t have to take care of upgrades. It just works and it works really well. You just need to take into account that you’re paying X amount of money every month to have that and it’s definitely something that the newer generation seems to be embracing. Also, some bigger companies are more and more investing in that kind of a solution. They see that long term investments for IT are hard to make. We have a general lifecycle of about five years I would say. After that you have to refresh your whatever backend, so why not go with the solution networks from multiple platforms that can adapt to the multiple business strategies, that you might have, because who knows where you’re gonna be at in five or ten years.

– Yeah. You’re right that the new generations are very in touch with what it means to not own a software, just rent a software or use a service. Now, for Magento obviously this is a change of paradigm. Magento used to offer everything at once? So, with this whole new wave – how is Magento turning or how is Magento adapting to this new paradigm?

– Magento saw the same thing happening. They have been moving to cloud hosting for a while, with mixed success, some might remember the period where they offered Magento hosted solutions for Magento 1. Now they are back with another Magento cloud offering, but it’s always a very monolithic cloud offering. Basically they’re hosting your eShop on Amazon, and that’s it. More recently they have become aware of this issue, and are gearing their architecture towards more service-oriented offerings. They promoted a new director of commercial architecture – Igor Miniailo – and he will start to transform that whole Magento architecture into actual services. What we call a module now will be an independent service that might still act as a part in a monolithic set-up but can be replaced by an outsourced search engine or a hosted SaaS solution.

Magento will also see some benefit there when it comes to revenue. Now they’re just giving you a big SaaS solution that costs, let’s say, $30,000 a year. They might go after even smaller or bigger merchants when they have this, let’s say, menu where you can pick, let’s say, just a very basic search that will cost me $500 a month.

– And you can’t customize it right? 

– Exactly. Then I want to have a stock keeping, which has to be like state-of-the-art, can do everything, maybe something like in B2B, right? You don’t really need a powerful search engine, in some B2B cases, because you just want this or that tie-clip, but they might offer you a very expensive or very advanced B2B stock-keeping platform.

– All right! It makes a lot of sense if companies need to evolve. It’s evolve or die. Now, I want to switch a little bit the conversation, not too far, but I want to ask you about Shopware 6. It is open source. What are the benefits of working with something like that?

– Shopware has been around for ages. It’s not as old as Magento, I’d say, but it’s getting there. It has been a closed system for a long time and more recently they’ve been starting to open source. They got onto this whole Service Oriented Architecture pretty early and that’s Shopware 6, basically. It is built as a headless system, they do offer a simple frontend that you can slap on it, they also teamed up with Vue Storefront to have a dedicated PWA-enabled headless frontend on it, but they very much focus on the backend. That’s what they’re good at, that’s what they understand, they want to do that. So, you see already that they have the service isolation where it’s easy to replace certain components, they have a very comprehensive API coverage, they have APIs that not only get and set data, but actually perform actions and that’s a big difference.

If you think of creating an order, you have to have products, you have to have an address, you have to have a shipping method, all these things are not simply set, right? Because the shipping method might be dependent on the kind of products that you have. You don’t want just APIs that set data for you, you want to have APIs that encapsulate business logic.

– That’s your product.

– Right. That’s what they’re doing very well. There’s, I think they call it alpha right now, we should see a release candidate at the end of the year, and that’s something I’m very much looking forward to. I’m already spinning it up at home to play around with it a little bit.

– Okay, Sander let’s say I’m a big player. I have tons of money. I’m ready to go to the eCommerce world…

– I’m already your best friend :D.

– … I know … :) …  well, that will be the first thing! I have a very decent budget, and I’m ready to go, but I still have no clue of what to do. So, what will be the first step, beyond the money, for me to set up a proper eCommerce platform.

– One of the first questions I would ask is “are you gonna go national or international?” That’s a big one, because you have to deal with language. Especially in Europe, its languages, its currency. Even between Germany and Austria, which are two german-speaking countries, there’s a lot of difference. Maybe Switzerland is even a better example, where you have still differences in the language…

– …French, Italian and German, and all in one country.

– Yeah. Totally different way of shipping goods. You want to choose a platform backend that can support it. Now, for example Magento would be a good solution for that, also Hybris or Demandware. You have a couple of solutions out there that might fit that. Another question that will be very important to ask is what kind of integrations are you doing. Now, you’re a bigger player, you’ll already have Salesforce in place, you might already have Pimcore in place, or Akeneo for your product management, you have a warehouse or maybe even more than one through Europe, so there’s a lot of implementations come from that. Then running in Salesforce you might want to go for Hybris by default, because you know they tell you that it comes with it, but it’s still good to do your due diligence and really look at the various options out there. Especially the cost of ownership! 

Whatever budget you have, take one third of it and put it in the bank. Keep it for the first year, especially when you’re gonna head out to new markets, you’re gonna need that money to do your marketing, keep the shop running, right!? You’re not gonna turn a profit on this new market right away, so you want to have a buffer. You don’t want to blow all your budget and have management coming after you because you still have to improve stuff after lunch. You’re going to deal with things that you didn’t foresee yet. Translations is an important part there, so whenever you have to re-publish some products or whatever you need to get translations, you need to get artwork. That’s very expensive. There you have your budget right – it’s about two thirds of what you should have available to you, and then we can start to decide on the platform that is best.

– I’ve seen it a lot here, especially here in Europe, a lot of successful companies that start with no technical team, which keeps growing, and the technical team comes later. So, let’s say I’m a CEO of a successful company with all this money. I could go two ways: I can try to hire a whole team for my company to manage all this, or I can literally just give this to someone else, to do it for me. How do you see that? As well-spent money?  What are the pros and cons of choosing one or the other? 

– That’s an interesting question. This is my opinion right, but I think that if you’re a C-level executive with a very little affinity for technology, you are gonna have to have a lot of trust in your CTO to do the right thing. Otherwise, there’s gonna be a lot of frustration, because you might not be aware of what the IT team is actually doing. IT is a lot of times “Yeah we’re working on this upgrade” and it takes ages and you can’t really see any end results from it because the frontend is still the same. What am I paying for, right? Well, with agencies you kind of get what you pay for, the way you want to see it, but they keep you more in the loop of “Okay we’re working this and we’re doing a workshop”. I’m not saying it’s cheaper, it might even be more expensive, because they’re used to entertaining the client, keeping him engaged…

– …throw a couple of buzzwords okay, they’re working…

– … and a lot of powerpoint slides. It’s a company that you know, you feel more in control because you’re paying the money and they are supposed to do what you want them to do. I think that it’s for C-level something that they have to be comfortable with, one way or the other. I’ve been in situations where I was part of an IT team where C-level might not have been fully aware of what we were doing and that’s a question of the value, of what was going on there. Whenever you need more resources they will be like “Okay, but are you gonna deliver more?”  Not per se, we just need to invest in this, right? So, there are different situations and I don’t think there’s one definitive answer to that. 

– Okay, that makes a lot of sense. Let’s say we have companies that they’re completely settled. Imagine Zapos or Zalando. The eCommerce environment is moving very fast and a lot of companies tend to chill and not do anything. For companies that are established already, what should they be looking at or what will be the steps that they are gonna take to evolve into something better be prepared?

– In a lot of companies we see that the IT team is very reactive. Logistics wants a new warehouse – OK, we’ll implement the new WMS. We’ll hook it up to the system and then we move on to the next project, because it’s already waiting. Well in Zalando, and in some other companies you get this, almost an incubator, where an engineer might try different technology on the weekend, he becomes passionate about it, he starts advocating it on the work floor and all of sudden you see a small team popping up. I think Zalando called some guilds and basically if they have two or three people that are interested in a technology, they allow them to investigate this, to start investing time in this, and I think that is what defines these companies as it drives them forward.

– Definitely! Companies need to invest in innovation.

– Right, so one of the benefits of hiring a solution provider, like a bigger agency, is they should be able to do this on their own, right? They have developers that do their hobby projects and then get a room on Friday afternoon. That’s a famous one…

– … the hackathons.

– yeah, the hackathons where they can just try out any new technology, and that’s where a lot of interesting projects come from. So, Vue Storefront, which I work for, was actually a project for one client and one of the lead developers on that invested heavily in the technology, on PWA, what it stands for, the benefits we have from it, and he basically created this whole product around it. Now it’s something that we sell separately, which is a pretty amazing journey – if you ask me. That’s one of those clear instances, were investing in technology can lead to a beautiful solution that now enables merchants with less budget to still have a PWA-enabled online store.

– That’s evolve or die. I met tons of data scientists from Zalando. You might think that these guys sell shoes, where is the place for data science, but I would say it makes a lot of sense when you think about all the innovation behind the company, and why they are where they are.

 – Yeah, I think, in general, a good way to know what kind of technology you should be investing in, is to look at other platforms that your target audience is investing time in, and then look at the issues that you might have. So, for example, why PWA is, for me, one of the hot, up-and-coming technologies for eCommerce is people are starting to get used to the more offline experience. It should be snappy, right, I mean Spotify has it, Instagram has it, so when they go on your eCommerce shop it feels clunky.

– They expect the same speed right? 

– It should be snappy. We see a lot of B2B customers that tell us “Okay we’re out in rural areas, we’re out at trade shows and the Wi-Fi is crap”. I mean, you’ve been to conferences they have for some reason like at least one dial-up connection for 500 attendees, right? 

– Yeah, like 566…

– … exactly. The guys out there selling products, pushing their services they have this iPad with them and then they’re like: “we have this new product” and it’s loading…

– … You have to wait a little bit, don’t worry

– … Yeah, it’s amazing, you’re gonna…

– … You’re gonna love it…

– …and with PWA we’re able to cache that offline just before they go into the conference, they have all the images, they have all the prices there, and they’re gonna be like: ”yeah, this is fast, right? You love this new product” or whatever.

– I guess that’s one of the things that this whole microservices environment is creating, this innovation.  Let’s have something people want, but at the same time pushes people to want things faster.

– Actually, that’s spot-on. I think that the whole Service Oriented Architecture enables you to adapt technologies faster. There’s a bigger investment upfront, where there’s more orchestration, how do you take control of deploying a certain application, or how do you do the logging. In the monolithic environment, it’s a bit easier to do it, or its feels easier maybe, when you go into services and they come with Docker and Kubernetes as hosting these solutions, there is a bit more tech involved in general, but once you have that right it’s so easy to start swapping out parts. I’ve worked for merchants and it was time to switch to a different fulfillment solution because there were more warehouses, and what we did was simply built a new application with the same API’s – backwards compatible – and add some new features for the new warehouse that the other systems didn’t need to know of, and just swapped it out in one go, and nobody noticed it even. For me that felt so powerful. Normally you would go into the system, start to swap out modules, had feature flags, …

– It’s time and money.

– And we see it with a lot of companies expanding into more international territory. Let’s say you’re webshop from Germany, right? You might be looking at the Netherlands, which is a fairly big market, so it’s safe to put down a new instance of Magento or Hybris, but maybe going to, let’s say Greece, where you have a couple of customers – you don’t know why, but let’s give it a try – you might actually want to plop down some Shopify there just to try it out. You don’t want to invest too much money, you just want to have a fairly simple solution for that. Now you’re gonna plop down this very simple solution and then have your Service Oriented backend that handles like fulfillment or payments, that you can hook up to it. I think that is very interesting, that you’re exposing your business logic to the outside, but you’re making it much more democratic, you’re enabling other smaller platforms to use it, while not having to invest big time in setting up this whole infrastructure.

– Even beyond smaller companies, beyond settled companies, everybody thinking of global expansion. Not everything is everywhere, right? It’s part of your plan, you want to be everywhere. Definitely, it’s a good step to take instead of investing all of your budget all over the world, but we’re selling still in the same places, but it is the test and the smaller services allow you to test differently.

– … also when it comes to a replatforming later on, you’re now in, let’s say ten countries, and you started 15 years ago in Germany on some IBM WebSphere. Since this is a huge system and then in other countries you weren’t able to put that down (because it’s simply too expensive) so you also have Magento. When you want to start either moving it all to one platform or you want to start replatforming some of the stuff there, it’s comfortable to have the option to do it one country after another. If you have to do a big bang, let alone the technical implications of that which is daunting, you also have the soft side, the people have to be trained with the new system – which is a challenge. I’ve been through that, I’ve done the trainings, and these are people that are not used to working with different systems, so it’s it scary for them. You have to make sure that your whole logistics is catered to, or is integrated into, this new solution, and having the comfort of this step-by-step, that’s something very valuable, I think. 

– I’m sure not only for CTOs. What they report to the CEOs is like we’re not gonna blow our budget. We’re doing everything, step by step, so it makes a lot of sense. 

– Yeah, and even being able to go to your colleagues to say: “okay, so you’re doing the logistics, don’t worry, you can still use whatever obscure fulfillment config you’re using in that country, because we’re just going to switch over this country now, and then go on to the next one”

– Sander well first of all thank you, thank you for being on the show today. Where can people find you next, I mean I know you’re basically in every Magento event so where can they find out where you’re gonna be? 

– Right, so I’m not gonna list all the conferences I am at. I would suggest people to go to Twitter, follow me there @sandermangel. I post there a lot of the events I’m going to, some articles I’m working on, in general what keeps me busy. That’s a good place to connect with me. 

– Cool! Well, thank you one more time again and thank you everybody for watching. Please don’t miss the next show because we’re gonna be on the road; we’ll be in Berlin very soon. Thank you very much Sander. 

– Thank you for having me.  See you next time. 

Reach out to Sander at Twitter: @sandermangel

Get the insights from the industry experts in the eCommerce Talks podcast.  Start listening >“></a></span><script charset=

Published July 22, 2019