Composable without Compromise w/ Filip Rakowski

foreign [Music] [Music] [Music] I always like that intro so Philip how are you man welcome well I'm great and thanks so much for inviting me team it's always great to talk with you and you know this same here when we don't do it in cafes with beers we do it virtually right let's just make it happen as often as we can yeah it's 2023 so you know all you need is just uh putting on some VR Googles and we can meet in a cafe as well oh boy I have a quest too so who knows maybe we can try that I have to put it on again I haven't used it in a long time oh my friend has a VR headset but to be honest like the only things that I saw was things like beat saber or this boxing game which is terrible Graphics terrible Graphics but it's quite good for Fitness really like I had hate cardio and this is a good replacement anyway I had that game I probably should just do it I haven't really played It Anyways on to um the the business of the day so we are doing composable without compromise and we have for 2023 actually changed up our format a little bit and we came up with six questions that we're gonna ask every guest and they're going to be the same questions and then we can see how different guests actually reply to these but before we dive in can you introduce yourself who you are what you do where you live stuff like that number uh yeah so I'm free prakowski I'm a CTO and co-founder of view storefront and basically most of my career is e-commerce and VJs so if you don't know this what we start from this I think the easiest way to describe it is basically a set of tools or a platform to build e-commerce front-end so these days we no longer have this luxury of just putting kimono Leaf out of the wired with everything out of the box some call it luxury some Quality Cars right now we have to compose everything and the problem is it's very hard to craft a front-end that has like everything out of the box because there are so many vendors we're going to talk about this today anyway but there are so many vendors that is basically impossible to have something that is suiting all of those things so we start from that's exactly that we really accelerate the process of building surface fronts we mitigate the risks and basically try to do as much as possible for you yeah I think so I think it fits and what I like about this it's like from Developers for developers right and I think that's something we all want an accelerator that doesn't suck um anyways um let's go into these questions I'm gonna put a question on the screen and then we go over it and a new answer and then if anybody in the audience has a comment or a question or whatever please put it in the comments I'll put that on the screen as well so let's go to the first question so what is an unexpected or surprising outcome from transitioning from a monolith to composable architecture yeah and like the hard thing about this question is like defining what is surprising outcome because the truth is like when anyone is transitioning into composable these days you're doing this for the first time so almost everything is surprising except the things that were pitched by the vendor or by their agency but to me like from the vendor perspective the thing that was surprising that I've served on the projects is that when you're migrating from a monolith which is a single code base to composable with a dozen of code bases probably and a lot of more complexity you actually need fewer developers and the reason behind that is that most of the things that you have you're just not maintaining anymore oh lovely I think my neighbor is starting to to drill uh but there we go yeah so so that was the biggest surprise like you even though you're moving to much more complex system even though there's much more moving Parts you actually in fewer developers because the business people they are responsible for maintaining a lot of those things because they are just Services they just deliver you an API and as a developer you you don't even have access to what's inside uh on the opposite side when we take probably the most popular popularly broad example of a monolith which is Magento you just have a huge huge huge call base with dozens of modules and you know if you want to change something you're not clicking through an admin panel you have to develop this and this requires a lot of work and this is something that's you know on the intuition level you could say that moving from a simpler system to a more complex system will require more developers but it's actually the opposite wow that is an awesome answer I did not expect this and now that I think I think you're right because there's just less I don't know I'm not gonna pollute your answer with whatever I'm thinking this works oh and I see we actually have a comment so let's have a look oh there we go that is one half thank you and welcome and thank you for watching let's go back to the second question now so um what would it be if you had to advise agency or SAS company executives yeah so actually this is this is something I even had to talk about recently on V Toronto so what I see in our industry in our industry in every industry is that it is very much buzzword driven and marketing driven and you know I say don't trust buzzwords don't trust marketing claims dig deeper understand the broader context understand the nuances because the truth is this is what what matters this is what really matters like the nuances usually decide if a solution is good or not so you won't see this at first glance you won't see this by you know having a problem going to the internet it could be the problem finding a tool that is solving that like that this does not mean that it's you know a solution to all your problems there are no such things there are no things that are solution to all problems uh every solution has drawbacks really and your goal as a decision maker is not to blindly adapt a technology tempted by its potential value but more like understand those drawbacks and really carefully wait uh if the return of investment is actually worth them because uh I will give you an example from The Headless space and Magento space where I was quite active in 2017. so this is exactly the time when the term headless was actually forged and started to become mainstream and it happened because Google was pushing Progressive web apps very heavily uh and yeah they were associating Progressive web apps in general with like modern application single page applications even though it doesn't have to be like that this is how it was associated in Progressive web apps they were promising amazing performance they were promising amazing engagement and you know at that time we had a lot of people that were quite unhappy with their monoliths so imagine on the one side you have a group of people that are very unhappy with their current Solution on the other hand you have huge company with a huge reputation uh and respect saying hey this is a solution to all of your problems so a lot of companies really a lot of companies blindly migrated to headless without either even acknowledging like what is the exact value of headless architecture because they were promising mostly performance it's not performance it can be performance but it's not the biggest selling point so we had all of those agencies migrating all of those vendors migrating and then what they were realizing is that hey like you still have performance bottlenecks on uh on headless architecture what you get is actually a clean start so you can ruin it again and this is exactly what they were doing because they had zero domain expertise they just blindly adopted so the Greenfield ruining I love it yeah and what we have right now is a lot of those a lot of those Merchants a lot of those agencies they're migrating back to monolith because what they realize is that the problem was somewhere else and there are so many challenges with the Headless architecture that's migrating from the monolith to headless it does purely for performance reasons in most of the cases I'm not saying in all but in most of the cases is trying to you know is choosing the wrong solution to a problem because there's a lot of challenges a lot of drawbacks of headless architecture and I would say migrating there purely to improved performance is not worth it for most of the cases there you go um I think you speak big truth here because we're looking at like especially where I work at uniform right where we look at all these headless things and how do you actually then put them together and then still have great performance but also have content that just be happy if you don't have any sort of orchestration layer you just put stuff together in code you're building something that's a nightmare so that's exactly like now we're like two three years after the store that you were telling and it's still happening and I've seen people go back to Shopify and then actually use their out-of-the-box templates again right because it's easier somehow so you have to be quite mature to be able to do this exactly and I think it was continue hold on okay come on I mean I think it was even worse uh with that market because we were all early adopters and there was really no data to validate if this is true or not so the only way actually to have any kind of validation was trusting someone who we hope has a domain expertise which was Google at that time and you know we all see where it went honestly I I've built my fair share of pwas and we used all these kind of fun tools but the amount of smartness you had to throw at it to actually make it perform was just it was almost not worth it right all right on to the next one this one is fun so if you could wave a magic wand and change something in our space what would it be well and this question is very much connected to to the previous one actually because to me you know this whole idea behind composable architecture is absolutely great like for large e-commerces for large companies that earn a lot of money on their software even a very small adjustment can literally result in millions of dollars of Revenue so it's great that right now companies are able to really like fine tune their stock and find tools find vendors that exactly fit their needs but it comes with a cost like everything comes with a cost and this is also something I'm trying to advocate for to make a more conscious decisions and this cost is complexity so making all of those things work together uh making sure that they are not only providing a good experience when to the developers that are working in the front end but also good performance that they are connected together well that there are no security risks that everything is properly balanced regarding the infrastructure requirements whoa it's tough It's really tough and I think you know despite composable Commerce being great this is a huge huge downside but my magic want in in this case is view storefront and I think you could say almost the same about uniform like our goal is basically you know to create those orchestration layers and those accelerators and those abstractions that actually you know uh do the heavy lifting for you and allow you to really focus on getting things done and on the other hand there's also my alliance like right now I would say that the ecosystem is wild there are no standardizations there are no like you know uh contracts for how should you integrate a search provider or how should you integrate a CMS provider maybe on CMS there's some kind of consensus but I'm still seeing some weird things or there's not a lot of things even in that space to be fair yeah yeah after payments in the Commerce side they're so it's so granular that's really hard and the the how you explained it is actually kind of funny because like you said I kept thinking about the DXE and how uniform attacks this problem and it kind of simplifies it but there's there's never a fix because it's always hard because if you do an amazing orchestration with uniform but you still screw up your front end you're still not there right and if you do an amazing thing with few storefront and you screw up however all the data gets in there it's also not working so there's it's never fully correct but we can get better at it um the one I don't think we can expect from the Mac Alliance is them telling us how to do it there's not going to be any standards I think it's all to be honest I at least like the tech Council I know that they are working on some kind of standardization of course it's super early right now uh but you know I would love I would love to see such things because right now you know you you know this very well as well because we are working in the same space building very similar products it's a nightmare like you could have five payment providers and each of them is integrated in a completely different way some of them are on the front end some of them are on the back end some of them require a middleware some of them you know require you to send the pigeon with a confirmation like it's it's crazy really I've seen some crazy stuff in iframes and like oh man there's a lot there exactly yeah what you want is to like sanitize everything so it's always the same I like that yes I mean in this case you know my two magic wands are my alliance like I'm a part of the concealer and I'm pretty excited about all the things that they are doing because they are attacking real problems and another one is you know a pretty big amount of VC money that was sent to the storefront to actually solve those problems as well awesome right on to the next so what bottlenecks have you seen incomposable architectures that could be any sort of bottlenecks like form tag to process to whatever mm-hmm I I mentioned one at the beginning which is like the overall complexity and I'm kind of bringing this up quite often but I wouldn't be myself if I wouldn't bring up something different which is performance uh to be more precise front-end performance I think in a simple words it's broken these days it's broken these days because of the way how we are beginning the front ends is broken because uh of the concept of single page applications being so JavaScript heavy uh and I would love this I would love to see this being solved and luckily especially you know last year we had a wave of software that is approaching this topic from a different angle like quick quick is probably you know the best thing that would happen to web development when it comes to the performance but also Astro which you know on first glance is a meta framework like any other but at the same time they have just one difference that is making it a completely you know a game changer and the difference is is that they are going static first and dynamic by default instead of what everyone else are doing so everyone else are implying that you know every component that you have on your application it has to run JavaScript even if this is just a piece of text you have to download the component you have to execute the code Etc it costs a lot and it slows down the website a lot so you know I'm this is a bottleneck that I would like to bring up because I think it's very concerning but at the same time the future looks very bright yeah I couldn't agree more and I've I've used Astro on a few occasions now and then you can actually choose to sometimes have some JavaScript and otherwise not it's almost like going back to the old days where you just add a script that can attach some JavaScript and that's the way to go and what do you think about next js13 actually moving to react server components that are by default server rather than client isn't that the same thing or does it feel different to you I mean the concept is the same day execution is different uh I to be honest like I think I preferred the Astro approach because it's just much simpler it's yeah it's you know uh it's focused more on static of course you can have an SSR mode but in general I believe that it should all be static and it's it all can be static really uh maybe with some small exceptions but in general most of the web can be static yeah because I think um next 13 wants to be SSR Dynamic by default and static when you choose right so that's that's it is different yes likely I mean I just you know a disclaimer uh I know Nexus 13 features only from YouTube yes I haven't worked with it yet uh though I know Astro features very well so I might be favoring it just because I had such a great developer experience yeah um let's be honest same for me I had an amazing developer experience with Astro then again I've recently learned next on when it was version 12 and it is great as well anyways on to the next so what feature of old technology would you like to bring back it kicks Nostalgia for sure uh uh but you know if I would look at all the nice things I remember from the past and the things that brought me the biggest joy that I miss right now it would for sure be a weasy week building like when I was starting web development and I was searching pretty early because I was like 13 or even 12. wow even earlier I don't I don't I don't really remember because it was first yeah I think something like something something like that I was making C websites about cats but the point is I had zero knowledge about programming really like I just just downloaded Microsoft front page because you know I was the keys and I told myself I want to create a website and I started hacking with it and you know over time I was jumping to the code section more to the then today we see section just because I wanted to learn but this fun and this productivity coming from a wissy week tool it was great and this is something I would love to see more and luckily we see more of this right now it has a fancy name no code everyone are wowing everyone are wondering if this is the future but hey we had no code before but then responsive design kicked in and it became much harder to actually actually create a no-code tool that was taking this into account so we had a short break and right now it's coming back so I would love to see more no code and with cmss these days like storyblock like builder.io I think it's really possible to implement on almost any website also uh I know that uniform is cooking some things around that yeah we're coming out soon with some really interesting visual editor because people need visual stuff let's be honest it's the way to go but just walking back to like the Dreamweaver days remember it would output tables and it would like like have space or gifts and stuff like the crazy things it would do that was pretty fun man I remember these times I I remember them very well so I was using this Microsoft front page I had this you know classic uh table based website I was using actually a crop of my cat as a repeated background I remember it quite well it was fun it was a lot of fun and all of this without knowing anything about HTML CSS there you go um maybe a sub question on this let's say we are looking at tools like webflow and Builder and all these like more granular options that you get to choose from when you build with these tools do you think it goes too far for Content editors to actually work with it do you have to be a developer to get it what is your opinion here um I don't think you have to be a developer to get it I think like exactly like the purpose of this tools I think is to bring uh the possibilities that were reserved through the developers to the people that don't know how to code and I think this is great though I think that people that are making those website they should have at least amazing knowledge about programming about performance about accessibility because you know like the developers craft components and they can create great components and marketing people are using them but it doesn't mean that the website will be good if the components are good actually you could exactly for example you can have you can have a very very long list that is basically hanging your website when rendering or you could have a lot of components and each of those components is fetching data like you have to train your marketing team actually maybe not always about how things work and how to optimize it search the robot at least how to measure and how to see if something goes wrong and when they should actually reach out to the development team yeah this is kind of why I like the storyblock slash uniform approach which is not super granular but close enough to make it visual I think it's like contextual editing I think that's the safest bet here anyways go to the next this one is also fun for me so we're not um like we always ask questions like how would you explain x y c to your grandma that's we're not going to ask it like this so how would you explain composable architectures to a 15 year old who is a digital native who has Tick Tock and understands all the swipes right yeah how do you explain it to them so the final thing is I have a sister and she's 17 years old so almost and knowing how she operates uh knowing how knowledgeable she is with basically everything digital to be honest I don't think I would need to uh think a lot about fearing figuring out how to explain this to her because they know how servers work they know a lot about how the web Works which is great but you know what I would like to if I would like to explain this I would try to explain this more from the problem perspective than from the architecture perspective because technically they are much better than us like we are dinosaurs really oh yeah our let's have fun here until we can because we'll probably be products of the past very soon uh so what I would the analogy that I would like to use is hey hey it's probably a sushi set like a pre-made Sushi set I I like to make sushi from time to time although the last time was years ago I remember doing that and you basically had two options one you're going to the grocery store you're buying a pre-made Sushi set and it has everything like it has everything that you need to make sushi except the fish itself and you know if you are a sushi connoisseur like for example Avenue you really you will really notice the quality of that like you don't really you don't really have any choice about okay maybe I would like minority from a different uh a different type of nori or Nori from a different supplier or maybe I want a different price uh because it tastes better with a salmon like you can't do this you just have a one set one Sushi set it will always you know have the same taste and from one angle is great because you're just going there buying it and you can make it but if you really want to cook something that is super delicious and special and the best possible this is just not the way to go you have to buy all the ingredients separately and this is exactly what composable is to me I love it and it's actually really close to my sandwich analogy that I spoke about I think last year at some conferences the sushi is better it also fits the the generation C better I think yeah so awesome um so I have one last question and this one is going to be fun but I'm not sure how it will work here we go there's an experiment so do you have a fun question for the next guest and I actually don't know exactly who our next guest will be yet oh it will be about composable stuff right about architectures and things like that yeah so I actually had you know I was thinking yesterday about that so I would I would give this question so will content Management Systems become a competitor to the orchestration software at some point because from my perspective they are going into this direction all right great question I have answers and opinions on this already um I know there might be a oh here Sergey has um something here McDonald's as a headless Commerce That was supposed to be one of my blog posts yeah man like this analogy keeps working everywhere and I I love that we keep using it it's awesome um so when we talk about this um like will CMS actually be competitors to orchestration what about high graph there are literally are it all exactly however exactly but you should know I just had a another um live stream just before this where we were talking between the differences between going from dxp to dxc to this more composable thing and if you are a CMS and you're going into the composable space you become much more of like an API aggregator kind of thing and a composable needs a little bit more I think so they might become a BFF like a back and for front end right but that is not also a fully composable orchestration system so you need more things together to be that full thing and I think personally CMS will go into space but never fully become it because it's still a CMS at the center which is not the idea of composable yeah that's true on one hand but on the other hand if you if you take a look how the current industry is evolving like everyone at some point will have a product for the orchestration because it's unevitable like you don't want to be the part that is easy to replace right yeah yeah because for sure the cmss they're already doing this so right now imagine the the architecture in the future when we have one orchestration stacked on another orchestration on another orchestration and at the same time the companies they will be there you know they will be partners and competitors at the same time so I'm very curious how the market would solve that because it's fascinating from the partnership's perspective it's fascinating from the software perspective it's generally fascinating I think that's a great question and I will be asking it to the next person um what interests me here is um how much Freedom are they going to give when they go and do this orchestration business right I think that might be it because lots of CMS that do this now actually solve short-term problems in their CMS to have a bigger scope with an audience and become sticky but when you look at a place like uniform we don't solve the things that are tiny problems right now but the big problems for next year and so if you compare it to the uniform approach will generally be the most idealistic one and it will always work for your future but do you need it right now do you need like an in-between tool quickly right so there's going to be a very interesting thing to say yeah I agree and I also fully agree that you know the orchestration layer it shouldn't be tightly coupled to any part of your software like we are kind of moving from a moment that was platform Centric around e-commerce the moment where e-commerce is just a commodity one of the vendors that you're plugging in so right now the center is actually the orchestration maybe also the front and which kind of orchestrates but definitely not an e-commerce platform so ideally you would like to be dependent like if because this is the part that you're kind of locked in really so ideally you would like to be dependent on something that is not changing that much over time and that is not influencing the way how do you do business right e-commerce platform it is like they could change pricing they could change features they could change direction but orchestration layer like what like what would be the pivot in the future more features on top I guess yeah like extra insights or webhook specializations so we have a few um comments I'm not going to put them on oh first we have Lucy this is not the Super constructive um part to our question but what's up Lucy thank you for coming so here we have Mickey dude that is some fancy program La profile picture looking sharp my friend so um he says I think likely there will be a number of incomplete Solutions stacked together which makes a lot of sense to me as there are now right yeah exactly exactly this is what I wanted to say it is exactly like this now like you know maybe not for every software but for a large part of this like everyone are early adopters it's like an industry of startups and new companies sometimes all companies being completely new things which is kind of the same yeah and so next one from Alex um as long as a CMS is the cash cow for headless CMS Fender they have a horse in the race um defeating the idea of Fender agnostic nature or a competition composition platform yeah very true right um but what we are seeing now with uniform is that even though the cms's um don't always like us taking a bit of business from them they are also realizing now that if they want to be in the race for bigger rfps they need an orchestration part next to it so we have these partners that are sometimes annoyed and sometimes lovers and then it's always moving and then we get deals which is a very interesting approach to to this whole thing so I want to go back next year and and do this discussion again to see what changed yeah I'm also curious as I said like this is fascinating from the perspective of like there's clearly you know a lot of conflicting things on the one side a very good uh potential for partnership and doing things together on the other hand some conflicts of interest because you know the companies they're overlapping more and more they're starting in different in different places but you know as soon as they grow to a certain size they more or less overlap with a lot of features exactly um it's just going to be whoever markets it the best and who is the least Fender lock in like I think will will win this race um thank you so much for your time we did it in like 32 minutes 50 seconds which is solid it's actually how I planned it so I'm very happy that you answered the questions how I like them to be answered so how can people follow you and find you yeah uh so there are many ways to do this I'm on Twitter uh my handle on Twitter is Phil rakoski so uh oh that's that's a nice trick team I'm gonna try this as well okay go for a drink yeah I put it in my name that works yeah exactly this is what I wanted to do but okay never mind so uh you can follow me on Twitter SV rakovsky uh you can follow me on LinkedIn I'm actually starting my journey on LinkedIn right now I don't really like the things that are happening on Twitter so why not and I'm quite impressed so probably I will post more in there LinkedIn is becoming a market for me also it's very interesting yeah so we will meet each other on LinkedIn and I'm also running a developer newsletter this is I've used her from developer sentient newsletter but it doesn't mean that it only has news about the storefront uh it's on it's only Mickey rating it so I'm throwing a lot of stuff related to e-commerce to web performance and to be honest like whatever I feel I want to put there so uh if you go to my Twitter profile you will find the link for the developer newsletter as well awesome um I am subscribed so I I've read it it's awesome so everybody when you're watching subscribe um sorry why you don't advertise your own newsletter I could but maybe that's I'm not sure if that's for the stream but you can if you want so I'm I'm running a newsletter called the DXE industry newsletter that I'm running through Linkedin LinkedIn is becoming this platform for us it's great it already after the first iteration it got like 400 subscribers which blew me away it's crazy it's slowing down now but if you want that um you can find it on LinkedIn just type DXE industry newsletter um for the rest if you want to know more about uniform you can go to uniform.dev or go to uniform.2 Discord because we would love everybody to be in there so we can have a lovely community in chat and do things like that and um I think we have reached the end um thank you so much for everybody watching thank you Philip hey you changed your name there we go yeah I found it everybody take a quick peek at Philip's name and then there you go um thanks again my friend um speak soon thanks so much it was really an honor and pleasure as always to talk with you and we see each other in Amsterdam I guess bye-bye yes in like four or five weeks nice cheers man bye cheers