Deliver the right offer at the right time with decision management

Transcript

Good morning, afternoon, evening, depending on from where you’re dialing in. Welcome to Experience League Live. It’s great to have you. I’m Sandra Ockman. I’m Senior Technical Marketing Manager on the Experience League team here at Adobe, and I will be your host today. As always, you can communicate with us in the chat on the right, ask your questions, and we’ll try to answer them. And of course, feel free to comment anytime. So let’s give this a try. Let us know from where you’re dialing in from. I’m located in San Jose, California. And yeah, I mean, today’s show is all about decision management. I have three amazing guests today. We’ll talk about the latest intelligent offer ranking capabilities. We’ll dive into adding offers to the native journey and optimizer channels like SMS, email, mobile apps. We’ll also talk about external systems and integrated applications like Adobe Target and how they play together. But before I introduce my guests, for those of you unfamiliar with Experience League, it is the top stop shop for everything self-help. You will find documentation, how-to videos, tutorials, and courses for Journey Optimizer, and of course the other Experience Cloud products on Experience League. And if you would like to continue the conversation with your peers and experts from Adobe and from us, with us, not from us, with us, after the show, visit the community forums and simply go to Experience League.Adobe.com. But stay here, do that after the show. And now without further ado, let me bring in my guests. My first guest is Jason Hickey, Principal Product Marketing Manager at Adobe and he’s joining us from Boston today. Hey Jason. Hey Sandra. Brooklyn, New York actually, not Boston, but big city East Coast. Okay. Okay. Sorry. I thought you were located in Boston. Okay. Then let’s see if I get this one right. Next, let’s bring in Brendan, Brendan Poifier. Brendan is Principal Product Marketing Manager at Adobe. And you are here to talk about… Lehigh, right? That’s right. Oh, yay. Got that one. Last but not least, and this time joining us from Las Vegas today because he’s at a conference is Ben Tepper, our Senior Technical Evangelist at Adobe. Hey Ben. What’s up? So happy to be here. Thanks to all the fans for dialing in. Yeah. Thank you for coming. And we have Christopher from Sweden. Nobody else told us where they’re from. Come on guys. Tell us where you’re from. We want to know. And well, but before we get started, let me introduce you guys a bit more, not only your titles. So Jason, you told me you went to four different elementary schools in three states over five years. Okay. Either you misbehaved really badly or why? What happened? Yeah. It’s the perks of growing up in a military family. We move around a lot and you learn how to make friends very quickly when you’re changing schools so often. Yeah. I’m a Navy brat myself. So yeah, I have almost the same amount of schools, but yeah. Wow. Cool. So, but what’s your home state? Where were you born? Well, born in California in the States, but Kentucky is kind of de facto home state in terms of where my family is and what home base means. Okay. But I’ve been in New York for 12 years now, so I’m a native New Yorker. I’m going to start claiming it. Perfect. I don’t, I have no idea why, but why based you in Boston? I’m in Boston, Sandra, maybe that’s why. Yeah. That’s why, that’s why you got me mixed up with joining from Las Vegas today. Jason, tell our audience, what do you do at Adobe? Oh gosh. Yeah. I mean, I think part of my job is just telling the world about, you know, data-driven personalization and getting people excited and about what we’re building and why we’re building it and who we’re building it for. And so doing stuff like this and, and really kind of telling this story to the world. Glad that you’re here and that you’re going to be talking to us about these topics today. Shout out to Brett for… Yes. Yes. Okay. So before we move to Brendan, I have to show you guys something. Doug. We have a producer in the background, Doug, so I might call him out. Big shout out to Doug who’s helping us today. Okay. Let me, let me run this. I hope this works. And the Brendan Poitier! Come on down! You are the first four contestants on The Price is Right. Now here is the star of The Price is Right, Bob! So my God, Brendan, you are a celebrity. The Price is Right. Okay. That was Bob Barker. So he was still hosting. So that must’ve been before June, 2007. So how did you, how was this? I mean, they, they randomly pick people from the audience, right? This is not staged or so. It’s really random. Well, yeah. Good question. They, they kind of, they do interview, the producers go through the lines of the people waiting and talk and then they pick out people based on the interviews, I guess. I don’t know. Maybe I was crazy that day or something, but they, they chose me. Unfortunately, I wasn’t able to win anything other than a Clemson prize expresso machine. Well, okay. It does help with your job nowadays, right? Yes, a espresso machine. Yeah. It’s, that would have been my next question if you actually won something, but Hey, that must’ve been an awesome experience. But Brendan, so now tell us, what do you do at Adobe? Yeah. So I’m in the product management team for Adobe Journey Optimizer and I focus specifically on decisioning. So whenever you’re trying to solve that selection problem, we’ve got multiple options to help deliver the best experience and you want to do some, some personalization on it and maybe whether that’s experimentation or ranking with some machine learning algorithms, whatever it may be, but finding that best experience or best piece of that experience is what I focus on. Oh yeah. We are, we will have a lot of questions for you and Brendan will also show us, you know, some, the latest feature in the tool, but yeah, last but not least, another celebrity. We have Ben here and I have something that I can show you for Ben as well. So Ben, you were, can I actually, can I do it? Can I do a run in like Brandon did? Do that. Okay. Yay! Come on down Ben! Come on down! Okay. Ben was an extra in the Lifetime movie when you met Gates. You played a student. You even have an IMDB, which is amazing. How did you get that role? And did you have any lines? Did you actually say anything? No, I didn’t say anything. I didn’t say anything at all. So, but I was, there was a shot. So I was, I was working and living in LA at the time and I had some friends who were working on this film and they had a bunch of us kind of come for this big student scene shot where I think it’s literally when William and Kate meet in the classroom, takes place obviously in Europe and London and England. We filmed at UCLA, so very similar. And just kind of stood there and moved around for a bit. And the funny thing was at the end, when we kind of watched the film, there was this probably like five seconds where they just have this shot of me and this girl smiling and laughing for no reason. And we’re all like, huh, that is really interesting. So yeah. That’s Kate and William. So that was, that was in 2011, right? At least. That was, you’ve done your research, Andrea, it was. Yes, I have. And I actually need to watch this movie now. I have to see you in that movie. I mean, it’s, it’s, it’s a hot topic right now, you know, William and Kate and the whole thing going on in the UK. So yeah, I definitely have to get up to speed with my knowledge and watch the movie and see. But that was, that was before you, you joined Neo Lane, right? Which was then acquired by Adobe in 2013. So Ben, tell us, what do you do nowadays with Adobe? Yeah, well, after my acting career kind of went, you know, peaked in 2011, I worked in a product marketing for Adobe Campaign and now I’m an evangelist across our DX stack. So talking to customers, showing the value across our solutions, and I’ll highlight some of that for you later today, but excited to be here. Great group of people on the call. So we’re gonna have a lot of fun. I think I am at least. I mean, your entrance was, was amazing. I really like that. And I mean, I have so many more questions about your fun facts, but let’s get. I have a lot of fun ones. Yeah. Let’s go to the real stuff. Let’s go to the things people are here for, the decision management. So Jason, Brendan, Ben, give us a bit of an insight, you know, for, for those people who are not familiar with decision management and who aren’t using it on a daily basis yet. What is decision management about? I do have a slide here if you would like to show that. Yeah, let’s, let’s do it. And you know, we want to get into the good stuff as soon as possible. So we’re going to go live in product and Brendan and Ben are going to show us some, some really cool things. But you know, I kind of wanted to contextualize journey optimizer. It’s really fun. It’s the newest product in the Adobe experience platform. It’s just over a year old. So we released this product back in market last, last June. And so it’s really exciting to be at, you know, such the early stage development of, of all of these features. And you know, we’re talking about offers and offer decisioning and decision management. So you know, I thought it would be kind of cool to just talk maybe with, with Brandon a little bit here about what do we mean when we use these words? Why did we pick these words and, and you know, how might you use them in, you know, in everyday life? I mean, I think this slide really encapsulates it for me is when we use the word offer, you know, it’s more than just a promotion or a discount, right? And so when you actually would log into the product, which we’ll do here in, you know, five minutes, you’ll see it says decision management. And we really did that purposefully, you know, because there’s all kinds of different decisions that you can make in your business. And there’s all kinds of different offers at different stages of a customer’s maturity and their lifecycle with your brand. So Brandon, why don’t you kind of talk me through, you know, your thinking process as you were sort of naming these things in the product and, and what you want to make sure gets across with with this whole concept? Yeah, thanks, Jason. So when you think about decision management, and what is how are decisions relevant to experiences, right? There’s, there’s content, obviously, and that’s what we’re going to mainly focus on right now. But there’s also when you want to send or when you want some something to be eligible, there’s also what’s the right audience, there’s, there’s many different factors on that you might want to do decisioning on. And the first thing that we focus on on was the content. But as you see, as continue to mature in journeys, you’ll see some of these other things take on decision management workflows. But specifically with the offer decisioning, what is an offer as you posed? A lot of that’s kind of the first question most people ask. And it’s, it’s, many people think of it as just the discount or promotion, but it’s much more than that. It’s really any purpose of a message you want to deliver to someone. So whether that’s a birthday message, or maybe a new product that you’re releasing, or even some loyalty program that you want to kind of nudge someone one of your customers towards to so they can get more value out of a website or a mobile app. It can be across any channel really, also, it can be on a, like I said, a website, a mobile app, an outbound message, like an email or SMS message, even a gas station pump, if you want to put an offer on a display screen, like some of our customers do. So the gas station pump is is pretty interesting. It’s not a place where you would normally think about but I mean, basically, what you’re saying is, any sort of internet connected screen or surface can be engaged with and have real time offer decisions put into them. Yeah, that’s exactly right. And when you’ve got a lot of different decisions to make, or a lot of options to display to your users, you want to make sure you’re finding the best one based on whatever your goal is. It’s not always to make a sale, it’s to make it’s to increase engagement, increase retention, or get someone to maybe watch a video, something like that. So so that’s right. So when you have the the profile, and all the data you’re collecting, within the experience platform, you can use that data to personal hyper personal, which, which offer to send to someone. And we’ll kind of get into more of that a little bit later on the different ways to do that. Yeah, and I think it’s the next good segue to the next slide, we only have two slides for you guys today, and then we’re going to jump in the product. But as we get into this one, Ben, I’d love to hear from your perspective, you’re out there traveling the world, you’re in Vegas, right now, you’re talking to customers, what are you hearing that people, you know, are looking to do with this system or what kinds of, you know, goals like, you know, we can walk through the who, what, where, when, why, how kind of framework, we’re going to spend most of today talking about how, which is in the middle here. But but let’s start with the why. And and why do companies want to do this? And and like, why are they they looking at us to help solve this problem? And, you know, what are they trying to actually, you know, improve in their businesses? Yeah, well, I think you guys had a really good point about the flexibility and the terminology of decisions, right? If, you know, I’m at a financial conference right now, financial tech conference, right? So we’re thinking about things like how do you drive loyalty with you have one account, you want to drive another one, right? And it’s that moment where you think not just necessarily, how do I get more financial value out of an individual, but how do I give that person a better experience? And so the use cases we’re seeing are, you know, yeah, I think the best offers that you see as consumers that we see, we don’t even know about necessarily, right? That’s personalization. That’s almost subtle. That’s almost kind of just so in line with what we’re looking for, whether that be at the ATM at the branch, again, I’m in that financial mindset today. But on the site in the app, you know, it’s not always as blatant as sign up now it’s sometimes as subtle as, you know, imagery or characteristics of a message that really drive that personalization home in ways that we understand an individual is looking for. And then I think the other use case that we’re constantly seeing is the testing aspect of things, right? How do people react? Does the offer or decision that we put out there resonate the way we expected it to? Do audiences perform differently than we expected them to? And taking those learnings and putting them right back into our use cases, right? So it really depends on the industry, but we’re seeing it in a way that it’s really all about, you know, what is my goal, right? And my goal is not necessarily to, you know, drive revenue. I mean, maybe it is, but my goal is to get to there by driving personalization experiences. And that’s where the decisions come into play. Yeah, no, it’s awesome. I mean, so what I’m hearing too is, and this is pretty agnostic to industry, you know, think about with an acquisition use case, it might be about finding the right incentive to get somebody to sign up for the first time or to open a checking account or to fund the account that they actually had, right? So it’s at that part of the funnel. For an existing customer who maybe has purchased something or engaged recently, but not super recently, it might just be a, you know, the goal might just be to re-engage or to increase the lifetime value. And then sometimes it’s churn reduction or prevention, right? Like, oh, we messed up. We canceled your flight. We did something that, you know, is totally on us. But what is the right sort of message to keep you happy, keep you loyal and make it right? Like, I think that that’s part of what an offer could be too, right? And what it means to make it right might be different from person to person. It might be drink vouchers for somebody. It might be free baggage for somebody else. It might be just like a mea culpa for somebody else. Right. And that could be personalized to the individual. Yeah, Jason, I think it’d be a good point that it’s like, you know, depending on what’s happening, you need to be ready as a marketer for all these different sorts of scenarios, right? You can be doing a promotion system at one moment and then something goes wrong. You kind of pivot everything across your marketing and sales efforts to align to that that moment that just happened, like a flight delay or like inventory shortage or things like that. Yeah. And then I think you go ahead, Sandra. No, no, good. No, I think it’s interesting that both you’ve been and you Brandon touched on when we were talking about what or kind of where sometimes it’s in your face, right? It’s a it’s a big banner on the homepage of the website above the fold. You can’t miss it. It’s the big rectangle box. Right. But then we kind of talked about, I think you mentioned maybe it’s the call center agent. So you know, maybe it’s not so in your face and it’s not an offer in the traditional sense. It’s actually and you know, we’re going to talk about, I think using our API is to get into some of these non visual surfaces where maybe it’s actually just a prompt in the, in the script of the agent. So they know sort of where to guide the conversation or, or what topic to bring up to you. And I think that that’s like really powerful when you start thinking about how my sort of journey to use that turn of phrase as a consumer I might interact with visual and non visual mechanisms, surfaces, screens, right? I might have the app open. I might have used the website, but I might call, I might do those in, in any kind of unpredictable order that you’re not always going to be able to predict and you can’t always map that out as a, as a step one, step two, step three. So I think that’s part of the magic, right? Is being ready in any moment on those sort of unexpected inbound requests, but, but have that same shared context across your, your interaction points. Yeah. So it’s, it’s, it’s, it’s really across all channels and in every direction where you can apply this or have to apply it actually to give a common experience to, to your customers. Yeah, that’s exactly right. So what the way we’ve built offers specifically is to address that where you create representations of each of these channels for that offer. Right? So if I, I have that, like I said, that purpose of that message, right? Maybe it’s 10% off or maybe we pay this to this new feature in our mobile app, whatever it is you want, you’d want to have a representation of that, that message that you’re trying to get out. Like you said, Jason, whether they’re calling in and you want the call center representative to have that, or whether it’s a kiosk that they’re, they’re on or a website or getting an email. So having that, the right representations across all those channels and then being able to have those all go up to that same offer for reporting purposes, for automation purposes to know which one’s performing the best is, is critical when you’re trying to make that best decision in real time. Yeah, no, that’s awesome. And I think I’m about ready to jump in the product. I’d love to see you demo it. I think the last thing I kind of want to just hit on before we do that, Brandon is a little bit of the how. And so when we launched the automation management capabilities we had some ranking logic that we had in business rules and constraints, but why don’t you tell everybody what you’ve been working on and some of it’s not even all the way released yet. It’s coming soon. And so we can kind of unveil it here, but tell us a little bit about what the last couple of months have been like for you and what you’re releasing and how sort of exciting that is to put this into the hands of our customers. Yeah, thanks. I think just a level set on how we make decisions today, you addressed some of them. There’s a manual way based on maybe your organization has specific priorities that change and you want to be able to configure that manually. You can do that through business rules, through offer priorities. We also have ranking formulas, which allow you to have dynamic weighting of those offers for each individual profile based on what you’re collecting in their profile and also based on context data that you might be sending in like the weather. To give you an example on that, if I go back to that gas station example we were talking about one of our customers, they will give an offer of a product based on the weather outside. So if it’s hot, they’re going to say, hey, come get an ice cream cone. If it’s cold, they’re going to say, come get a coffee or a hot cocoa. And if it’s somewhere in between, get a soda while you’re waiting, pumping up your gas. So you can use context data like that or also anything, loyalty status or anything that you’re collecting in the profile that you want to decision off of and maybe boost certain scores based on those profile attributes. That’s ranking formulas. But the other one that you’re kind of hinting on here is the AI ranking models that we have. And that’s what I’m going to show you in the product. How to do that. So we have, go ahead, Sandra. Let’s go in. I’m already opening the product. So yeah, that’s great. And then before you click in, I want to check in with Ben. We’re starting to talk about AI. What’s the excitement level you’re hearing about AI? What are you hearing out there and how excited are you about having some new AI features in the product? Yeah, I think it’s huge. It really helps expedite the process of decision making. And in this case, I think AI stands for always interested in a demo. So I’m excited to see this at least. Yeah. And one of the things you’ll notice that we worked really hard on is kind of trying to bring AI to the everyday marketer or business user and not have it stuck maybe behind the data scientists. So it’s available for all people based on whatever your technical aptitude is. So if you look in offers, here’s within Journey Optimizer, you’ll see under decision management here, I’ve got two different navigation items. The components are the building blocks of what I need to build for my different offers. As you’ll see for our kind of demo fictional organization, Luma, which is a fitness apparel brand, I’ve got some different placements that I’ve created of where I want these offers to show whether it’s, you know, a card carousel, maybe on the homepage or mobile app, there’s maybe a weekly email I’m sending out or a home banner. And then I can give those things tags, I can create rules to for eligibility purposes, or use segments that I’m I’ve created an AP already. And here is where I can create my different ranking logic, I have the those ranking formulas, like I mentioned. And then I’ve got AI models here. And so let me just show you how easy it is to make one of these. So I’m just gonna say this is a personalized AI ranking model for Luma. And you’ll see I’ve got a few different model types. And you’ll see this continue to expand over time. Right now we have auto optimization, which everybody has access to right now, which is using Thompson sampling to continue to as you add and take away offers to make sure we’re always delivering the right offer based on performance. And then we also have this personalized optimization. This is the one that is available now you just you need to talk to your account representative if you want us to turn it on for your organization. And this Brandon, if you go back to that list, you know, I’ve heard of these and there’s, you know, some general terminology, I’ve heard like, non contextual bandits and contextual bandits and things like that. But I think really what you’re saying here is the optimization model type is out of a set of options, it’s going to pick one best option. And then the personal is like the personalized optimization is out of a set of options. Each of those options might be the best option for somebody. Is that about right? That’s exactly right. Yep. So auto optimization is more kind of global, what’s the best for all your users, just like you said, personalized optimization is, let’s actually look at each profile, let’s do some look alike modeling based on what users have been, how they’ve been engaging with the offers, and find the best one for for the individual user that needs one, an offer. So for the personalization, or personalized optimization, that would also improve the more the the users and the customers interact with the brand. So the more information the system gets of how the user interacts, the better this model will become. Correct? Yeah, that’s right. And the more users that you have, and the different kind of profile attribute mix that each of those users have, the more targeted it’ll be able to become. And so let me show you how that works. Just one more question. You did say that the first model is available to everyone right now, the second as well, or is that something that not everybody has out of the box? So if you have access to offer decisioning or decision management, if you see that, then you have access to this first model auto optimization. The second one, we just need to turn it on for your organization. If so, read to your Adobe representative to get that activated. Okay, that’s exactly. So here, you’ll see that the optimization metric we’re using is the click to impression ratio. So the number of clicks that this offers getting based on the number of impressions or the number of times it’s being displayed. And so we’re going to optimize off of that. And I need to determine which data set or where do I want the data? Where do I want the model to be building, building from? So where where is the model going to find my clicks and my impressions. And so I can just pick a data set here. And you can have multiple data sets if you have that data in different places. It’s just telling the model where to kind of suck in that data to build itself. And then here’s as you walk through the metric right now, we’ve got it for offers clicked. That metric will also expand over time in the future. Yeah, absolutely. We understand. I think this is kind of the, this is the first, the main use cases that as I talked to customers that they’re wanting. So this was the first one we wanted to make available. But as we will expand this to be, you know, any custom event that you want to optimize off of. That being said, it’s probably important to note that when you’re working with machine learning models, it’s only as good as the data that you give it to build, right. So when you so you kind of have to think through more. You know, an example of maybe a metric that would be hard to build a model off of is retention after six months, right, you’re not going to start getting accurate data until the model has actual conversions of retention after six months, right. So you want probably more shorter term metrics, things that the model can build right away and start giving you options of the best best offer based on the data you’re sending in. A quick remark to the audience, guys, if you have any questions, put them in the chat, we will address them. So go ahead and put in any questions you have that you would like answers, we will definitely answer your questions. So don’t be shy. Okay, Brandon. Yeah, no. So the last thing here is just, I’m going to tell it what segments that I want it to be looking for from a segment membership standpoint across these profiles. And I can choose up to 50 of these different segments that I that I want it to be building off of. As you see, it says here, you can have up to 50 segments. Again, this is another area we’ll continue to expand this and to more segments and then even attributes itself, exclusion rules, things like that. But you’ll but right now, I Yeah, good. Quick question. You said up to 50 segments. I know this, this is an extremely new feature. But do we already have a feeling for what the best practices? I mean, you know, 50 segments is a lot is that what you would aim for more segments is better? Or is there? Is there a good feeling for Okay, you should probably stick with 10 or so or does it just depend? I would say for the most part, the more the better. So the more personalized we can get right. I mean, that means that you need to have actually, you know, 50 segments at least that are that you’re using as an organization. I think those segments that you use as an organization are great ones to to use in a model like this. Like if you’ve got loyalty status is people that have done x, but not y in the last 90 days, those types of things are great things to put in here. Maybe people that have been to your website three or more times in the last 30 days, things like that. segments that you find useful maybe as you have created life cycle within your organization to get people to convert on different things. I recommend using those segments. But yeah, more is always going to be better generally. I think the interesting kind of implication here is that different teams at the at the company or even across different regions, maybe it’s you know, one country versus another country where they either have, you know, whether or not it’s restrictions or just kind of internal guidelines, like you get to pick and choose. And it could be different team to team or region to region and based on the models that you want to build. Exactly. Yeah, those are great ones too. Yeah, and there’s also context data that the model also looks at out of the box. So like what channel it’s coming from, web browser information, any of that data that when you when the feedback events of the clicks in the display events come in, that automatically gets captured by your experience events you’re sending into a EP, there’s some data there that the model will automatically take to. So all that combined, we can now create this model. So I’m going to save and activate it. And now Oh, try that again. I had my browser open for a long time. So let me just create it again. This is there we go. Yeah, I just had it open a little bit too long. So that saved now. And then let me show you how you actually use it. Now that we’ve the model will start to build after I’ve created it. So where I want to use it is actually in the decision object. So if I go into my decision object, and just I’m going to click Edit here, because I’m going to add this model to this decision. But again, to level set on what’s happening, I’ve created my offers already. And I’ve created offers with representations for each of these three placements, one for the home banner, one for the offer card, that is going to be in a carousel, and also one in a weekly, weekly email. And I add my offer collection, which is all the offers that have representations that I want, as the consideration set to pick the best one. So you can see that for this collection, I’ve got six offers here. Many of our customers have hundreds, right? There’s there can be a lot more, but for this demo, I’ve created six, I can add eligibility rules that I only want certain users to see. The these offers here or I can, I’m going to leave it at none. And the here’s where I select the ranking method I was telling you about earlier where I have the offer priority, which is that manual ranking, I can use that formula, which is the dynamic expression using profile attributes as the variables in that expression. Or I can use the AI model that we just created. So add that. And then you see this is the one that we just created here. And I can choose that model. So now and then I can go ahead and finish this. And now, anywhere that this decision those three across those three placements, it’s going to use that model to determine what’s the best offer for a user. So you see, it’s actually relatively easy to use AI modeling in your decisioning for your content. That’s I mean, this is this is really awesome. How can I how can I test this? How can I see? I mean, AI is like this big black box. I mean, you throw things in and then it works with it. Can I can I somehow, you know, test the outcome? And if it’s a question, it’s doing it. Certainly, you can use, you know, the API is directly so some people like to use postman to see what one comes back, we included a simulation workflow here where you can do exactly what you’re saying you can test. So the way to do that is I’m going to use choose which test profile you want to select from. Actually, I’m not sure if I put the test profiles in here, but I’ll just talk through it. So you if you have test profiles in your data sets, then you can choose which profiles you want to simulate the offer from and then you add your decision scope, which is the placement and the decision. And then you can you click view results here and it will run the decisioning engine and give you back the best offer. We actually have another year while you’re you’re doing this. So I knew yeah. Quickly, while you asked that I’m just doing here. I’m just going to show in a different sandbox where I have test profiles. So you can see how it works. But go ahead and ask question while I’m showing this. Okay, so the question is, is there any business use case you can talk about how adding a segment in AI model can bring difference from just adding only data sets? Can you ask it one more time? It’s on the screen as well. Any business use case, can you talk about how adding a segment in AI model can bring difference from just adding only data sets? So what’s what’s the advantage to using the segments in the AI model? Yeah, data sets. I mean, so as we talked about where we would eventually like to get to, I mean, eventually, I’d like you should be able to, I think, will allow you to just look at any segment and any attribute. But as we’re iterative, we, the models, we tell the models what to look for right now. But eventually, we’d want to expand that out to be able to kind of encompass any data that you want to send in. So that’s what the segments doing is it’s saying, look at each profile, what is their segment membership that’s on that profile. And as a new user, say, Jason comes to my website needs to see an offer, I’m going to now look at his segment membership and then go match that up across the different segment mixes that I have in AEP and find the best one that has performed for that that group of profiles, and then deliver that to Jason. So and I think the data set helps you refine or define the scope of measurement, right? Like where are your metrics and how and where like, is the measurement and then the segments that you’re adding into the model is actually what the model will look at as it’s building out what we would sort of call features in an AI context. And so this is effectively feature management, where you’re deciding the features that the model will use to build its trees from and its decisions from. Yeah, I guess that’s a good point, Jason. Maybe I misunderstood the question initially, but the data sets are are really to show where the feedback data is coming from, right? So when a user has engaged with that offer or done something, saying, Hey, model, build your model off of the data in this data set. So and then the segments are something different. It’s not. That’s not the data set. That’s on the profile itself. Cool. I think I see now that we’ve built the offer, we’ve built the model. How does this sort of manifest to an end user when they’re building a journey or building a campaign? You know, if I’m actually authoring in one of our native channels or even an integrated channel. So I don’t know if that’s been you or Brandon, but I would love to see sort of what the, you know, what the payoff is of setting these up for a content marketer or an optimization manager or somebody who’s doing the sort of the authoring of experiences. I would say let’s hand over, Brandon, you’re still showing us something, right? No, I think Ben is the… Let’s hand over to Ben. Can you see my screen? Yeah. Yes, we can. All right. I love it. Love it. Okay. So yeah, you know, Brandon, thank you for kind of walking us through how you set those up. I want to give you three examples of how you can start to use those offers across channels. So remember natively Journey Optimizer has in-app messaging, push notification, SMS, email. I don’t have that much time today because Jason went way over his time. I’m just kidding. But we don’t have time today to go into every channel. So I’m just going to show you a couple of those, right? We’ll do push notification because I’ve got it up here. Any place that you are building content, you have the ability to bring an offer. So I’ll give you an example here, right? Think about a push notification. We get these all the time. We can trigger these based on a variety of behaviors. If I come into the editor for this here, right, you see obviously a bunch of different types of attributes I can bring in for my data profiles. But since the focus today is these offer decisions, I can also bring in any of those offers that are aligned to this panel of representation. So when Brandon was showing you the different offers that he had, he had web, he had carousel, he had something else mobile, anything that would align with a text-based offer, we could easily bring in here and start to personalize here, right? So if I wanted to bring in my welcome text, let’s say when someone enters a store or things like that, I can add that in here. We’re not going to, but we could do it here. So this is one way that we can start to add those in. The other way is if we did it in a text message, we can do the exact same thing, right? And the key here is let’s say your offer is 20% off for the next 24 hours, right? That is going to look very different in a text versus a push notification versus an in-app message or things like that. So what I’m going to do now though is switch over to how we can do this in an email example here. And I think, you know, I like this one because it’s very visual. Now in this case, we’re designing the email right here in Journey Optimizer. If you haven’t seen this before, there’s a variety of ways you can drag and drop to build different content, experiences, columns, etc. But right here in my content components is my offer decisions kind of widget, if you will. Now, again, we’re going to look at that exact same inventory of offers like the ones that Brandon put together a few moments ago. Again, because this is an email, anything that aligns to either text, image or HTML is going to show up here. So let’s go ahead and pick an email image and see if this is the right one here. I want to do, here we go. So let’s say I want to pick one of these offer sets here. So one of these groupings of offers that we might have, I want to do the loyalty one. And I’ll add that into my email experience here. Now what you see is that, remember each one of these offer levels, we’re using that ranking to determine who is eligible to see and exceed and see that offer. In this case, pretty simple one, loyalty based, right? I’m sending an email to, let’s say, a million people, right? Some of them are not members of our loyalty program. Some are just starting. Some have been around for a long time. I can see in preview what each one of those offers is going to look like in here as well. And you see over on the right hand side, I’m actually visually seeing that as well. So this is one example, you could add multiple of these into a single email campaign, or any of the other campaigns that you’re building as well. So that’s a couple of examples, right? Push notification in the message itself, SMS, we can do it directly in the text and then email, you know, such a visual platform, we can do it right here as well. And then one of the things I’m interested in is, as you were just sort of dragging and dropping and, you know, this is so modular, what Brandon showed and what you showed, it could be the same person that’s doing both sides of these tasks, but it also doesn’t have to be like, like, if you’re an email marketing manager, and you know, you know what offers you want to use, you don’t actually have to be the person who set them up initially and created them. What do you kind of see out there? Or, you know, is there like a center of excellence where maybe there’s, because it’s omni channel nature, there, there’s some of that centralized logic, and then the channel marketers across all these different channels, web, email, push, SMS, in app, like they can kind of take those and go run with them. Yeah, I mean, you know, it depends on the organizational structure right now, it depends on where organizational structure you’re planning to go to, because a lot of businesses that we talk to are kind of thinking about how do we how do we develop our set of a center of excellence, like you’re saying, I mean, I think the most common use case is you’ve got somebody who is building offers, and then you’ve got people who are building for channels, right? So it’s sometimes different personas, sometimes the same. I think the most important thing is that those offers are centralized, right? So that if I’m an email marketer, right, my organization is divided up in a way that we have email marketers, we have mobile marketers, we have social marketers, that for each of those organizations, I have access to the exact same offers that my counterparts and co workers do as well, right? So that no matter when let’s say Jen comes to the site or into the mobile app, she gets that same offer no matter which channel she chooses. And I think that’s what makes it kind of real unique. I don’t know if that exact answers your question, Jason. It relates to a question Thomas has in chat about, you know, can we control the number of customer touches, you know, across channels? And so, you know, that’s a pretty useful feature for a center of excellence. So I don’t know if Brandon, you want to cover it or Ben, but you know, around like, impression capping, fatigue management rules, and those kinds of things where we could actually say, how many times to show this offer totally or by channel or across channels and things like that? Yeah, let me see if I have one here off the top of my demo. But when we’re building this down here, yeah, when we’re building out an offer, it’s kind of easy at the bottom of my screen here, you can define what is the priority of that offer being shown? And then at what point in time do we stop showing it based on the number of impressions that we’ve had of it, right? So making sure that you know, okay, we get it, Jason does not want the new shirt. Let’s move to the next kind of tiered offer based on that priority that you see and based on that we know that he’s seen it two times. So we know he’s not interested. Brandon, I’ll let you answer anything else there. Yeah, you got it. I mean, the only thing I’d add there is that you can do it across all placements, which is probably best practice most customers. Yeah, you could do it for specific placement if you wanted to. If you want to, you know, just cap it like only on email, or only on this website. Or just like I said earlier, across all the placements. If they’ve seen it two times, let’s move on to the next one. He doesn’t like the shirt. Cool. And then again, I feel like Gene might be reading our minds. Gene asks about Adobe Target. And I think that is sort of where you’re going next or soon about how the offer decisioning is actually integrated with Adobe Target. Might be cool to show that off. Perfect. Yeah, great. We have about 10 minutes. So yeah, let’s move on to… I need three to three and 2.5 more. Absolutely. So I’ll just switch over to Target. Two big integration points here. We’ll obviously dive into the offers in just a second. But I also want to highlight that those audiences that you’re building in platform, Experience Platform, do and can automatically show up here in Target as well. So you can see here, I’ve got obviously, you know, a lot of demos that we do. But you have all the different…oops. Well, you got it there. We don’t have time to go all the way through it. But you have all of your different audiences coming from Experience Platform. You can start to leverage those for your A-B multivariant or experience level testing. What I will do though, is I’ve built out a super-fine webinar demo activity. I’ve already opened it up here. So we’re in the visual editor here. What we can now do at any point in this site itself, or of course, you can bring in different sites if you want to, is we can, you know, always be able to click on the containers and add and insert before or replace. But now we always have the option of adding an offer decision here as well. So what does this mean? What does this actually look like? So if I come in here and select that hero image, let’s say it’s a pretty static image right now, but I want to link that to instead a more dynamic offer like one of the ones that we built before. So in this case, I do have this Luma hero page offer. Go ahead and insert it. And you can see it actually is a little bit different. It’s not perfectly formatted because I’m not the best at HTML. But you see here that it showed up. And what’s more is that as a user of Target here, I can actually look and see what those different representations are, right? So one key thing is that every time we’re building a collection of offers, we must have a fallback offer, right? What happens if all the criteria that we built, someone doesn’t fit that? Well, this is the default, which is the one that you’re seeing on the screen now. We also have one that can be custom for, let’s say, people who we know are interested in menswear. Again, this is an HTML offer that we built. So you’ve got all the styling, you’ve got the text, and you’ve also got the imagery in here. Same sort of example for, let’s say, a women’s banner. So you can really easily bring in any of those decisions or multiple of those decisions into any of the pages or experiences that you are building here. And if you want to get really fun and integrated, you can always go to experience targeting and bring in one of those audiences from AJO and Platform as well. So I’ll stop right there. Yeah. Yeah, we have one other question from Thomas, and Thomas asks if there is a message hierarchy that can be established within offer decisioning that allows for certain messages to be delivered while others suppressed based on priority as not to bury the target in communications. Good question. Yeah, so if I switch over to my offer library here, I do have this concept of priorities, and you can see here we do a lot of demos, so you’re not obviously going to have 10, 100 level priority ones. But this is that priority that we will determine when does that show compared to the other ones. So let’s say that Sandra is eligible for Luma women’s tops and Loosen Up, whatever that is. She’s eligible for both of those, she’ll get the women’s top one. But let’s say the capping on this is two. Once she sees that twice, she’ll move to this next one. Is that specific to the- So there’s a full interplay here between eligibility and priority, and then capping and fatigue on top of that. So when you sort of add those together, that gives you that hierarchy and gives you the ability to bounce between different offer collections and decision scopes. The other question is, is this per channel or would it be cross channel? So for example, I understand if I’m on the website, I am served this offer twice. What about email? If we’re using the same, let’s say, model, can I make sure that I’m served across channels only twice with this offer? So if I go onto the website, I see it, I receive a marketing email, I get the offer again, and then I won’t see it again. Or is this just per channel? I think Brendan was just getting at this. It can be either one of those two.

Cool. So I know we’re almost out of time and Ben, you’ve shared some really cool stuff around journey optimizers, native channels, where you author there, and then integrated channels across the Adobe ecosystem like Adobe Target. How would somebody execute that kind of gas station pump use case that we talked about? And I don’t know, Brendan, if we go back to you, but we talked a little bit about the APIs that are available. So if you’re not actually using an Adobe tool to author this decisioning, what are your options? What can we do? It might be a cool way to close it out to kind of talk about it. That’s it for me, so I’ll definitely let Brendan take it.

Brendan, I know we might not have demos ready for this, but just interested in what you can talk through on the APIs and what’s available.

Are you kidding? APIs are the best demos.

Let’s move to conversation mode, sorry. Are you still showing something on the…

Let me just talk to it for 30 seconds or so. Just a question if you’re doing something.

Yeah, if you build a representation and when you put content in your author library, you have a few different options of content type that you have. You can do HTML, you can do JSON, you can do text or images, and those images can come from our assets library that we give you out of the box, or it can be a URL that maybe you have in your digital asset management system somewhere else that you want to get that content from. So depending on what the API request needs to be, that format of that content, you just create a representation, title it for gas station pump placement, whatever you want to call it, and then you just add it as one of your representations for that offer. Remember, the important thing is the purpose of the message, and you can create however many representations for whatever channels you need, including those API only channels, and then you can just access getting the best offer from hitting the API that we have directly to populate the gas station pump or kiosk or any other IoT device. And then depending on the channel, we have the real-time API and we have batch APIs, is that correct? That’s right, yeah. So you can use, we have a few different APIs that hits the kind of the hub of our profile that has access to everything you have on the profile. We also have edge APIs that are more focused on scale and low latency for more inbound type use cases. So, and then you can, like you mentioned, we also have a batch API if you’re wanting to get the best offer or offers for all the profiles in a given segment and maybe export that to some other messaging vendor for delivery. That would be the batch use case. Cool. I think we went through it, Sandra, we got just a couple minutes left. We have it and the audience, there are no other questions at the moment, so either we overwhelmed them or we answered all the questions. This is really good. I think we went pretty much into depth in some of the topic areas as well.

So we have two minutes. We still have one segment left to cover. So the unrelated…

Ben, you wanted to tell us, give us a… Yeah. So, I mean, I don’t know about you guys, but when I’m cooking, there’s a lot of steps that I will skip because I think that they’re ridiculous or some small ingredient. But my tip is if there’s one thing that you do not skip on, it is when you are using things like almonds or hazelnuts or walnuts that you always toast them. It takes between eight to 12 minutes, depending on the nut, depending on the methodology, you can do it in the pan, you do it in the oven, always toast your nuts. Perfect. Can I get the cool tip on the cool tip? Yes, do that, Brandon. In the process of toasting them, make sure you don’t leave because they can get burnt real fast. Yeah, it’s a quick thing. Yeah, that’s a very good call out. Another tip for the cool tip, do not put any fat in the pan. If you roast them in the pan, just the pan as is, put the nuts in, no fat. We’re talking in the oven, right? Not on the stovetop? You can do either. Skillet’s a nice option, oven’s a nice option, microwave if you have to, I guess. Which could really work? Have you tried that? Microwave? No, I haven’t tried it, but you can try anything in the microwave except metal. We have a disclaimer, don’t try this at home. Yeah, you don’t know what’ll happen if you put walnuts in your microwave.

And I mean, if you’re lucky and you’re using an app for cooking and they’re using decision management, they will tell you what to do. Okay, so I wanted to thank you all for being here. This was a fun session and this was, I think we covered a lot of topics, a lot more than we were hoping for, which is great. I really enjoyed it. I hope our audience enjoyed it as well and had all their questions answered. If not, please come back to this webpage, the YouTube page. We will be posting a link to the community forum where you can ask additional questions. You can continue the conversation with us and with your peers. And yeah, big thank you to our audience. Thank you to the three of you for joining me. Thanks to Doug and the show. And I hope to see you all soon. Take care. Bye. Thank you everybody.