Get the most from your behavioral event data in Real-Time CDP and Journey Optimizer

Transcript

Hey, everybody. Welcome to Experience League Live. My name is Daniel Wright. I am a technical marketing engineer here at Adobe. I’m excited to be hosting my first experience league live episode. Get the most from your behavioral event data in Real-Time CDP and AJO. We’re going to be talking about competed attributes and you are welcome to ask questions using the chat pod in the right hand side of the YouTube page. So go ahead, ask questions at any time and we’ll try to repeat those to our guests. And why don’t you start off if you want to just throw in there where you’re from. If you’ve used computed attributes. And why don’t we get started? I am excited to be here today with Ratchet Gupta. Oh, and Laurie Mishra. Hey, everybody. Great Ratchet. You are a product manager for a computed attributes. Why don’t you go ahead and give your give our guest your full introduction. So awesome. Yeah, I’m Ratchet. I’m a product manager on Adobe’s real time CTP and Unified provide service. I was one of the core team members involved in building this awesome feature which would help you deliver better personalized session and audience segmentation experiences by aggregating your behavioral events data. And today I’m really excited to be talking about, you know, how you can use this feature and how this can be helpful. Thank you. Awesome. Raja and I saw that your fun fact is that you’re ambidextrous. Now, what’s the most heroic thing you’ve done with your superpower? Oh, I can actually play tennis with both of my left and right hands, so that’s super cool. Oh, dangerous ratchet. The racket Racket? Yes, I am here. Laurie, please tell us a little bit about yourself. Yeah. Hey, everybody. Laurie Mishra. I’m a product marketer for Adobe Real time CDP. I’ve been at Adobe for six and a half years, which is crazy. I’m having a great time here working closely with Ratchet and bringing this functionality to market, and we’re super excited to deep dive a little bit into that today and help you get more out of the behavioral data that you have in platform. Great. And I saw your fun fact was that you once had a beer with the mayor of Chicago. Can you tell us a little bit more about that? What was going on? How many people? Just one beer or multiple And. I would have had multiple. He only wanted one, but he was campaigning and he just happened to come into a bar and restaurant that we were hanging out. Me and my friends and ended up sitting at her desk. He bought us some beers, even got his chips and glass. So, you know, he earned my vote. That’s great. And I understand this was the mayor who is known for saying to never let a serious crisis go to waste. And so so let’s use that to transition into computer attributes. What was the the problem that Adobe was trying to solve with this feature? Yeah, I think with this functionality, what we were seeing is that customers were using behavioral event data in platform for all sorts of personalization, but maybe getting access to it and doing some of the aggregation and computations they wanted was a little bit harder than it should have been for them. So that’s kind of the context we had in mind when we brought this to market. And I have a couple of slides I can show you all some of the things we heard from our customers while we were doing research in terms of what was difficult for them and what problem they wanted solved. So I think we all know this. If you’ve especially if you’ve done anything with Adobe, you’ve heard a lot about personalization, but basic personalization is not enough. And getting the most out of the behavioral data, the real time signals that you get from your customers really is the difference maker when it comes to personalization. So say you have somebody who visited your site, maybe you look at the last time that they visited your site and based on what they were looking at before, you can try to figure out what experience to serve them. But what if you could up level that impact and maybe also look at the last product that they looked at in your app in a separate device. What if you could incorporate some signals from a data partner about that same visitor and learn a little bit more about them and use that to personalize? And what if you could also look at the value of the purchases that they’ve made in the last month and use that information to pick the right products to showcase to this person when they land on your website. Of course, we all want to do this right. All of the marketers, our customers that we talk to definitely want to do all of these things quicker and more simply. But what they might be doing today to get at this information is probably a little bit more challenging than they would like. So say you’re a marketer or a data engineer or someone who is responsible for our platform and you want to summarize and use timely customer behaviors to personalize experiences. What you might be doing today is trying to get access to those behavioral insights and computations from other systems. And what that naturally introduces is some complications and latencies in how you get that information that live somewhere else. Create a computation, and then use it in marketing and advertising. Second, you want to generate actionable insights from customer behavior behaviors quickly, but maybe to do that today because of where the data sets live and the complicated workflows, you may have to call upon a data scientist, somebody who can write a cycle query to create those insights for you and then bring that back to you into our marketing system. Of course, introducing latency. Another thing that’s important with behavioral data and really the use of any customer data is being aware of the usage policies that you have for your organization. So just keeping track of the opt ins and opt outs that a customer has to say, you know, yes, you can reach out to me. And yes, these are the channels that I’m open to being contacted, but maybe not these other channels. It’s important to respect that kind of adhere to those preferences and also your larger organizational policies. But today, what might be happening is if your behavioral data is fragmented and in a couple of different systems, you have a lot of people touching that information to create aggregates and insights you can’t totally be sure about if you are able to respect the preferences of your consumers. And then lastly, you want again, that behavioral data available so that you can use it in segmentation and maybe you want to send some of those values down to other downstream systems, but it can be difficult today because you might have to be coming up with complicated segmentation rules to get at a more nuanced picture of your audiences. So to, you know, say, for example, get at the four bubbles that you see around our guy in the purple shirt. You’re maybe coming up with a few different segmentation rules to arrive at a an audience list that can meet a more complex criteria. MM I know some customers have been using the query service feature in the platform to aggregate values and output new data sets and accomplish that type of functionality. What does computed attributes bring to the picture? Yeah, great question. I think our goal is always to have workflows and options available for a lot of different personas. So like you mentioned, Daniel, we have query service and data distiller, an Experience Platform that allows a more technical persona to write SQL queries and aggregate their own behavioral insights and attach it to different datasets and then use those datasets for activation. What we wanted to do with computed attributes is make that same workflow and goal available at a profile level and in a workflow that a non-technical person could execute for. Just trying to get it, you know, as many use cases and as many personas as possible. Mm hmm. So it was, you know, with with that in mind, that context in mind, we built computed attributes and we’ve computed attributes, but we’re allowing real time CDP and Journey Optimizer customers to do is to summarize profile behavior into computed profile attributes so that they can enhance segmentation, personalization and activation use cases. Which really cool about this is these aggregates that you’re creating are stored at a person level. So say you want to know how much I spent in the last month on ever Lancome, which I kind of have a problem spending too much money there. You can calculate that exact value and attach that dollar value to my profile and use that to personalize. So the really there’s three key things here that we want to highlight. The first is, like we’ve been saying, this allows you to enhance your 1 to 1 and batch personalization use cases so you can use those actual aggregates, those actual values to determine what kind of experience somebody is going to get or what kind of experience a whole audience might get. The second is this simplifies how you might be doing segmentation today. Segmentation can be binary today, right? Somebody meets your criteria or doesn’t. What this allows you to do is incorporate the actual value into your segment creation, which means you have fewer segment logic that you have to come up with. And the third, like with anything Adobe Experience platform, all of this functionality is delivered with trust. So say you have data sets that you have applied labels to and policies to say this can’t be used for personalization. Computed attributes will respect that. That policy that you’ve set so that you can be sure that the right data sets are being used in the right channels. Hmm. Now this. For a long time that I’ve been at Adobe, I started as a Target consultant in some of those use cases you mentioned and the last slide being able to show things like or capture things in store against a profile like the last product viewed. We would use a feature called profile scripts to to accomplish that, but it is really contained to usually just a web channel. What did they do on the on the website. But with it seems like this is similar functionality in platform and since it’s at that platform level, you’re not restricted to just events that are coming from the website. You can incorporate all types of sources of event data and also take advantage of these other platform capabilities like the data governance and consent preferences, which really takes that ability to a new and new level and depth of of power. So that’s that’s really cool to see. Yeah, that’s awesome. I couldn’t have said it better myself. And it’s interesting to hear that context from you, Daniel, but what we were doing before and to your point, you know, one of the main value products of Experience Platform is being able to bring together information across channels, across datasets. So not just web, but but really anything that you think is relevant to your business leads me in nicely, actually, to a very quick visual I have. So today you have the real time customer profile. So if you’re a customer of real time CVP or Journey Optimizer, you may know that the profile is comprised of profile attributes. You know, things like maybe location, maybe gender, maybe buying preferences. And then you can also incorporate event data into those profiles. What we’re doing here is allowing you to aggregate that event data in the form of computed attributes and store those as another attribute in the profile for that person. And once you have that computed attribute, you can use it for segmentation, or you can actually also use it in audience composition, which is something we brought to market three or four months ago. You can send that specific attribute downstream to Real-Time CDP destinations. So say you calculated the number of flights somebody has taken and you know that because of the qualification criteria, there are now a gold member. You can pass on that value of gold member to a web personalization use case. So when somebody logs in or lands on a on an on your website, you can say, Hey, thanks for being a gold member. And the third is similarly, you can use that attribute for in Journey Optimizer as well. So Ratchet will demo all of these use cases so that you can kind of see how this really comes to life across the platform. I have a quick example here to sort of reiterate how this works today and why a marketer or why a business might find it useful. So we have a scenario here. We have a marketer here, Lisa, She is a marketer for a national grocery brand, and she needs to increase sales for a new vegan product line. Thanksgiving is coming up. So, you know, she’s got to get the sales up for Tofurkey and reach her vegan customers. So. Well, Lisa. The only time of year that tofurky sell. Yeah 100% which I have to be honest I made that comment about tofurkey, but I’ve never had it myself. So I don’t know what it tastes like. But it’s what’s funny is what they look like is it’s like tofu, but then it’s like pressed into a mold. So it looks like a little turkey. Yeah. So, yeah, it’s hilarious. So Lisa here has to get this sales up of Tofurkey because Thanksgiving is coming up. So what she might do is she might create a computed attribute and get a count of Google products. All of her customers have purchased in a well-defined time window. So what this will mean is that as somebody purchases vegan products at that count increases and that count is attached to a person’s profile. And because it’s attached to that person’s profile, she then chooses to use that attribute in segmentation logic. And he or she creates an audience of website visitors who have purchased more than five vegan products in the last 30 days because in her business, pretty good indicator if you bought five of any one category in the last 30 days, it means that you’re interested, you’re interested in that category. And once she has that audience created, she can then send it downstream for web personalization. And maybe she’s using Adobe Target to deliver that personalization or really any other personalization engine that she wants so that she can say, okay, once an audience qualifies for this criteria, I have said show them visuals and highlight the products from our latest vegan product line so that I can drive more sales. Cool breeze. Yeah. Pretty simple example here. Daniel, are you seeing any questions or comments? I do not see any questions or comments yet, but folks watching live, feel free to use the chat pod. Do you have any questions for Lori or Ratchet? I do have a vegan joke for you. Lori. Do you know what? Do you know what vegan zombies eat? You know. Grains. And grains. And sugar. And also since it’s sort of it’s a low pass Halloween, do you know did you know that most ghosts are vegan? Yeah. No, I did not know that. That’s right. They’re super natural. Oh, Well. Where are you with the sound effects? You told me you’re going to be ready to go. All right. Oh, okay. So we do have a question from C Jara. How does CDP radio compare to Adobe Audience Manager? Yeah, great question. So with Real-Time CDP, the focus is making sure that you have a strong first party Data Foundation Audience manager is great for use cases where you’re still utilizing third party data signals and you’re activating two cookie based destinations. But of course, with Google soon to phase out third party cookies starting in January, we think it’s important that our customers start shifting to a first party data centric strategy and Real-Time CDP allows you to easily bring in all of your first party data sources and then also bring in third party signals from from our data partners that you might be working with as well. So you can create unified audiences regardless of the type of data and then activate them out. And Real ID Journey Optimizer uses a similar framework as well so that you can centralize your first party audiences and reach out to them across engagement channels. That’s great. So why don’t we switch over to Rach and who has a demo for us so we can see what computed attributes looks like and what it can do. You ready for that? Yep. Awesome. I am ready. Just let me know if you are able to see my screen. Yep. We see it at Muslim. Awesome. It’s a lot easier. We can use cases already making me hungry anyway. So yeah. I think it was helpful to learn about how computer active you were taught. Useful by this part about having this feature as part of the platform capability, but will be good to now talk about how this actually works. And one of the key things that Laurie talked about was the aggregation of events and being able to write those aggregates directly to your profile. And we have some out of the box functions, you know, more simpler, like some more count, but also complex functions like when Macs are most recent. So just to give an example of how you could use these functions and we’ll go into the user interface in a bit, these are the functions available on the UI, so you don’t have to write any kind of sequel or even like put any complex rule. You are able to now create these rules via user interface and it’s super easy for you to now, for example, target your profiles based on the last product somebody viewed, right? So just to give an example on the most recent function, you could, you could build a rule that kind of says that, Hey, give me a for all the profiles, give me the last product they purchased in, for example, a shoe category. Right? And what it would do is it would, you know, aggregate the events for all the profiles for each profile and look at the latest shoe they bought and write that value to the profile. And now you are able to use that value either in your segmentation or in your personalization experiences. So that’s actually, you know, kind of go into the UI and see how this would work. So I right now I am in the Experience Platform page and under the profile stop you will see a new tab called computed attributes. And that’s where this feature is results. This is an entry screen where you have all the attributes that we have created and you have some of the descriptors so that you know, you know, what kind of attribute was created, When was it last evaluated, whether the last evaluation was successful, how frequently these calculations are updated and we have some of the lifecycle statuses because each computed attribute could be in either a published state, which means that it’s available on your union schema to be used, or it could be a draft, which is like this is our definition, which I’m kind of working on. I still need to work on it and, you know, get it approved before I publish it. Or it could be if you wanted it published a computed attribute, you can also deactivated, which means that you know, you don’t want to use it anymore and the deactivated computed attributes don’t contribute towards your entitlements. And then you can know kind of like when you created and when you modified the computed attributes. So let’s actually look at an example of how you can go and create computed attributes by UI. So you would essentially go head on, create computed attribute button. And once you go here, we already have one of the attributes created so we can actually have a look. So here we are trying to look off for each profile. What was the last time that transacted? Right? So here it’s about give me the last time somebody bought an apple product, right? And essentially this display name and the field name are the fields on the union schema, which will be auto created for you so that you can use them either in your segment segmentation or in the personalization. Have the second piece here is the definition piece right here. You will specify some of the filters on the events that you want to do. For example, here you want to aggregate on all the elements where the product was offering Apple. And you know, there was a patches event, right? So we having the patches there exists and the product is a five. And lastly, what you would have is something which is called the functions. So these are the five functions which we spoke about. So you have some count mid max and most recent and this specific use case because you want to look at the last time somebody transacted, you can look at the most recent function which gives you the latest value. But you have to tell give me the latest value of what? Right. So you have to basically specify the banks. I’m here and and you can take whatever timestamp you want. Here. We have taken the time stamp of the experience it and once you have specified that the the last piece is about specifying the look back. Right. So how much in the back do you want to aggregate. And these look backs are super flexible in terms of you can specify in terms of hours, days, weeks or months. And each of them, the durations have their limits, but they provide you the flexibility to do more like, you know, more frequent refreshes, which is like monthly refreshes, which are like you want to aggregate, for example, for six months and you want to refresh it on a slower frequency as compared to, for example, if I looking back in the last 24 hours and you want to refresh it on a more faster pace. So once you have actually created this computed attribute, you can either save it as a draft or you could publish it. Now, once you publish it, what happens is it creates a new field on the union schema. And basically when you go on the union schema, if you see the individual profile, this is where under the system computed attributes, you will see all these attributes which are created for you. So that’s just a quick intro to how you can, you know, go and create computed attributes on the UI will be now interesting to actually look at how these attributes can be successfully used for to unlock the use cases for your segmentation or personalization. So let me switch back to the to the next, just to give you some glimpse into how you can simplify your segmentation workflow with the behavioral data. So how you would previously do before this feature was if you have to target profiles, are audiences based on their by chase frequency or how much among the patches you would have gone on, you know, and created multiple segments which are kind of, you know, segmented by the number of times that do not how many, how much of the total value they spent. So you would end up creating multiple segments and activate each of these audiences as to your destination are Adobe Journey Optimizer journeys. Now with this feature, what essentially translates to is you can unlock the same use case with far lesser segments in a way that now you’re able to create a one segment which says, Give me all the profiles where that done at least one patches in the last 30 days and you can create essentially computed attributes which can store the values of how much the patches are, how many times they produce. And then as a single audience, you can activate not only the profiles but also the values linked to their purchases. And this simplifies your segmentation a lot. And just to give you a quick to work into how this looks like, when you actually go into your segment Builder, this is the example of a rule builder. And here you can search for computed attributes. And under the system computed attributes, you you see all the attributes that have been created and you can just drag and drop the field scale and specify the rules of how you want to target those audiences. So, for example, here we have, you know, done the pull in that attribute of some of the data used in the last six months. Are, you know, the last transaction value that they transacted on. And now you can segment based on these more granular event based data, Islam are right now. Let’s get to the more cooler part of it, which is the personalization. We have spoken about creation of a company not to be using it in a segment. The third thing is about now being able to use this to personalize better either why I describe destinations, why our target, or why Adobe Journey Optimizer. So let’s look at how things were, you know, looking like previously. So you know you have the events experience event student profile with a certain amount of detail of course, and you would run a segmentation which would be limited to the amount of data you have with the GDL and the ability to access the behavioral data, which is the events directly in your personalization is or might be limited to the segments that you are building. So in a way you are not able to directly, you might not be able to directly pull in the event data at the time of personalization. And that’s where the computer activity would feature fits in. But this feature now, even if you have a TTL applied, you can aggregate for a longer duration. And the reason is because computed attributes are incrementally calculated. So once they see a profile and entering an event and bring up the file store, the counter keeps on ticking. Which means that even if you go underneath those events and if they are already qualifying for the rule specified, it will still keep on counting and aggregating your European out. And what that translates to is that you are able to now not only segment based on a longer lookback of data, which means you are not only able to target based on, you know, like larger in size, but also you are able to now use your event data directly in your personalization and as you could have guessed, this is because with this feature you are entirely essentially transformed into an aggregated profile attribute. And since you can use these profile attributes in the pasteurization, what that means is that you are essentially using that event data directly in your customization. All right. So that involves. Like an example of using it directly in the personalization is that you can you know, if you use lawyer’s example, if they looked at the tofurky on the grocery store website and save as the last product that they viewed before they left the site and that was your computed attribute, the last product viewed you could use that in the messaging. So with AGL you could send them an email that had a picture of the beautiful pressed tofu into the turkey shape. You could show that on the website through Target when they came back to the site. Next time like that. That’s what you mean by using it directly right there. You can be very specific with those attributes and use them in the messaging. That is hard. It wasn’t Gargamel. You are spot on. You explain it way better. Yeah, exactly. So you are now able to go more granular into what exactly pertains to the extent that you cut the email, you send them out on a push notification, you send them out, you could actually send that value personalized for every profile. That’s correct. Hmm. Cool. Sorry, I didn’t mean to interrupt. Oh, I know. That was a great question. Thanks for our thanks for asking. And now actually, let’s see how what does that actually mean in the user interface? So I have a sample journey, put it up for you here. And as I usually journey flow, you have a read audience. You would basically quantify some audiences using a segmentation. You have some condition which will basically decide to split a journey and eventually you want to send out an email. Right now with this computed attributes, you can use them in multiple places here. So just to give an example, if you are splitting a journey and the Stokie example, for example, you could have a segment saying, Oh, give me all the people who did cheese in the last three dates, but in your condition, you want to now split based on the computed attribute, which was the last product patches. Now you could simply go into your add expression and under the field groups you have all these computed attributes are readily available and refreshed at the latest time for you. So now you could actually look at like for example, the last transaction last July, they shipped some of the usage and whatnot. Now in this example, I’ll find out whether you can pull the last product they are choose and selecting a journey condition. And you could now split your journey based on if they purchased product of category regen versus a product category meet. You can ask video journeys. It doesn’t end there even in your personalization and email experience. When you go and you edit your content, you are now able to put that exact product value from the computed attribute directly in your personalization experience. So that’s where the real value comes that you are now narrowing down in each part of the journey based on the events which would have been very complex and with other ways to write a sequel. Are you now to combine, get a sense or think about the user preferences? So it does all the heavy lifting at the back end. So all that portion is abstract for you, not. All right. So let me actually go to one use case. This is just one example of how a bank, you know, who was one of the read out customers for us, they use computed attributes to fulfill their personalization use case. So what they were essentially interested in is how much that each provider spend on travel in the last six months. And they created a computer. Dr. Boot, which is essentially the sum of which uses the purchase type as travel. And, and they use the computer attribute in multiple ways. One was to essentially segment their profiles based on their purchases. And by doing that, they were able to personalize on the web to display a travel oriented credit card offer because a certain profile, you know, had spent more than X amount and they feel that they might be a good candidate to offer a travel credit card, but also because computer attribute can be used directly into their personalization experience, which means that they were also able to personalize their emails by things like, Hey, thanks for being a very valuable member and your total purchases was X dollars and last 30 days, which which could potentially wander and users were like, Hey, okay, I did X amount of purchases, maybe I am Y dollars from my target and I could like become a more premium member so you could unlock those kind of use cases. They were able to unlock those kind of use cases by using this feature. All right. So Daniel, I will take a buzz to see if there are questions going in the chat. There are some questions about using this segments and Target and or analytics. You I think you mentioned how if you build a an audience in CDP, you can share that it using the computed attributes, you can share that to Adobe Target. You can also share the attribute value itself to Adobe Target also, right? That’s, that’s correct. So like any profile attribute, you can also project these attributes on the edge. These are calculated in profile in platform, but they can be used in the edge, which means that even want we target can either use a segment which is using a computer that we are that could directly use a computer attribute to personalize on the web. Mm mm. And by sharing that attribute, that’s how you can potentially show the image of the Tofurky or whatever the product was that that individual customer was you determined, was interested in. What about, what about analytics. I believe we now have the Experience Cloud audience or the Experience Cloud destination in real time. CDP Have you heard of anyone using customer attributes or an audience based on customer attributes with that destination? Yeah. So customers bring analytics data in the city all the time, right? And with this computed attribute, even, you know, this can aggregate the data, which is like you’re not a big street up and send that to the destination. And if I’m not wrong, if the analytics destination can accept an attribute, so they should be able to use that company attribute there as well. So we need to look at what the specification of the destination is and most of the CTP destinations can accept provide attributes. So yeah, they could use computed attribute as well. Potentially. Hmm. Yeah. And I feel like computed attributes were released around the same time that that new Experience Cloud Destination was released. So there might be a lot of, um, you know, new things to try out using those two features together. Can you tell us about the, the beta process for this? What, what types of things did you learn from customers during the, the beta period, what things they wanted to use this feature to do that you had been or perhaps you had been covered already? Yep. Sounds good. So we had our very great and successful bid of about ten large enterprise customers and they were across, you know, retailers, banks, media and entertainment companies. And we have seen like the value across different teams. So just to name a few, they were able to, you know, the customers were able to use their events directly in the post ization. As we mentioned, there was a use case around being able to retarget the anonymous profiles because your profile data could have, you know, your events could have like partner I.D. or they could have like, you know, other identifiers like email addresses, right. Which which are not readily available as a profile attribute on the profile. So they were able to now look at those attributes and pull that into a profile attribute. And them either through a Series B destinations or Chinese. So that was super cool. The third thing was we observed like a large retailer being able to reduce hundreds of segments into just a handful of computed attributes, say they essentially wanted to, you know, they basically built segments based on what was a lost product. Somebody touches on a lot of metrics like how frequently their watches, how recently they purchased Sprite or the total value they purchased. And now they were able to build these three or four arguably computed attributes which essentially reduce their huge workload off managing and creating these large volumes of segments. And finally, we also observed a huge value in terms of the time to value which one of the customers they were super happy with how this reduces the stress of, you know, managing your data sets are transforming them, applying the due governance labels and a lot of the, you know, heavy lifting that goes into bringing this behavior up into profile and making sure the behavioral data in the right is in the right firm to be able to be used in your customization. So now, because this entire thing is embedded in a workflow, they don’t have to allow think about like hundreds of different things. They would have to do otherwise. So yeah, that was the feedback from the beta. Hmm. So we do have a question from a room. This sounds especially relevant, you know, for something like, you know, holiday shoppers is asking, is it possible to create a custom date range in the computer to attribute, for example, a timeframe of August to September instead of having it be a rolling last two months? Yeah, I could see it being useful for, you know, the holiday shopping season. You know what compute an attribute for people, people spending from last November, December. Is that something that’s possible now or. Yeah so that’s a great question and we have heard some of their use cases where customers want to have like custom ranges and and this is part of the potential future enhancements. We we plan to do so in the current in the current iteration. What they can do is they have like a rolling window and there’s a look back. They can specify two months, three months or six months. But yeah, the specific date ranges is something which is on our mind and we plan to bring this as part of the future announcements. But thanks for bringing that to discuss. Yeah, yeah. And any other things that you have on the roadmap for the future. Oh yeah, definitely. So we are always striving to improve our capability and by no means this is the final thing. We are working on improving this further and there are like three different dimensions in which we are working on improving this. The first one being that upping the amount of the functions we have. So we do have like five powerful functions, but we can’t expand the list by doing things like being able to contact, you know, like make a list of product SKU patches instead of just like giving the last product. Right? The other thing is like in terms of the look backs how much of aggregation you could do. So things like custom date ranges, but also things like can you do like a lifetime value or can you do like longer lookback windows? Right? So getting getting there is the second damage and the third one is in terms of the data complexity. So being able to handle your complex data, which could be either in terms of arrays, it could be in forms of your relations, relational entities and stuff like that. And like I look up, you know, on my identity things so those are the things which we have heard from customers so far and we are working on bringing in the feature roadmaps. MM hmm. Yeah. And I just noticed today in the Experience Platform community section of experience, somebody asked about had an idea to extend the computed attributes features so that it supported the lookup tables or the schema relationship. So, so using the community to post those types of feature requests is is a great thing because those get funneled up to folks like Rosa and Lori who can then help prioritize those items for the product roadmap. Lori, what are what our customers missing out on by not using this feature? Well, we’ve seen really quick adoption of this. You know, we I think we released it about six weeks ago, five weeks ago. And the the trend line of adoption is is fast, which is great because I think that indicates that it’s really solving a real problem. Some of the things I think that I shared before are pretty cumbersome to do. So you might be missing out, you know, without something like computed attributes, you might be missing out on timely event data, right? By the time you have somebody in a separate system create aggregates for you, send that back to you so you can figure out what audiences to attach that to and send it for personalization. Maybe that data is stale, but with this, it’s it’s all in kind of one interface. The event data is coming in and as it’s coming in, the the counter is ticking, like Richard shared. So I think that’s from a marketer perspective, from a brand perspective, that’s probably the biggest value. And you know, there’s also some other functionalities here. Like I think the fact that you can have those left back windows means you can purge some of the less useful event data from the system or kind of reduce the bloat in the system. And that helps you with things like your license management, which which, you know, we’ve had most customers care about. So how do you kind of just keep the stuff that you need but not hang on to anything and everything because all data is not useful, so computed attributes helps you accomplish that. MM Great. So let’s, let’s wrap up this recording. Recording of today’s session will live on, on YouTube will also host it to experience leak as well and we’ll create a community post for this session. So if anybody has any additional questions that came up out of this session, you can post those there and we will respond. There is a lot of documentation for computed attributes on experience. Leak. If you just use the search feature and search for computed attributes, you’ll find that pretty quickly. And with that, we like to end with an unrelated call. And for this month’s Experience League live session, I believe that Lori has a Thanksgiving cool tip for us. Lori I do, yeah. So if you’re hosting a Thanksgiving dinner in New York, cooking a turkey, something I read on the Internet that I’ve started to do when I host dinner is I separate the skin from the bird and I crisp up the skins in a afterwards between two cookie sheet. So you get like maximum crispiness. That way the bird can cook at the temperature it needs to without burning your skin. So if you care about crispy skin, you know, there’s your tip. Sounds delicious. So I want. When should we come over? I’ll see it Thursday and. All right. All right. Thanks, everyone, for joining us today. And again, go find the community discussion and post any additional questions you have. Thank you. And thanks, Ross and Lori. Thank you. I.