A Common Data Strategy for Marketing Orchestration

Kirill Epshteyn, VP of Marketing Platforms @ Macy’s

Learn how Macy’s works cross-channel with various internal marketing teams to send the right messages at the right time.

Resources Engage A Common Data Strategy for Marketing Orchestration

Video Transcript:

We're going off the cuff. Yes. All right, everybody, I'm Bryan Dunn. I run product here. With me is Karylle Epstein. Karylle, you want to introduce yourself too-- Yeah, absolutely. But first I want to say thank you [INAUDIBLE] is going to make my job here much easier. I hope it's going to go smoothly from now on. Well, I'm responsible for the marketing technology at Macy's, so everything above the content and content management. Today's marketer depends on the data, analytics capabilities, orchestrations capabilities, and delivery across all the different channels. So what marketing systems at Macy's does is provide expertise in tools, systems, process, animation, across all those different areas, and bringing them together in order to enable marketer to drive their experience and drive communication with the customer. So and-- you guys have obviously a big, established brand. You have hundreds of touchpoints with your customers, in store, on the web, digital, mobile. How are you guys thinking about mobile and how that fits in? Well, mobile is unique. Mobile is still new. Historically, probably we're not that unique. We thought about mobile as an extension of our web program and then extension of our new program. And I think we've quickly idealized that. The opportunities in the mobile space just make it unique and differentiate it from the-- or app space specifically differentiate it from everything else we had before. And that's where's the uniqueness in approach and the uniqueness to the marketing program specifically or overall mobile strategy as a whole. I will key on a couple of things that makes that uniqueness. First of all, is the frequency of the customer engagement. Actually on the flight here, I just googled and researched. And just one of the articles that I saw right away speaks to that average customer today or average app user or mobile user today interacts with their mobile phone at least 50 times a day. I don't think that any e-commerce websites can compete with those numbers. Any one of those interactions is an opportunity. It's an opportunity for marketing, organization for-- of for the brand to engage with the customer. Second, expectations, so experience expectations. There was a lot of, actually, conversation-- this concept was brought up before. The experience in our mobile space is much more personalized and quickly evolving. There is a lot of brands out there that set the bar very high by rolling out new features and capabilities continuously. And those features and capabilities quickly become a new norm in the eyes of the customer. And it is our responsibility to-- even if we are not-- if my app or Macy's app is not the one to define the new norm, we have to adapt and catch up very quickly, not to become an outlier. Basically, as the expectations become adopted by the customer, we need to meet the customer expectations. Just to give a quick example that personally for me, kind of a thumbs-- fingerprint authentication. Once I get to use it with couple of sites, I've immediately forget the corresponding passwords. All of a sudden, that constant challenge of remembering what is my password here, becomes less of the pain-point for me. Now, I expect it. Now, it became my expectation and has become a usability expectation that I dearly strive for. And the last one is personalization opportunities. Contextual signals that we can gather in a mobile space is unique, like geolocations, deep customer profile-- knowing what's in their wallet-- knowing where they are, how they navigate where they are in our stores physically. It's just new signals that we never had a ability to tap in before. So it's kind of translating an enormous opportunity for the marketer, to be-- to drive value to the customer and to drive relevance to the customer. And in with that basically monetizing the opportunity. So that was a lot, right? There's a lot of trends, there's a lot of these different signals, there's a lot of analysis. How are you-- how do you-- beyond just looking at trends, how do you understand what's working with when you-- when you're challenging yourselves to come up with whatever that contextual customer experience is? How do you know what's working and what's not working? What kind of approach do you take? Yes, Brian told me just don't give me one answer, one-word answers, when you get there, so I decided to see if it-- You scared me. Yeah, good. I wanted to scare you for a second. I guess two things-- one is very simple. We talked about the trends are changing. And then even other apps out there, other brands out there-- setting up the new bar, setting up the new expectation for us. So we need to stay current with those brands continuously, research the market space, see what's coming out there, correlate-- see if those new features of capabilities correlate to what we're trying to do with the slide, and see if we can adapt. The second thing is more complicated. It's a continuous test-and-learn strategy. This is a new space. Like, there is a lot of-- which somewhat lacks the standards or those standards that are changing so fast that it's always new. So I highly recommend-- and that's what we do internally-- is a very focused and continuous Everything we do from the marketing communications or product features perspective is being tested for the direct APIs or direct objectives that we had for this feature or campaign, but also as a broader impact. Just to give, once again, as a simple example, engagment campaign in marketing can be measured two ways, direct response, customer subscription, or obtain, or conversion, but also continuing to-- repeat views. Do we all-- just look at those two things, two-- two kind of metrics in the combination really helps us to understand opportunity for the acquisition and retention [INAUDIBLE] So as we think about our long-- test and learn strategy, we have to look at both for them to make sure that we are not driving one in jeopardy for another. So you-- the picture with the text stack that we've all-- I wish I have some talking points. Yeah, we should've thought ahead there. So it's a super-complex ecosystem, and you have all these signals. How does-- where does mobile fit within your strategy? And how does it interface with the rest of your channel? So in-store, the web-- what kinds of things are you guys thinking of as a retailer with a lot of-- a large brick and mortar presence to drive customer experience? Well, Macy's is an omnichannel retailer or omnichannel brand. So our marketing is focused from the-- designed with an omnichannel focus in mind. With that, mobile becomes both source and target. It's a source for the reach data or reach contextual data that is tied to the customer engagement in the mobile devices. Whether it's a geosignals, or profiles, wallets, and all this information that can then be gathered, merged, or resolved with the rest of the data available to the brand-- available to the enterprise based on past purchase behaviors, et cetera-- and be used to engage customers across all different channels. Whether it's ID targeting, or email, like it's in Google on and on and on. Basically, this is the richness and uniqueness of the data that helps us to be more meaningful across. But it's also the target base. If you think about-- it's another interface with the customer. And customer expects to be personalized and relevant for Macy's on the mobile devices as well. So as we think about our push programs in app programs, now we have that opportunity to get that reach data that is infused by both mobile signals and traditional signals to optimize our push messages and optimize our mobile program specifically. Got it, yeah. So that makes a lot of sense. It's super hard though, taking all the signals and putting them in the same place and really having that source of truth. How were you at Macy's dealing with collecting all those signals and aggregating them and allowing your teams-- the new slide with all the different team structures, you probably have. Macy's is a big company, right? You probably have combinations of all those things. I was trying to map myself to one of those. Didn't work? Yes, somewhere. So how are you compiling and allowing the teams that need to do messaging across channels to impact the customer? Well, we have a common data strategy. Common data strategy is focused on one, fundamental thing. Everything we collect from the customer, or everything we know-- or we learn about the customer needs to be resolved to the common customer at [INAUDIBLE].. In simple terms, everything we do we need to have a key that can later be used to the-- as a mapping mechanism to match to the Macy's customer, individual, or [INAUDIBLE] 1, 2, 3, 4-- that how this looks like in a database. You have to plan for it. So every time you set up, bring a new tool or process or you enable new platform, you have to think about how you're going to then bring this data into that common, centralized master view. Mobile profiles is great opportunity. In a mobile profile, especially when customer authenticates-- and customer authenticates on mobile more often than in other areas-- you can immediately create the link between their mobile device ID, as well as a profile ID, that can later be merged to your master ID, et cetera. So we are creating that link that later allows us to collect that information, bring it to our centralized location. It actually creates that-- what I call multi-attribute key, that now gives you one record that tells you what's your idea across all those different devices and experience that later allows me to orchestrate my marketing communication. Got it. And so-- so that was a lot on the inbound data that you're collecting. Going outbound, how you have this hub where you have the rich customer profiles that you understand across all the touch points, you still have to coordinate the messaging across everything. So what are your strategies for what belongs with a mobile team, what messaging is unique to them, versus your web team? And what things belong on a centralized team to drive engagement? Well, you asked probably the most complicated question so far. Oh, good. You got me there. Have at it. And I wish I have a simple answer. The reality is not. We are a complicated organization, I'm sure many of you guys as well. So there is not one place where all the decisions are being made. Some teams are focused to deliver to the marking and calendar or the standard communication that's being planned ahead of time and needs to be applied to all touchpoints. Some teams are tasked with a specific activities in a mobile space, and they are looking for the quick test and learn of contextual campaigns that can drive customer engagement in their areas. I think we need to-- a marketing and technology organization-- we need to allow for both. That's what I believe. I don't think there is going to be a 100% one way or another. There is always going to be that centralized campaign management with a desire to communicate the same message or the same content across different touchpoints and stay relevant to the customer. There is always going to be a uniqueness of each touchpoint. And it's not mobile specifically. There's always uniqueness in any program, any marketing tactics out there-- and that's trying to drive the specific tactics KPIs or specific tactics test-and-learn strategy and be able to experiment and operate much faster. You have to have both. You have to have tools and processes and capabilities that-- to allow for the integration of your centralized campaign managing platform with campaigns that been set up at each touchpoints. You need to be able to synchronize the data, not to lose the common-- not to lose the view of the customer. That's where the data strategy comes into play. And you need to have vendors that allow for the seamless-- or products that you use that allows for the seamless integration of data-sharing. And so when you look for-- so you embraced the complexity and just said-- It is what it is. -- we're not going to get a single tool that is going to solve all of our problems, so we're going to have different solutions across. And it's your team's job to stitch together all of the data, all of the inputs and the outputs, to make it feel seamless to your customer. Is that an accurate representation? Yeah. And just to clarify, it's-- we embrace complexity, and we acknowledge there is going to be a central platform. But there is also going to be the capability at each case to deliver outside of that central platform. So it's not just a completely distributed model. It's a still centralized model with an ability to execute at the H's. So what-- can you give an example of what you might execute centrally from the platform? And what you'd expect your teams to be developing on the edges? Well, let's take some mobile example. In the mobile space, we have different types of campaigns. Some of them are just Macy's one-day sale broadcast You receive the direct mail a week before-- and don't quote me on those numbers-- you receive the email two, three days or two days ahead of the event, you receive the push message that is delivered to all your customers on the day of the event-- so everybody knows what's coming their way. That's centralized orchestration. It's been done for the essential campaign management suite. There is a lot of modeling, customer analytics, that lead to that event to make sure that that offer specific to the customer is the best offer. In-- under the circumstances-- and this is one of those-- how would I call, broadcast programs, that is essentially managed. On the flip side, well there is-- I'll just give you an example where-- tied to our, I guess, test and learn strategy. In order to collect mobile data, we need to capture customer mobile preference for the location-sharing. More customers [INAUDIBLE] more data you have, more effective you can be with your messaging. So you have-- one of the programs that we are running on the H is kind of a messaging program to highlight, or convince, the customer to share this information with us. So that does not need to be structured in the middle. It's very contextual. It's probably-- it's most often being triggered by what customers do with their mobile device or where they are in the app process. The only other piece of information that is required to where the customer [INAUDIBLE],, so you don't prompt the customer to do the same thing over and over again. So this is type of the program that is designed at the agency itself. Like, we look at different places where-- in an app-- where we can prompt customers. We look for different trigger events. We look for the different value proposition offers, so it's not just about the promotion, but its what other value we can deliver to the customer, who shares the location with us? Let's see, navigations store, all kind of additional ability to simplify [INAUDIBLE] store program, things like that. But this is a program that is fully designed on the edge. There is no need to fully synchronize into the kind of a mothership, into the central company management solution. And it's-- and because-- at the H-- gives us ability to try different offers, cadence, sequences of the messages, places in the app much more quickly and much more robustly then we would be able to do otherwise. That makes sense. And how do you-- like, because you need all the flexibility here, you're like-- you want the best of both worlds thing, where you have the touchpoints on the edges to make the most sense to live there, plus centralized campaign management. How are you choosing your vendors, both from an analytics perspective and a marketing perspective? Like, what are the traits that they need in order to facilitate you to pull in the data and push the data out and synchronize and run the kind of campaigns that you want to? Well, we may have touched on that before, but let me elaborate. I think the most-- data-sharing and data cross-pollination is probably the most important thing right now. I have zero tolerance for these data silos. If there is a data size-- doesn't matter how sophisticated and great your tool is-- if I cannot take this data and easily integrate it to the rest-- with the rest of the enterprise and vise versa, use the data from the rest of the enterprise to leverage within the tool or platform, it's a no-go. It's no-go for me, because it will never grow. It will never grow as a program, because after the first set of use cases that are unique to this tool, my inspiration, my ideas will drive me to try to see how I spend, how I can deeper integrate that tool into the overall marketing strategy. And I don't want to hit a roadblock there. So I want to be-- that's my first entrance criteria. I want to make sure there is-- I have common APIs, ability to capture data, extract data, upload data, and integrate with the rest of the marketing ecosystem. I did want to leave time for a few questions from the audience. We have a lot of people in retail. And Karylle is somebody who's dealing with-- as he's-- as we've been talking, and he explained what his job is and the complexity that goes along with having a physical presence, a web presence, a mobile presence now, and some of the touchpoints between mobile and physical, he's really got-- you got your hands full, my friend. So any questions from the audience for Karylle? Come on. Who is a retailer in here? Nobody wants to raise their hand. Nothing? Mr. Birsey. So early on you mentioned that you have two different ways you look at marketing campaigns. You have the direct impact and then the like, longtail. How do you balance with your team between those two different types of metrics, or is there like, specific campaigns that are focused on direct impact versus others that are focused on longtail? And I guess talk a little bit about your strategy there. Well, this question, a little bit outside of my direct zone of responsibility. I work with the marketers. I enable the marketing organization with the capabilities to do both. So what we usually do as an organization, we-- obviously, most often measures for the direct response, for the specific KPIs in the campaign. But what we learn time and time again, if we lose the view into the broader picture or the longitudinal impact to the key objectives, we may shooting ourselves in the foot, especially in the mobile space, which evolves very quickly. And you can truly over-personalize or over-message to the customer. And even though for that specific campaign you may get your results which you're looking for, your retention will be-- you customer retention will be impacted. So I guess the answer is you have to do both. You have to be able to measure in a point of time and see the specific objectives that you were looking for, but you also have to enable from the-- enable yourself from the beginning to be able to measure in the longitudinal view and go back and see the impact to the customer, lifetime value, et cetera. We have one over here. Hi, Karylle. I know that one of the things that I think about when I think of tech-- marketing technology at Macy's was actually an article I read a few years ago about how you are one of the first adopters to test beacons in stores and other cool mobile technology, that I think a lot of people are thinking about but maybe haven't experimented with. So I was wondering, could you just touch for a minute on how beacons are working for you and how some of this other real mobile-specific technology is working out? Well, beacon is a great idea and great opportunity. And I-- we're still exploring. We're not in a final state. It's like scratching the surface. First you put the beacons in store, then you start getting the data and signals from them. Then you're like, oh my god, what am I going to do now? So I think we're somewhere like in an advanced stage of like, oh my god, what I got-- what are we going to do now? You can-- to me, beacon signals is customer navigation path in a store. One of the-- similar to what we've been historically doing on the website by collecting [INAUDIBLE] data and seeing what the customer has shared the interest in, like understanding the customer favoritism through searching categories of products or brands, et cetera-- by looking at their navigation-- now we have the similar ability to use beacon signals and well times as customers navigate the store to derive customer preferences. I think this is big, because now those preferences can be used across any type of a communication. And with Macy's customers still being store-- there is a-- the total percentage of Macy's customer becoming omnichannel is growing and growing very fast. But there is a huge bulk of Macy's customers that are store-only customers. And those beacon signals gives us ability to communicate and know more about those customers, outside of a historical purchase behavior that we had before. So now we can really understand what customer interested in, versus what customers bought. And that's a huge opportunity. Secondarily, it's ability to drive value to the customer, help the customer to navigate the store, help customer to find that department. So there is-- it's not about promotions. It's about driving value with the customer and engagement. So once again, those beacon signals really open doors and a lot of opportunities there, that we are still exploring. All right. One more? OK, I think we're out of time, so we'll stick-- there a break up next? Yup. Break up next. Thank you. [APPLAUSE] [SIDE CONVERSATION]