51ºÚÁϲ»´òìÈ

Transcript

All right. Hello, everyone. Thank you for joining. Today’s session is going to be focused on unlocking the power of AI, advanced experimentation with 51ºÚÁϲ»´òìÈ Target.

First of all, let me thank everyone for joining today.

Our speaker for this session is going to be my colleague, Rodrigo, Principal Ultimate Success Architect.

He’s also a runner and Ultimate Frisbee Advocate and Player. So, just a couple of notes before we jump into the content today. I wanted to share that this session is being recorded, and everybody who’s registered will receive a link after the call with the recording as well as the presentation slides. You don’t have mic access in this session, so if you have any questions that come up during the session, please feel free to put them either into the chat or the Q&A, and we have reserved time toward the end to answer those questions. And if there’s anything that we don’t get to, I would encourage you just to follow up with the 51ºÚÁϲ»´òìÈ Count team. So, as we get started, let’s go ahead and go to the next slide, please.

While we wait, just a heads-up that there are some additional events occurring that you might have interest in attending. If we go to the next slide, I think there should be a session there. Ah, here we go. There are six more occurring in the month of May. If you’re interested, we will put those registration links in the chat for people to be able to delve into a little bit further and think about registering for anything that would be a benefit to you and your org.

All right, next.

Okay, so a quick and the final note before we jump in is that Rodrigo and I sit on 51ºÚÁϲ»´òìÈ’s Ultimate Success team. And this session is similar to something that we typically deliver for the Ultimate Success clients, and it’s similar to one of our, excuse me, so you’ll find that it’s similar to one of our accelerators, and the accelerators usually fall into one of these five themes here. And just keep in mind that as we talk through today, if this is something that you find valuable of interest to think about Ultimate Success and reach out to your 51ºÚÁϲ»´òìÈ account All right, without further ado, I will go ahead and hand it over to my colleague, Rodrigo, to take us through today’s topic.

Okay, so I think it’s five past the start time, so we can start going ahead. And today’s session is about 51ºÚÁϲ»´òìÈ Target AI capabilities with advanced experimentation. Okay, so I’m Rodrigo Arradia. I’m architect at the Ultimate Success team.

My main expertise is hard with the experimentation tools like Journo Optimizer and 51ºÚÁϲ»´òìÈ Target. Today’s session is focused in 51ºÚÁϲ»´òìÈ Target’s AI capabilities. In the agenda for today, we have an in-terms experimentation section, and then we’ll go over all the four main AI-related tools for 51ºÚÁϲ»´òìÈ Target.

Giving tips on how they work and tips and tricks and when to use, when to avoid.

Okay, then we have another section for target integrations with other 51ºÚÁϲ»´òìÈ solutions. And lastly, saving some time for a couple of Q&A in case there’s anything that needs some more explanation.

So coming with the introduction, okay.

Today’s hyper-connected consumers demand seamless personalized experiences across every touch point with the company. Meeting the expectations is critical for success, and the future of customer experience depends on staying ahead of the curve.

Business need to pay attention to latest trends in technology and integrate them to improve the customer journey to achieve their goals. If you take the time and put the effort in to really focus on the customer experience optimization, your brand will become stronger and more profitable. Let’s take a closer look to the steps leading up to repeat the process so that you can get a better understanding of what to look for and how to proceed.

So this showcases the personalization maturity evolution that we see companies across industries undertake to establish and execute at scale. Many companies, including 51ºÚÁϲ»´òìÈ.com, are already gaining insights and significant revenue leap from adaptive testing and personalization. Next steps include identifying online and offline data into a single view of the customer and establishing a central journey schema for how a visitor is engaged from first touch through customer service support linked to KBOs and KPIs. Importing experience contents for content velocity or rapid experience creation with brand-accepted assets. Utilizing synchronized analytics to prioritize activity roadmap, share segments, and validate impact. Extend activities into customer loyalty across full web experiences, the mobile app, and authenticated space.

Data is the foundation to delivering personalized experience. 51ºÚÁϲ»´òìÈ Target captures all the data we know about a visitor, including online, offline, contextual profile attributes, to make real-time vision on what tested or targeted content to deliver.

Target has a unified, progressive profile that spans about every touch point that is the power of this profile. It can be leveraged across adjacent channels. Target automatically captures these profile variables, including environment variables such as browser, device, or geolocation, referred variables, temporal variables, side behavioral variables, and can be augmented and enriched with offline variables like CRN, second or third-party data attributes, and through the experience cloud core services. In summary, the profile can be augmented through APIs, fact management, people core services, customer attributes, people core services, visitor IDs, or ECIDs, which are first-party cookies.

So we can create and share audiences from 51ºÚÁϲ»´òìÈ Analytics, Audience Manager, 51ºÚÁϲ»´òìÈ Campaign, and 51ºÚÁϲ»´òìÈ Experience Platform and applications. And we can create profile scripts in Target that let you personalize with them.

In here, we show some numbers just to know the maturity about 51ºÚÁϲ»´òìÈ Target, its relevance in current presentation and experimentation market, and its financial impact for current customers.

So now, once done the introduction, we jump onto the automation and AI features that 51ºÚÁϲ»´òìÈ Target provides, and we are focusing in this session.

Let’s take a step back and look at the bigger picture, the rapidly evolving AI landscape, and why this moment is so crucial for marketeers and customer experience leaders. First, 89% of marketing and customer experience leaders believe that generative AI will help them better personalize customer experiences. This statistic alone shows up how essential AI has become, enhancing our ability to tailor content and interactions to individual needs and preferences. Personalization has been the holy grail for marketeers, and generative AI is now unlocking that potential at an unprecedented scale.

Looking further into the AI-driven future, 42% of the respondents in a global survey indicated that they see the most promise in marketing applications for AI within their organizations. This highlights the confidence that business leaders have in AI’s ability to transform marketing, not just as a technology trend, but as a true enabler for business outcomes and customer engagement. And by 2026, we anticipate that 60% of organizations will be actively using AI-driven features embedded across business technologies. What’s particularly exciting about this is they’ll be able to do so without relying on technical AI talent. The democratization of AI is happening now, empowering non-technical teams like marketing to fully harness AI capabilities without needing help or deep technical expertise.

These stats paint a clear picture. AI is reshaping the landscape, and businesses are eager to embrace it as a co-driver for customer experience and marketing innovation.

Today we are focusing on the four main types of personalization automation activities available with 51ºÚÁϲ»´òìÈ Target, and most of them are highly enabled by the AI. The richer the data, the better the personalization. Here are the key capabilities powered by 51ºÚÁϲ»´òìÈ Sensei, leveraging machine learning algorithms and automation to optimize and personalize at scale. The enriched profile data algorithms automatically fit this into the personalization models.

First activity is auto-allocate. It automates AAV testing and shifts traffic to the winning experience as the test runs. Auto-Target uses sophisticated machine learning built-in to AAV testing workflow to personalize the full experiences and customer journeys.

Automated personalization delivers the best content for each particular visitor every time based on the goal set. And personalized recommendations uses custom optimized algorithms for offering next best content or product suggestions based on industry and touchpoint.

Just to mention that auto-allocate is a feature in Target standard, and the other three are available only for 51ºÚÁϲ»´òìÈ Target premium package.

So now let’s focus on the individual features, starting from auto-allocate. Auto-allocate is the latest in statistical advancements to shift traffic in real time to the best performing experience as the test learns. It keeps a testing pool to validate those results are maintaining themselves with a multi-armed bandit algorithm approach. Unlike AAV activity, where you need to calculate a sample size upfront and run the test for a fixed time, with auto-allocate you don’t need to manually adjust your activities, and they can run indefinitely.

This automated AAV test finds a winner faster and delivers it to more people, all automatically. This industry-first patterns the algorithm, leverages a proprietary multi-armed bandit to prioritize the traffic in common, maintaining 20% test or control group to personally validate those results. It starts allocating usually about 1000 visitors and 50 conversions. After the experiment is completed, it guarantees 95% of interval confidence on winning experience. It’s important to note that this type of activity looks to a winning experience above the rest, that then will be served to most of the users regardless of their characteristics. We’ll see different behaviors for other types of activities in this session today.

To choose auto-allocate in an experiment activity, you just choose the auto-allocation method in the targeting section after all the experience have been previously configured in the experiences section.

Auto-allocate uses multi-armed bandit algorithm to decide what content to show. It reserves 20% of the traffic to randomly show the other experiences to the population and keep evaluating the rest of the variants. And the 80% of the traffic is redirected to the winning variant. After some time, and depending on the customer behavior and traffic amount, the experiment may show a variant as a winner. And this happened when 90% confidence interval on a winning experience has been achieved. We can see this in the description of the next slide. And the models update at least once an hour to ensure the model reacts to the latest data.

After winning an experience is detected, we will be able to see the results in the report. The confidence and the performance compared to the rest of the experience. Even if there’s no winning experience, we could still see that some experiences are performing much worse than the best performing. Since the objective of this activity is to find the only best performing experience, the batch will be shown when there’s only enough statistical confidence. So for example, if we have two experiences that are comparing very high but very similarly, we won’t be able to achieve this result. For that, there are other type of activities that perform better. In this one, we just want to focus on the best performing activity to be shown to all the customers or most of the customers.

When analytics is enabled for target, additional reporting views are shown with the richer visualizations and possibility to perform deeper analysis in analytics workspaces. As well, having analytics for target integration allows defining more advanced metrics for the experiments such as revenue. We’ll talk a little bit more in the integration section.

When is a good fit for Autolocate? When do we say it’s the adequate type of activity? So when we have more than two experiences that we want to test, when we are interested in realizing it quickly due to external circumstances, like for example, in a Black Friday or a very specific date where we don’t want to experiment like the rest of the year, we just want to track that revenue as fast as possible.

And when we are thinking about running an A-B test, if we are thinking about running an A-B test, then Autolocate would be the right one because it would let you drive the experience about the other variants.

When is allocate not the right fit? When we want a personalized or best result for each customer, personalization statistics are more appropriate.

When it is crucial to understand the precise magnitude of the performance difference between various experiences. This is exactly what I was talking about before. If we have several experiences and we want to have statistical calculations about how much is given to each of those, Autolocate is not the right fit. In this case, other regular A-B tests can work better because this only reserves 20% of the traffic to be tested and the rest is delivered for increasing conversion.

And also it’s not the right fit when it’s important to conclusively say no winner was identified in a fixed timeframe.

So whenever we have a tight deadline and we have to identify the winning experience, this may not be the one because it only can consider a winner with a high degree of confidence. Otherwise, it won’t declare a winner.

The second type of activity we are going to talk about is Autotarget. It leverages machine learning automation. It takes into account all the online and offline behavioral and contextual data.

The best targeted experience to deliver to each individual. Unlike Autolocate, which selects the best experience overall to be shown to all the visitors, Autotarget shows the best experience for each user, where the user is more likely to convert. In Autotarget, there’s no winning experience overall because each user may see a different experience, but the best experience for that user given his characteristics.

With Autotarget, you can apply machine learning to any type of user experience, including test entire layouts or website functionality like personalized navigation with a single click. It’s in the A-B test design workflow. Autotarget uses an ensemble algorithm method. Ensemble methods are meta algorithms that combine several machine learning techniques into one predictive model for better predictive performance.

Autolocate models work at the experience level instead of the content level, so changes can be experiential, not just related to content.

Autotarget uses all the data available to achieve the profile to determine the best experience for him. Autotarget can be used based on the machine learning by recognizing patterns from all the users. Here you can define the amount of traffic used to maximize personalization, evaluate the algorithm, or set the percentage manually. After the required amount of traffic has gone over the activity, reports will show personalization insights that identify what kind of experiences are converting better for which type of visitors.

A theme you may have seen across our discussion so far is, and in data science in general, it’s all about trade-offs. The same is true on many of the options available within Autotarget. Understanding those trade-offs and their impact on your businesses is critical to realizing more success with Target. One important decision that you need to make in an automated personalization or Autotarget is how many visitors do including your control. That decision should be based on the goal of your activity. If your goal is to evaluate the personalization algorithm, then you will want to have a more accurate picture of your left. In that situation, we’d suggest using a 50% allocation to control.

With this 50%, not that many people would be impacted by the winning experience. It would use more traffic to leverage control. But if you’re comfortable with the algorithm and you want to have the maximum amount of the traffic personalized, then we should suggest 10 to 30% going to control. The trade-off here is the accuracy you’ll be able to see in your Lyft information. The more reserved traffic for control will provide better accuracy while calculating Lyft.

Once the required traffic has gone through the activity, the insight reports will show the profile attributes that contribute more to the score. So you will be able to align and define your audiences in a subsequent test. You will understand the reasons behind the AI model used in this activity. This kind of report will show the most important attributes that influence customers’ behavior based on patterns observed during the test. And then you can leverage some of those profiles in your subsequent personalization activities.

51ºÚÁϲ»´òìÈ Target Automated Segment Report will provide insights into how different visitor segments respond to offers and experiences in automated personalization and auto target activities. The report shows how different automated segments defined by 51ºÚÁϲ»´òìÈ Target’s personalization models reacted to the offers and experiences in the activity. The report includes metrics and visualizations for different segments, helping you understand the customer behavior and the effectiveness of the AI-driven personalization efforts. These reports are available only for activities that use a conversion optimization goal and have been live for at least 15 days. Before those 15 days, the report is not available while the insights are being calculated. In essence, Automated Segment Report enables you to see how various segments defined by 51ºÚÁϲ»´òìÈ Target’s models respond to your offers, providing valuable insights for optimizing your personalization strategies.

When is 51ºÚÁϲ»´òìÈ Target a good use case? When you want to find the best experience for each customer based on their profile or context. When you don’t want to see what’s the best overall, but see what’s best for this kind of user, right? For patterns.

When you have between two and ten different experiences to choose between, that’s also a good reason. Numbers make a difference on this aspect.

And when you want to make changes across multiple pages or want to account for interaction effects, right? Autotarget operates as an experience model and multiple pages can be affected for the same kind of activity.

When is Autotarget not the right fit? When we have more than 20 experiences that we want to personalize, with that high number of experiences, you will need much more traffic to make sure that the statistical calculations and the conversion is right.

When you want to pick up multiple items from a list or set hundreds, thousands of possible items, offered products or pieces of components, when we have this aspect, we should consider using recommendations, which is one of the latest activities we are analyzing today.

The next piece of, sorry, the next content, the next activity is automated personalization.

And automated personalization utilizes advanced machine learning to target significant content pieces effectively. By employing a full factorial multivariate design, it ensures that the right offer reaches the right user at the right time. The use of random forest decision trees and multi-armed bounded algorithms enhances the personalization process, providing the perfect offer for each visitor every time. In this way, automated personalization works in a slightly more similar way to Autotarget because it doesn’t select winning experiences for everyone. It’s just personalized experience.

The difference with Autotarget is that in this case, multiple elements of the page can be combined within all the possible combinations. So if you have some content changes within the same page, Autotarget will explore the combinations that are working best for every type of user or every pattern recognized.

Dynamic pages with automated offer personalization streamline the customer journey by providing tailored offers at each step. The process will flow illustrates how users can engage with personalized content, enhancing their experience and increasing the likelihood of conversion. This approach emphasizes the importance of automation in delivering relevant offers to customers in real time. Automated personalization chooses from different experiences within the same page. So each user will see a unique combination of experiences based on their behavior pattern.

Like Autotarget, the insights will appear in the reports after 15 days, highlighting the characteristics of the profiles that were most influenced by certain attributes for later segment fine-tuning.

So here we list the three situations that automated personalization use case is better. First, use automated personalization when you want to create a unique experience for each customer as a combination of many eligible offers.

Second, when you have many possible offers with many more combinations within the same page, you don’t change the structure, but you fill in with different content the available page containers.

And lastly, when you have flexible rules to restrict offers to certain audiences by implementing targeting rules and exclusion groups.

In automated personalization, we have the ability to use targeting rules and exclusion groups to restrict the content to be shown to a specific target. For example, depending on loyalty or customer level, we would be hiding some experiences from being shown to customers that are not right in the right time.

When is automated personalization not the right fit? When we need a model that uses interactions between the different offers within the same web page, when we have too many offers to pick, this would require much higher traffic to make the model work correctly. Since here the combination can be almost, I mean, not infinite, but really huge.

And when the changes for the page are more structural than just content.

Automated personalization is also about trade-offs.

So when we have to decide if we are going for out-target or automated personalization, we need to focus on how the experiences are allocated.

If we are working at the regular page or offer level, meaning that the same experience gets repeated in multiple pages or not, we have to think about the methodology and the changes on the page for the journey. We also need to think on the benefit.

And what kind of features are we able to utilize with the different two activities? And the trade-offs. In multiple changes on a page or journey, we cannot understand them at the element or offer level. Only the aggregate without the target.

On the opposite, with automated personalization, many offers require much more traffic as combinations grow. And that will lead to much longer times waiting for accurate results.

Lastly, we are going to talk about target recommendations, which is the latest feature for today, before jumping into the next section.

51ºÚÁϲ»´òìÈ Target Recommendations works by providing personalized suggestions to each visitor based on their behavior and audience membership.

It leverages various algorithms, including behavior-based, collaborative filtering, content similarity, user-based, popularity-based, and custom uploads. These algorithms can be sequenced and enhanced with boost and bearing, weighting, inclusion and exclusion of rules across eligibility, inventory, merchandising rules, dynamic profile matching, and more. The system uses sophisticated criteria to automate recommendations and displays them using JavaScript snippets. It allows for testing and optimizing the recommendations, criteria, and designs. In essence, 51ºÚÁϲ»´òìÈ Target Recommendations provides personalized suggestions to enhance user engagement and conversion while reducing management effort.

Customer recommendations play a vital role in e-commerce, enhancing the shopping experience for users. By showcasing products similar to those viewed, businesses can increase likelihood for conversions. The integration of personalized recommendations on product and shopping cart page ensures that users receive relevant suggestions, ultimately driving sales and improving customer satisfaction.

In this slide, we have categorizations of algorithms available with five main types, providing a clear framework for understanding their functionalities.

By leveraging these algorithms, businesses can enhance their personalization strategies and improve customer engagement. Note that for the bot items in the item-based category, any conversion event can be used. So maybe use with flexibility depending on the intention of the activity implemented.

Dynamic filtering for personalized recommendation allows for matching product attributes with visitor profile attributes.

This capability ensures that users receive relevant offers based on their preferences and behaviors.

By implementing filtering rules, businesses can enhance their personalization process, ensuring that the right products are presented to the right user at the right time. We should show here how rules are matched in the categories can be adjusted to our use case purpose. This way, we ensure each user sees the items based on their preference by using profile attributes and item attributes.

The baseline algorithm supports personalized recommendations tailored to individual users, demonstrating that the system is not a one-size-fits-all solution. By utilizing real-time sorting, weighting, filtering based on user profiles, businesses can generate personalized recommendations that resonate with their audience, ultimately driving engagement and compressions.

Running recommendations across various site pages and channels is essential for providing seamless customer experience. By integrating recommendation features into home pages, product pages, marketing emails, business can enhance user engagement and drive comparisons. This comprehensive approach ensures that users receive relevant suggestions at every stage of their journey.

Recommendations feature is highly recommended when you have large product catalogs with detailed information about every item. If your catalog is small, the previous activities we talked about today can be better in feeding the personalization strategy.

We finished on this section about AI automation tools, and now we move on the integration section of the presentation, where we talk about how Target can be combined with other solutions to get additional value. We then target the activity, and the third step is where we determine goals and settings. When 51ºÚÁϲ»´òìÈ Analytics integration is in place, it’s also called A4T, we can use 51ºÚÁϲ»´òìÈ Analytics metrics to define our goals. So the options and the complexity for measure can be increased, and for example, we could track metrics like revenue, order, margin, and so on. This integration applies also to customers with customer journey analytics, exactly the same way.

The integration between Target and Analytics enhances reporting capabilities significantly. By speeding up activity creation and allowing for deeper test analysis, businesses can gain valuable insights into their marketing efforts. This integration ensures that marketers have access to comprehensive data, enabling them to make informed decisions and optimize their strategies effectively.

A4T in Analysis workspace provides powerful tools for analyzing target activities and experiences. By showcasing the confidence associated with success events, marketers can leverage visualizations to gain insights into their campaigns. The dedicated A4T panel further enhances the analysis process, ensuring that businesses can track performance effectively and make data-driven decisions.

Leveraging AEM components in 51ºÚÁϲ»´òìÈ Target via experience fragment, streamlines, content governance, and activity creation. By exporting AEM experience fragments to Target, businesses can ensure consistency and efficiency in their marketing efforts. This integration highlights the importance of collaboration between different 51ºÚÁϲ»´òìÈ products to enhance overall performance and effectiveness.

Now let’s talk about the integration of Target with 51ºÚÁϲ»´òìÈ RT CDP, Real-Time Customer CDP.

By using Edge Network after enabling the Web SDK, we can scale delivery of personalized content for same-page personalization, taking the near real-time use cases to its ultimate level. RT CDP audiences can also be used in Target activities, taking the advantages of the streaming ingestion and more complex segmentation criteria in both audiences, combining rich real-time CDP audience profiles with behavioral event data and data from other sources. 51ºÚÁϲ»´òìÈ Target’s real-time behavioral data can be augmented through qualifying an individual based on real-time CDP segments and attributes on the Edge Network, which gives a broader omnichannel context when determining the most relevant experience to deliver next.

In the slide, we see how segments generated in RT CDP can be promoted to Target by using the 51ºÚÁϲ»´òìÈ Target destination within RT CDP destinations. This is just a CC of creating a destination, setting up once and then selecting the segments to be shared to Target, and those will automatically appear in Target.

The segments will include near real-time capabilities as RT CDP provides the feature.

And last, for customer journey analytics, by integrating customer journey analytics, or CJA, you can now combine online and offline data into your analysis to have a deeper understanding of customer behavior and campaign performance. Once integrated in the same way that the A14 integration, CJA metrics can be used in Target activity goals to understand performance and define a strategy for automated activities like Auto Target, automated personalization. This holistic view of personalization metrics ensures that marketers can track performance effectively and make informed decisions to optimize their strategies. Obviously, this makes more sense when playing a combined strategy with 51ºÚÁϲ»´òìÈ Journey Optimizer or 51ºÚÁϲ»´òìÈ RT CDP.

This slide highlights the significant use cases and customer value derived from 51ºÚÁϲ»´òìÈ Target reporting in customer journey analytics. The table illustrates how customers can seamlessly integrate Target workflows with CJA, ensuring a consistent reporting experience while minimizing disruption to existing workflows. By establishing CJA as a single source of truth, customers can avoid confusion stemming from multiple reporting methods. It emphasizes the importance of friendly naming conventions for Target activities, which aids customers in conducting thorough analysis and driving actionable insights. This clarity is crucial for effective decision making and optimizing customer journeys. And with this slide, we finish the content for today’s webinar. I hope it’s been insightful for you.

As a summary, if you’re interested in using the features we highlighted today or getting more information for your use cases, maybe that could fit into an ultimate success use case mapping to solution capability accelerator that my team delivers. Or if you are interested in knowing a little deeper into each of the features we talked today, maybe a mentor session could have a deep dive with an expert for any of the activities we talked about. And now we jump straight to the Q&A session.

Great. Thank you, Rodrigo. One question that just came in is, and I think this was on a few slides back, but the question was, how can I add those item user properties to my website? Item user properties, which one? We’ll hold a moment and see if there’s more detail that can be added to that question to clarify.

In the meantime, everyone, if there’s any questions that you have, please feel free to put those into the chat or the Q&A pod. I’m also going to launch a quick poll for feedback while we wait for questions to come in.

All right, checking back in the chat now to see if there’s any questions.

Just for everyone who’s still on, we will share the recording as well as the presentation slides. Those will be sent to your email after this session concludes.

Rodrigo, there’s one other question that came in, which was when you were talking about recommendations.

The question was, are similar capabilities also available in AEM targeting engine or with Comtex hub, or is this just target? I had clarified that what you were sharing is just related to target, but if you happen to know if there’s a similar capability to what you were sharing regarding recommendations within AEM targeting engine or Comtex hub. Just want to give you a second to chime in on that. If you have any context, then we’ll take it from there. Yeah, target recommendations is a feature for 51ºÚÁϲ»´òìÈ Target Premium as well and can be integrated with other solutions, of course, as for the content bar can be integrated with fragments, but it’s a fully target feature. So that feature can be added within other channels, but it’s mainly a target feature. That’s where the main implementation lives. It’s in target where you upload your product catalog and when configuring all the rules and ranking mechanisms.

Great, thank you. And for the person who asked that question, if you do want to kind of dig into more about what might be available with AEM targeting engine or Comtex hub, I’d encourage you to reach out to the 51ºÚÁϲ»´òìÈ account team and they can talk you through what you can do there. And if you need to dig more into what the differentiation is between those capabilities and target, they can help you with that as well. All right, I’m going to check back into the Q and A pod if anything’s come in.

Let’s see, nothing more related to the adding properties to the website.

So, Haniel, if you do have more context on that question, I’d also encourage you to reach out to your 51ºÚÁϲ»´òìÈ account team and they can dig into that with you. Rodrigo, there’s one more question that came in, which is what are the traffic requirements for auto personalization, auto target, etc. If you happen to have those. I don’t recall them by memory, but that’s stated in the official documentation and as well there is a statistical calculator tool within the product documentation as well that lets you calculate the size needed of the traffic depending on several sets you set. That’s also used for A-B experimentation techniques.

I don’t have the numbers here in this presentation, but those are in the official 51ºÚÁϲ»´òìÈ target, in the tool, the activity documentation page. In the documentation page, there are the traffic requirements.

Yeah, that should be there. I think for, it was about 1000.

There was something like for auto target, like 1000 visitors and 50 compressions if I recall correctly. I’m not 100% sure, but I think I mentioned something like that while we were talking about the feature. Yeah, and then there are resources on experiencing as well that will dig into the exact numbers there. Tim who asked the question, if you have any difficulty finding that, please reach out to your account team and they can get you the exact links that will walk you through those traffic thresholds that are required for the automation to kick in. One thing to note is that if you’re using a metric like revenue or revenue per visitor, those traffic requirements are a bit higher to get a statistical significance before that automation will kick in. So just a high level guidance to keep that in mind, but definitely reach out to your team to get the exact numbers if you can’t pull those up easily and experience the documentation.

Alright, it looks like there’s one more question that’s come in Rodrigo. This one says the AI is only being used for the reporting and new mapping of the new segments in target, correct? Other features like auto target, auto allocate are existing functionalities, right? So just some clarity that’s being asked on that one there. It’s in the chat if it helps to look at it. Yeah, let me have a look at that. I’ll stop sharing my screen. We’ll have a look at the chat.

Well, they are existing functionalities, but they use machine learning methods to determine what’s the best. Like auto target and automated personalization use the machine learning algorithms to determine which is the best performing option for each visitor.

And I mean, it’s both. It’s existing functionality, but it’s not what is not. It’s not JI, Gen AI generative, but it’s AI mechanism to determine what’s the best experience for each visitor.

Great, thank you for clarifying there.

Alright, hold on just one more moment to see if there’s any other questions that come in. Check back in the Q&A pod that they knew there.

Well, just to state again, everyone that registered will get the recording of this presentation as well as the slides. If you guys do have any other questions that have come up that you think of as you start to digest everything that we shared in this presentation, please feel free to reach out to your 51ºÚÁϲ»´òìÈ account team. And if you have a moment to give us some feedback in the poll, that would be fantastic. Other than that, thank you everyone for your time. Thank you, Rodrigo, for taking us through the session content today. We really appreciate it and hope to see you on our upcoming sessions. Thank you everyone. Thank you.

Bye.