Moving to a composable architecture, with Levi's Zach Crittendon
Download MP3[00:00:00] Jasmin: One of the biggest changes an enterprise can make is what we call "Going composable": moving off a highly customized, monolithic solution to a best of breed technology stack. Today we hear from a major retail brand about how they did it.
[00:00:33] Zach: I think migrating off of a monolith is a very big challenge because you have something that does everything and now you need to split it into multiple different systems that have different areas of responsibility.
[00:00:49] Jasmin: Zach Crittendon led the journey to composable at Levi's, putting in place an architecture that supports pretty mind blowing capabilities today.
[00:00:59] Zach: We had shoppable videos. They were inventory aware prices were up to date, basically two days or so of coding to actually deliver that experience.
[00:01:09] Jasmin: Zach has been an advisor for going composable for many of our customers, answering their questions and helping them figure out what to do first. Today he is sharing his advice with you.
You are listening to people Changing Enterprises. I'm Jasmine Guthmann, and please enjoy this episode with Zach Crittendon from Levi's.
You were a change agent that helped Levi's to implement some of their first composable technology. What did the environment look like when you first got there?
[00:01:42] Zach: Yeah, so, uh, Levi's journey into, uh, e-commerce has been a fairly recent thing.
As you know, the company has been around for over a hundred years. And so Levi's in venturing online started with an agency. Everything was managed, you know, outsourced. And our, including our website. So I think in 2015 is when we first brought that capability in-house. And we're now on our third iteration of implementing and running our own e-commerce storefront.
And so our first and second iterations were with a Hybris monolith. So essentially the entire website experience from the. Homepage through the buy, browse, buy, purchase, order confirmation. All of that was done in monolith and that was a great way to get started. It's an all-in-one solution. It has, you know, a lot of best practices built into it and was sufficient to get us live on the internet, but it was not going to scale with us in all of the other things that we needed to do as a company.
[00:02:48] Jasmin: And when you, when you made that decision to bring that capability in house. What drove that decision?
[00:02:57] Zach: I think Levi's saw that it was becoming more and more important to have a direct relationship with our customer. It was really important for us to have that capability in-house and not be totally dependent on a third party to build and, and operate everything.
I think the benefit of that became very clear during the pandemic when essentially brick and mortar retail completely stopped and we were left essentially only selling through our online presence. And so having built that and deployed that and having a very strong online presence was, uh, hugely important for us in being able to make it through the pandemic as well as we have.
[00:03:34] Jasmin: And has it, how painful was it to move off Hybris? Into composable technology.
[00:03:41] Zach: I think it's something, it's not easy, but it's also not unreasonably hard. I think the important thing in making the transition to composable is to make sure that it is really the correct strategic decision for your business.
Then, in some ways it makes the case for itself as you go along. So I think in our case, where we were going was an increasingly omnichannel relationship with our customers, both, uh, web, mobile app, and in-store experiences. And the only way to, you know, there are some monolithic slutions out there that combine all three of those things, but in most cases it's, they're not the best at any one of those three things.
And so being able to go with a composable approach allows us to really make sure that we're delivering the best experience for our customers possible in each of those touchpoints, and able to really control the experience as well as the interconnectivity between those experiences with the compostable architecture.
[00:04:42] Jasmin: And I'd love to dive a little deeper into that. So tell me, sure. Can you gimme an example of. Going composable has enabled you to do that was impossible before.
[00:04:53] Zach: Going to composable has made it faster and easier for us to create essentially whatever experiences our product and user experience teams are able to dream up. In a composable architecture we're able to very easily combine things and deliver to different channels. So a great example is our 150th anniversary is coming up next year, and so there's a lot of, uh, interest to celebrate that and deliver experiences that kind of highlight our history and our journey. And so we obviously wanna deliver those in different channels and be able to highlight different products, different collaborations and things like that.
And so if this was in a traditional monolithic architecture, we would have the work of constructing that experience in a way that could be delivered to an app, constructing that experience in a way that could be delivered to the website, and doing things like integrating our brand content with our product catalog and all these other sorts of things.
Being in a composable architecture has made that all of that much simpler. And so in this case, we wanted some content to be reused between basically the same content on our website and our app, and it was quite easy to set all that up because our architecture is composable. So the effort, instead of a major engineering project to get that done, it was honestly 90% of the work was in the requirements and UX design, all those sorts of things. And, uh, once those details were settled, it was really only a couple of hours to get everything set up in our CMS, Contentstack and to ensure that that was properly integrated with our product information management system and able to be displayed on our website and in our app.
[00:06:38] Jasmin: That is amazing. And it gives you a lot more freedom, a lot more time. To actually strategize and be creative and think about where you're wanting to go, rather than having to pave the road to get from A to B.
[00:06:52] Zach: Exactly, yeah. And so I think being in a composable architecture, we were able to deliver an experience like this largely by reassembling pieces of things that we've already built.
So we're able to take those UX and product requirements and essentially check a checkbox to enable those again, versus a big complicated engineering project where you have to rebuild a lot of things from scratch. And in some cases our business users are able to create new experiences completely on their own without any engineering work whatsoever.
[00:07:26] Jasmin: What are some other ways that you try to make the most of your composable technology that people new to it may not be aware of? Still being stuck in their old monolithic ways.
[00:07:39] Zach: So I think the most important thing in composable is having a very clear idea of where you're going. Make sure that you have a, a good idea of what a strong, powerful, flexible composable architecture looks like in the future.
And that could be 10 years in the future, five years in the future. Think, you know, really big picture. And far ahead. And then the flip side of that is make sure that what you're working on today actually matters for today and will have business impact for today. So ideally, as you develop these capabilities, they should be developed and delivered in a way that can be put in the hands of your business users and visible to your external customers as quickly as possible.
Ideally in weeks, maybe months at the longest. To have that done in a way that also ties into the longer term architecture. So at Levi's, when we initially developed our, uh, content management integration with our PIM product content, we actually started very simple and only delivered the capability to do the homepage in the CMS.
And within the homepage we use a modular layout system, and I think we only had four different modules that could be used on the homepage when it initially went live. So fairly small and limited, but it was sufficient to do 60% of what needs to be done on the homepage and that that in itself was a win to be able to get something like that done in a few months.
But we also had a much longer term vision that these modules would be able to not just be used on the homepage, but also within our product listing pages, within our product detail pages, in blog articles, lots of other places that this type of content could be used. And so, Ensured that in that initial architecture, we structured in, in such a way that we could both expand the number of locations that this sort of modular system could be used within our site, but also that the set of modules themselves could also be expanded.
And so since that initial launch with only four modules for our homepage, we have expanded to, I, I believe, 25 different modules. We were able to do that without in, for the most part, doing any refactoring or re-implementation because that initial very simple use case was structured in a way that was intended to grow, to fill all these other use cases.
[00:10:12] Jasmin: So does that mean that the core of four models are unchanged today?
[00:10:16] Zach: Essentially, yeah, I think migrating off of a monolith is a very big challenge because you have something that does everything and now you need to split it into multiple different systems that have different areas of responsibility.
And so if you try to do all of that at once, I'm sure someone somewhere has done that successfully, but I could see a lot of ways that that could go wrong. And in our case, we definitely didn't try to do it that way. So the way that Levi's moved from this Hybris monolith to a composable architecture was again in phases.
Phase one was, you know, very minimal simple. A headless CMS for a single use case within the monolith - we only replaced the homepage. Phase two was completely replacing the monolith for a very simple, smaller version of our website. We launched basically our entire Eastern European website initially on this new architecture. And then phase three was building on top of that simple replacement to do a full replacement of our entire website.
So I think a huge part of that transition is finding the business case that fits it.
It's a really important part of that process to find a use case that does matter to the business and will give you, you know, a real world evaluation of what you've built, what its strengths and weaknesses are, but also to do that in a way that doesn't jeopardize your main line of business, so that you can have those learnings.
If there are issues, you know, have the time to work through and identify and resolve those before you bet the business on it.
[00:11:56] Jasmin: Absolutely. What are you most proud of? If you look back on how far you've come at Levi's?
[00:12:03] Zach: I'd say one of the things I'm most proud of is that I have no idea what's on the website day-to-day.
I find out, I guess in theory, I could go in the system and see what's coming up, but for the most part I see what's on the website the same time our customers see it. And you know, there have been some really cool experiences, really cool brand moments that have been built using these tools. And you know, on a day-to-day basis, they just work. Our marketing team has for years, maybe five years, had a vision of being able to mix our content and brand experiences with our commerce experiences on our website. Your website is often segregated into three different areas. You have your homepage and your landing pages. Maybe a blog where you have more creative flexibility, more editorial control over telling specific types of stories. You have your category search and product display pages, which are very templated and focused on. Things that come out of your product information management system, and then you have the transactional parts.
And so by going with a composable architecture, we have been able to do a, I think a very good job of weaving those experiences together. So we now have a large number of places where we have what we describe as content in commerce, and then also areas where we have commerce in. So an example of content in commerce would be in some of our big collaborations we have, you know, Levi's with the Simpsons, Levi's with Nintendo or Mario.
We're able to take those, uh, product listing pages and enrich add videos, add uh, blocks of text, carousels. All sorts of different, basically any of the content that we could have on our homepage, we can have in these product listing pages, and it allows us to take these very special pages and sort of dress them up with whatever the broader brand story we're trying to tell about that.
We're also able to use that same capability to dress up our regular men's jeans, women's genes, things that are more traditional product listing pages, but places where we still want. Tell the brand story show, you know, not just a list of products, but how people are wearing the products, styling them outfits that they might be putting together.
And so thanks to the composable architecture and we're able to make it very easy for our business users to create. These experiences that combine the commerce and the content sort of in one place. And then we've also enabled the opposite, which is marketing experiences where it's, you know, very specific photo shoot or video shoot, editorial article writing that is very specific to a, a certain campaign that we're running, but then sprinkle into that experience.
Inventory aware products. And so that's a really important thing that can be very challenging, which is you have, you know, you might have a group shot of three models each wearing a full outfit, and you have then potentially, you know, five or 10 different products. In that one photo that you could highlight and say, you know, this is the hat she's wearing, these are the jeans he's wearing. This is the belt he's wearing, so on and so forth. The challenge to that is any one of those products could be sold out or no longer available, back ordered, you name it. And so if you hardcode all of that, it's basically a ticking time bomb to an experience that ends up very frustrating to a customer where they see, oh, that's the hat she's wearing. Like, really cool. I wanna buy that hat. And then you click on it and it's sold out. And so by again, having this composable architecture, we're able to create content experiences that are fully custom editorial. But also and allow the business users, uh, marketing team to tag specific parts of an image, let's say, with the products that are related to that image.
But then as that experience is delivered to our customer, we're able to to realtime calls to our product catalog. So if a product, if we know that a product is sold out, we don't highlight it as one that is, you know, featured, here's what she's wearing and how you can buy it. And so that's an example of a content in commerce experience where we're still able to not be trapped in the template of a product display page.
Instead, deliver a very rich custom bespoke brand experience, but have the commerce part of that brand experience basically just a click away. And that's something that, uh, our team had been dreaming of for years before we moved to this, uh, composable architecture. And, you know, it wasn't the first thing we delivered. It wasn't the second thing we delivered, but. Because we had that vision from the beginning that that's where we wanted our architecture to go. We were very particular in how we built all of those pieces so that when it actually came time to enable those, it was actually just a, I wanna say, uh, so when we initially delivered the commerce and content experiences.
I think that was maybe a one and a half week dev effort to do that. That was a fairly quick and easy, fairly quick and easy thing. Now, that was only because we had months of foundation we had built upon to do that. But actually that final piece on the top of saying we can put any marketing content we want in seamlessly into a product grid.
You want this content on the second row, this tile should go in the middle of the third row, any of that sort of stuff. We were able to do all of that in just a week or two fairly quickly. And then actually the first time we delivered, commerce in content experiences where we basically, we had, uh, shoppable videos where it was a carousel of videos we had, and you could tag it with the products that were being featured in that video.
And those, you know, they were inventory aware prices were up to date. Currency, you know, you could. Tag the same video with the same products. And in Europe it would show British pounds for the visitors in the uk it would show Euros for visitors in France, so on and so forth. And that it was all sort of automatic because it's, you know, coming from the commerce catalog.
And so, that experience was actually able to be built as part of a hackathon, so mm-hmm. Basically two days or so of, um, of coding to actually deliver that experience. Again, building on the foundations of, of everything that had come before it. But actually that final piece ended up being quite easy because everything we had built previously was leading towards that capability.
[00:18:51] Jasmin: Very, very, very cool example. How you need to have that initial vision or strategy or long-term thinking. And then really every day have the team basically chipping away at, at the individual components or modules that will then eventually come together as this beautiful orchestrated, uh, experience like you described.
Exactly. And no wonder that, that that is what you're most proud of because you should be.
[00:19:23] Zach: I think, I have never heard anybody say this, but I think the choice of the word composable, I think is really meaningful in the sense that composition, uh, like a music composition is a series of notes and, uh, chords and those, you know, come together into bars and, uh, movements and eventually you have an entire piece.
And the choice of that word is very significant because it really is each, any note that's out of place can ruin the whole thing. But the piece itself cannot exist without its individual notes. And so, and the right note in the right place can take something in a completely different direction. So I think it's, uh, you know, each note.
An impact on its own, but is also able to sort of, uh, harmonize, combine with everything else in the piece to, you know, deliver that experience.
[00:20:20] Jasmin: Thanks for listening to People Changing Enterprises. We'll be back next week with a new episode, helping you make your mark.
![Moving to a composable architecture, with Levi's Zach Crittendon](https://img.transistor.fm/_Ns3_xt0Cn2dfCLJZPOvRRA19X54pAuudEBzjzYvw_I/rs:fill:800:800:1/q:60/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9lcGlz/b2RlLzExMzk0NDkv/MTY4MDYyOTkxNi1h/cnR3b3JrLmpwZw.webp)