Read the full transcript
Nothing like following McKinsey to get a real-world audit of what I’m about to present to you. We can grade after the fact. Very hoity title, Taking Personalization from Vision to Reality. The subtitle here could be our first year of embedding a program within the large enterprise. So I wanna walk you through kind of what our real-world scenario is, using Dynamic Yield but, in general, really thinking of personalization and all of the principles that McKinsey has laid out for us. Let’s keep going. Let’s start off. Get our kind of legs underneath us about who is URBN. So, I’ll refer to myself in URBN, not Urban Outfitters, because we are a portfolio company. The brand you’re most likely aware of are Urban Outfitters. Anthropologie and Free People are primary retail brands. We also have some special brands, BHLDN, a wedding brand. Terrain, which is a regional garden brand. And then our personal favorite in Philly is the Vetri family of restaurants, the tastiest of the group. So we’re really out there exploring different verticals. And as you can imagine then, this gives us quite an opportunity to think about the customer and what that might mean within our ecosystem. So how are we setup? It’s a little bit different just because of the multi-brand environment. The primary brands of Urban Outfitters, Anthro and Free People truly are independent brands, all a happy family, but ultimately making independent decisions about their programs, their technology over time, obviously how they’re gonna market to the customer, but we do have a centralized team. we call it a D to C team, the direct to consumer team, and that is a share resource under our chief digital officer that is really partnering with the brands to identify how are we gonna build programs like personalization, what’s the technology we might need to power that, Dynamic Yield, great example, and then what strategy. So, in a similar sense, how can we get that going? And that’s gonna be a big focus of what I wanna talk about today, is really how do we try to get that activity going within the organization. We’ll touch on process briefly. But the key dynamic here is just that this isn’t a centralized team that’s pushing the buttons of go. It really is that partnership. So each of the brand teams we are in kind of the forming, storming, norming, performing phase of how do we get groups going, the independent brand organizations to run those programs. Because as we just heard, it’s about how can we do this in the speed of business. So aligning with the marketing programs, the merchandising calendars, that’s not something that can be bouncing back from the central team. We wanna get this competency out into the brands, and the brands have amazing resources that can do this. So it’s really just helping them identify the areas and the process with which they can start to execute. So let’s go into a little bit of what we went through as we brought in Dynamic Yield. And I think for the context of the presentation I wanna be clear. Again, as we know, personalization is many things across many channels. Really a big focus for what we were doing that was a big organizational jump start was focusing on the web. We have a common platform and we really wanna do increase the capability that we could do for our customers in that channel, realizing that over time we’re gonna be able to expand that to our additional environments which is a very robust actually branded app environment. Email, as we all know, is very solid. But we really did wanna define how could we get this going on the web because it is such a primary touchpoint for us. So, the big question usually, and this relates very specifically to Dynamic Yield, why and what were we looking for? The first piece really was that audience identification. So again, at the brand level, good understanding of some key audiences that were existing in email segment, but also just wanting to know how could we access more of that customer data, but the key thing being how can we use it and enable it on our web channel. So the ability to segment and target primarily in web, but over time we wanna be able to do that across the channels. That was a key reason we started. Second piece was the experience delivery. Usually these are desperate capabilities or concept. So, you have the understanding of who that customer is, how do we do something in that moment to change that experience or deliver what is that more relevant piece. That’s a huge component of it. So those were two pieces. And the last one was, as we look across the organization, the brands historically have made decisions that work for them from technology or partners. We wanted to see if there was an ability for us to sort of normalize what is not only the technology across the brands, but that would also help us at a resource and a strategy level, so that we could have kind of a common understanding of the capabilities in a single platform that are working across the brands, as well as what does it take to staff and manage. So really being able to leverage what we can learn, and develop, and understand across the brands could easily then be applied to the sister brands that we have. So that was a piece we were looking for. I think we all probably have the same thing, too. There are point solutions that we have in either different channels or different functions. We wanted to see if we could bring, consolidate some of those so that, again, the resources could manage broader parts of the customer experience through a potential single interface, recos, messaging applications, search, those type of tools. So jumping into what were the year one goals, because it was a big ask and a big opportunity for the business, but first thing was really just doing that. How can we get this literally stood up within the brand. So, helping a team formed around that and getting a process going. So, identifying who and how those functions can work within the brands, and we’ll talk a little bit more about that in a moment. Organizing them, and then kind of from the central perspective how do we enable. So, partnering with Dynamic Yield or different services in our own engineering group. How do we enable what are these increase capabilities around identifying the customer and then delivering the experiences. And the second piece where I wanna focus will focus more time on, is the program design. So, scalable process, so important we put it in there twice, but developing the roadmap and pipeline. So the bottom line here is once we have this capability enabled, we have this team formed, how and what should we be working on? As we just heard, there are several different entry points around identifying opportunity, and we’ll walk you through how we help contextualize that for our different brands. I think I’m going pretty quick here, too. So teams, we talked about this. This is the success for us as an organization. It’s how do we have people that are dedicated and understanding to what we’re doing. The biggest thing is having a lead. A lead is not necessarily the same thing as, I would say, executive or leader in the traditional hierarchy sense. This is truly just somebody who’s thinking about the program, worrying about the program more as their primary kind of issue. It’s hard to maintain momentum if you don’t have somebody who is ultimately working to advocate and get what you need to have done. So that’s a really important piece that I think is applicable across anything as you’re trying to start it here in terms of the team. But from there, where I think we found a lot of successes, there are already so many wonderfully skilled resources within your business. It’s case of how do you form the right group to be able to execute. So in our case, it is having a leadership component so that there is an understanding that this is an important piece within your business. But at a tactical level, your web developers, people who know your site. QA, so kind of in that change management process who can help understand what’s going on. Analytics is a partner to bring insights or different things, as we talked about, who can you lean over the shoulder and look at the dashboards to get those questions done quickly. Marketing and merchandising are really the lifeblood, in our case, of what is the program, because ultimately they are setting how we as brands are going out to the customer. We have different promotions. We have different seasonal things from merchandising. We have the different case of when product is coming out. This is kind of an interesting event that I’ll talk a little bit more about the end. But a lot of how we started out and how we’re still performing is really from the business out to the customer. So in other words, our calendars are initiatives, not so much right now how are we taking the signals of the customer to understand where they’re at in the journey Honestly, I feel that’s okay, because we wanted to be in the view, in the voice of the brand to get buy in so that our teams get confidence in how this can work, and so that can kind of propel more activity. So that’s a key thing, too. What’s really been valuable for myself is looking at our three different brands. There are really different strengths amongst the teams. There’s some really good analytics folks in one team, web development folks in another, content and creative amazingly strong in another. And that’s been super helpful in having conversations amongst our portfolio of saying, “Here is why brand x is succeeding “or having more activity in one sense. “Here is why this was seen “as an opportunity within another brand.” But I think just having an understanding of where your strengths or weaknesses are is always a good piece to really look at where you can increase your team capacity. I was kind of laughing with Claire. The last question there around how would we hire if you had a choice. So, it could be just a different perspective from where we’re sitting. But I was a conference recently down at Wharton, the Wharton Customer Analytics Initiative. We did a partner program with them, their MBA crew. And this is question from people who are trying to get hired as well. And I think the perspective we have is we look at what is the trajectory of both of the role and the resource that we’re bringing in. So, in other words, the reason I’m a little more bullish on getting a bit of a generalist to begin with is there is so much competition in the market. If you wanna go out and hire a great developer right now, what does that mean? That’s a very tough thing. If you wanna go get the best analytics person or somebody who is working in BI, that is an amazingly crowded space right now. Not to say we can’t and shouldn’t, but those defined roles are very hard to fill. And so, we’re seeing there are some potential in somebody who has maybe half of the literal skills that you’re looking for now, but you know has that capacity to grow. And I think the other piece that we find really valuable about this, that role, that technology, that environment, that’s gonna change in six and 12 months anyway. So it’s who’s gonna have the capability and the capacity to grow with your expectations, with the technology infrastructure that’s gonna be building around them. So we have a little bit of a different vent on that if that’s helpful as well. And then the process. I won’t go too much into this, because I’ve not six sigma on that side. I’m learning a lot in this. The one thing I’ll say that was, I think, a real aha for us, and I was quite thankful for it, we had a great A/B testing program. Let’s call it that change management process was already embedded in the organization. Ultimately, that is a key piece that can sort of slow down that metabolic rate that was referred to in terms of if you don’t have the confidence of the business of how the change management is happening, if you don’t have the right resources on there, there’s gonna be that friction point where you’re trying to do more or enable the brands to be a little more self-sufficient. So, we’ve tried to build upon that and to do that. The one caveat I’ll have there, and we run into this at times internally ourselves is testing as a program can sort of muddy the waters a bit about how people are gonna think about what are the opportunities for personalization. Testing can at times just sort of answer a one-off question, and then you’re on to sort of a unique, separate, distinctly different question. We wanna kind of think a little more, I think horizontally in terms of how is this gonna work over time, what are opportunities that are a little longer in base than just what might be a one-time question that we can answer and move on. The other dynamic, as we talked about from a central team perspective is we, we have a different dynamic there around supporting our platform, is we have to then learn how do we intake a different set of request, and we’re still working through a lot of this candidly. Working with the outside development team. So we do get support from DY, specifically around building some different templates. I thought I love the jalopy picture that was up here. I think that is such a critical piece. How do you get things out in a proto typical or innovative way that might not be perfect but is representative of the concept? And that’s where DY has been a great partner in terms helping us take some of the templates they have, obviously, branded very specifically for us, but allows us to get in the market and understand what is the impact. But that was a new dynamic for some of that development work to happen outside of our organization. So we work through what was that development piece. And then the ultimate goal where I consider myself having success is the more that the brands can do without them ever asking me or DY or any of our partners for anything. We wanna be able to empower them with the capabilities that their team feels confident in using, and that stability comment where they’re not freaking out any of our platform engineers or partners around sites, be it in performance and those usual things. So that’s really the other piece for success that we look at centrally is are we enabling more capability over time so that the brand can move at the speed of business basically. Alright, let’s go into a little more visual concepts here. So, identifying opportunity. Same thing with the progression here. There’s so many entry points, thinking about the web, for us being a primary area. Just through thinking about how we could approach the brands of identifying the new opportunities, the key piece we are looking at is personalization is so broad. You should be able to see success in your role in almost any part of the marketing, merchandising, digital business. There’s something that you should really be able to fundamentally not just understand, but that can help promote your own success. So, a big part of where we started was simply saying, “Across our brands, “where are areas that there are pain points “but we can really help the business “understand what matters. So, certainly, with marketing teams, how can we help them with the return on ad spend? So, looking at marketing programs there. Is there an acquisition piece? Really just reframing what they’re already doing in their current roles and saying, “How can personalization, “this ability to identify that customer “and change that experience, “how can I help you do that better “and you become a hero?” Looking at a calendar, I think this was brought up a few times, but one of the most overlooked pieces is every business, whether it’s as forward looking as you like or not is sort of not the point, but the calendar is there of what you’re doing, already communicating to the customers in other channels, again, what product might be going out if you’re a retailer, that’s just a natural layout of how can you apply those concepts to it. And then the other is making it easier. It’s a very, very literal win within an organization. If you can help somebody save five minutes or push five less buttons. So, looking for opportunities there. And then this is another one. Again, going back to the concept of we’re looking at the web channel, but there are great things going on across the different marketing channel. So email has amazing segmentation and audience is going on all the time. So we’ve had great conversations of saying, “Web team, here is exactly how email “is doing this independently. “How could we apply some of that same segmentation “or some of that same communication strategy online “because now we can mirror what that segmentation is?” Retargeting principles for product. We know you have an interest and affinities develop. How can use that to modify what is our product recommendations or other messaging, and then always analytics. We have a great analytics team across most of our brands, and they’re developing insights in our CRM that allow us to understand who’s high value, who’s low value, what is the legacy purchase power of that customer. Really looking at where we’re already using the information, the insights or the programs, and just saying, “Let’s do that exact same thing “where it isn’t being done now.” And this is what some of those look like. I wanna share four concepts that I think hopefully align with what I just said there. Using data was something new for content. Event promotion is something to just really highlight how that was sort of underserved, we’ll call it area within our business, and then loyalty program as it relates to audiences and how we… That’s a good example of how we’re tying to think about that more long term and horizontally than just one off. So let’s start with the homepage audiences. Urban Outfitters. When I started with the Urban brands, I had never once heard the words, “We have too much content.” That was an amazing blessing when you think about personalization. A pain point you never heard me talk about here was creative. The creative teams across the Urban portfolio are amazing, and the quality of content that we produce is amazing. A natural entry point for us to come back to the business and say, “How can we do this better, smarter, more efficiently “while also having more impact to the customer “and increasing that?” So, from a homepage perspective on Urban Outfitters, primarily speaking, you’ll come in and you’ll see that it’s a women’s majority what we’ll call story. But the merchants and content teams do have additional stories. Our primary ones are men’s, women’s, and apartment across Urban Outfitters. So we got the thinking about how can we potentially deliver what is this additional content that exist, and make relevant rational decisions to identify the customer and deliver the right content. So we’ll see if we can get this going. Alright, so this would be what might be a native homepage. And you see that it’s kinda multi merchandise there. You had a main hero for women. You had a men’s module and a women’s module. As the customer start browsing around our site, with Dynamic Yield’s help on this, we’re developing a click by click affinity model that’s helping to start to build a rudimentary score of what is this person interested in. So, as a starting point, that could’ve been a new customer. We don’t have a perspective. But as browsing is happening, we’re starting to build up the profile. And now as they continue and go back in, we’re seeing this the second visit into men’s. There’s a very literal score that’s being developed in the backend. But you’ll see once we return to the homepage, in that session, we have updated based on the realtime intent of that customer, and we can do that now across all of our different primary categories. What’s nice is this isn’t just in the session where we’re gonna just start and build that. When you come back from this session tomorrow, that would be your default setting, as men’s, because that was the last time we saw you, we knew that was your interest, and then we’re gonna start to, again, keep evaluating what is your intent within that session. So a very focused but specific area where we solve the business opportunity of how do we tell multiple stories on our homepage, but we also we think, and we’ve seen it through the performance, we’re able to improve what is the customer satisfaction and the discoverability of the product by doing this for customers as well. Second piece, still on content. So that was the homepage where we think, obviously, a much broader exploration starting from the customer. We had a challenge at the next level down, say you go to the women’s category, or the men’s or music and tech, whatever it might be. This is truly one where you can imagine how many different lower categories there are. If you’re in men’s, you have shoes, and pants, and hoodies, and hats. Dad hats, I still haven’t figured that one out, but it’s its own area. And so the question is, the merchants all wants space, but as a website, especially on mobile web, there’s only so much room there. So as opposed to doing the affinity scoring because we saw that actually we couldn’t go down to that level of granularity and kind of keep a broad enough group, what we did was we used Dynamic Yield’s capability for just what you might call just the basic optimization. So here’s an example of the default page. You can see very modular. So remember, there were kind of those six pieces, the vintage fleece and the tearaway track pants are up top. I always forget how many clicks I have here. This could be a second session. What we’re basically doing is loading in even more content than the template supports. And through the performance of each module, it has to have specific attribution of clicking on there, is helping to determine sort of in that bandit methodology how we should merchandise the next session. So it’s learning continuously through the week. Now, what is impressive from a scale perspective, and I think one of the big wins that we’ve had, both the home page and these category pages, which there are about five or six of them are, they’re updated every single week. So we’ve gotten this to the point where the content team and the creative team, they have bought into this. They’ve seen the value where this is now part of the natural workflow where the CMS is still updated as is Dynamic Yield to make sure that we’re getting there. The big win out of this also has been the increased understanding of what is working in terms of these pages. So going back, the default, the default layout might have been New Arrivals, but being able to go back now to the business and say, “Hey, it actually was Anorak jackets this week, “and here’s the data behind it,” the creative team has really started to buy in which has been a great win for us is the creative team is now thinking in kind of that data driven way of why is this is working, who is getting there. We’re excited to really start digging into, I think, some of the audience-level understanding of who’s resonating. We’ve had recent ones where it’s been snowing in Philly or New York, and 80 degrees somewhere else. We’re starting to look at, are there regional differences that we should be understanding as well? So, California might be loving New Arrivals, whereas New York in snow was hitting more in the hoodies. So we’re starting to look at some of those segments below to understand how we can also feed that back to the business, which can be fed out to the stores to other components within our organization. Market matter. So this is kind of a straightforward one, but I think, again, a good example of applying it to a business area that didn’t always have these capabilities available. So, there are are store events for key markets, so not all product at times is in all of our markets. We wanted to increase that exposure. And then I think the key thing is a lot of times we had some systems that could do some one-off, but how can we have just a whole myriad of these in our system at once. So, some brands have a lot of different fashion shows by different markets. There’s a great, Free People has a movement program or a movement category of products that’s amazing, that’s in certain markets, and they do a lot of really nice programs there, and then there are other random things that are coming off. So our ability to get the message in front of just that customer, so as not to be distracting to those who don’t live in New York and can’t go to the hot pot style dinner, is an important factor there, and started to give, again, voice to I think the events team or different parts of the business that hadn’t been able to necessarily use the digital channel as frequently as they wanted to promote their events as well. So, loyalty. This is just kind of one piece out of how we really think the data and the understanding of customers is so paramount for web today and all the other channels to morrow. So, onsite behaviors will all get out of the box. We are loading in CRM data to Dynamic Yield. So, as we come in, our brands have different levels of recognition, but we have a pretty high stitching rate of our customers. So we can use our CRM data quite a bit. But one element that we’re passing in is some loyalty component. So, the opportunity we had there is two of our brands have loyalty programs that kind of function differently. And so, we know that that’s a primary business unit. They have teams there, so we work with them to come up, what are your goals, very obvious there around more acquisition, more usage, but we came up with one of those business use cases, not just what are we gonna do, but the business use cases around how can we push that. So, we have two examples. So, for Anthropology, a big piece is wanting to increase the enrollment of, both brands have this, but Anthropology had a real opportunity. So, typical messaging around cart thresholds you might call it. Well, through some analysis, we looked at the opportunity where there’s a free shipping threshold, but you need to be a member to sign up. And what we realize is throughout that flow, we never overtly told the customer you need to go sign up. And so here, it’s a combination of giving them the validation that they’ve hit the threshold, but also alerting them exactly what they need to do to be able to get there. So it’s hard to see, but it’s a signup and enjoy after somebody gets the success message, and that’s based on this understanding they’re not a member. If you are a members, there’s a different message that comes there. So it’s using the information of in or out of the program to begin with, and then multiple messages to try to, again, if you’re a member, it’s just gonna streamline you through the conversion. If you’re not, we want that sign up, but then the success event there for the customer is they’ll get that free shipping. Then we’ve talked about kind of stealing what might be going on in other program. So our app for Urban Outfitters is heavily utilized, and the rewards program is really robust there. Nice flashing updates of exactly where you’re at with your score. That’s an example where we’re in the process of identifying how can we bring that score to the website so that basically the web could replicate the same thing and service that the app is functioning. It’s important because with the rewards percentage, we don’t explicitly say kind of where you’re at necessarily from… Literally, you have 92 points, but it’s important because we know if you’re in a certain range, next purchase would give you what is another bonus. So it’s important for the web. Once we get this enabled, it will be able to trigger additional messaging throughout the shopping journey to say, “Don’t forget to buy as you will “hit that next tier within your rewards program, “and get the benefit that you’re after.” so increasing what is kind of the explicit understanding of the program and those benefits. So what’s next for us? We talked explicitly around the web channel. For us, we’re in the processing of getting the app enabled, and as well as the email. So we actually just had the entire product approved from DY down. Everything was in motion, but it was really helpful to, again, kind of maintain and build the momentum for us internally, hearing again how we’re now going to be able to continue a lot of the capabilities you saw work those across what is all of our channels. Something else that’s on the roadmap and we’ve just started to hit, I think, kind of a ceiling of how are we thinking about it from a reporting perspective, but I call it multi-campaign experiences. So you look at something like that homepage entry, or the gateway or something we might do in cart. We wanna be able to tether those together, not necessarily orchestrate everything. But how do we say great if we’re going after a loyalty member and I wanna have eight different experiences that they might interact with? Being able to tether together to have a clean control group and really understand the aggregate impact is something we’re really anxious for. And I know that’s a product feature that’s being delivered very soon, and we’re excited to use that. And then overall, it’s expanding the brand adoption. So, thinking about loyalty or you think about events, you think about the creative team. Those are specific constituents amongst the brands. We know there’s a lot more that we can do for our partners in marketing or in merchandising so we’re just continuing to evangelize what we can do and try to build the momentum. One of our brands had an amazing offsite where they brought all of the partners together and they just had an amazing session where they thought about the who. That was it. It wasn’t worrying about can we execute this experience or anything else. It was just a focus of what are key audiences that matter, and how can we start to think about those experiences that we wanna build for them across our different touch points. And develop customer plan. So this is what I alluded to earlier. Everything right now is very much business out to the customer. We really are looking to increase our intelligence around what is that lifecycle journey. As a multi-channel retailer, we know that the mobile web is so paramount in how it’s serving the store, how can we understand what are those activities that happen on that channel and really build things that are gonna facilitate maybe non-transactional moments for or customers as well. So, that’s something that I think is just in the infancy, but it’s really that customer empathy that were really moving to next. Right now we’re building our business empathy, so we can get that momentum going. That’s a year in the life of the URBN Empire trying to get personalization up and running from the ground. Happy to answer any questions if there’s time. If not, happy to connect. Shoot me an email. Thank you.
Nathan Richter, Head of Digital Strategy-Personalization at URBN, shares insights on how to foster a world-class personalization program and culture of continuous optimization across the company.