How To Work Around Walled Gardens And Reach More Customers — Reeto Mookherjee | Why Traditional Attribution No Longer Works, How Brands Use First-

How To Work Around Walled Gardens And Reach More Customers — Reeto Mookherjee | Why Traditional Attribution No Longer Works, How Brands Use First- Speaker 2: Hello everybody and welcome to the Ecommerce Coffee Break podcast. I'm Claus Lauter and you're listening to the podcast that helps you become a smarter online seller. In today's episode, we find out how to supercharge your gross marketing with predictive AI. Joining me on the show is Reeto Mookherjee. He's the CEO and co-founder at Angla AI. So, let's dive right into it. Hello and welcome to another episode of the Ecommerce Coffee Break podcast. Today we want to share and find out how to reach more customers. We want to talk about walled garden and what that is and what all of this has to do with predictive AI. With me on the show today is Reeto Mookherjee. He is the CEO and founder of Hangular AI, a company pioneering predictive conversion software for digital marketing. With over 18 years of experience in AI and machine learning, he has held a senior leadership role at GoodRx, NBCUniversal, and Ingram Micro, and he is an Informs Edelman Laureate for his groundbreaking work in pricing optimization. So, let's welcome him to the show. Hi, Reeto. How are you today? Speaker 1: Good. Really nice to do this together. Speaker 2: Cool. Let's get started. So, performance marketing used to be very, very simple years ago. It has changed a lot and it's not that easy to get through or see results on MetaGoogle or TikTok. So, tell me what has changed. Speaker 1: Yes, sure. Actually, fundamentally, I would say the shift has started in around 2017 when there is a slew of regulations that started coming in. What started with GDPR in Europe continued at CCPA, all of these regulation changes. Then there was a major shift happened when one of the major tech players, Apple, made their privacy policy changes. They're saying that, well, platforms, it's okay, you can track users when they are using your product. It's not okay, or you need to ask for explicit consent, but track the same users when they are not using your product outside and they're going onto the internet. And that essentially disrupted the way the entire architecture, entire digital marketing ecosystem was in. Essentially, they needed to have data and data about unrestricted view of what people are doing inside their app and what they're doing when they're living it. And that line of sight almost got broken for all the iOS users. And when you think of Apple, it's more than 50% market share in some of the markets like the U.S. That's where most of the buying power comes from, those devices. So that disrupted the way digital marketing was working, and that happened around 2020. Speaker 2: Now, I think a lot of merchants are very aware of that. And then some workarounds came up, conversion API from Facebook and so on and so forth. But obviously, the access to the data, to the details that are really necessary to target your audience was not as good anymore. What kind of challenges came out of that for brands and how did they try to work with that? Speaker 1: Yeah, so, you know, as you can imagine, that was a major disrupt. And when Apple started rolling out and everyone, you know, there's a question around, well, how many people will give them consent? And it turned out it's a very low number. You know, the latest number of stats I've seen is 88% of users have opted out. You know, it's almost become de facto. People thought, well, if I ask for permission different ways, maybe people will give consent. It didn't work out. So what happened overnight is, you know, if you're thinking of CMO all the way to a hands-on keyboard media buyer, your media efficiency suddenly on the surface got lost, like your, you know, media cost or acquisition cost or efficiency. If you're running direct response playbook, your cost of action, desired action, which is usually a revenue event, cost per goes up, went up by 100, 200%. It's like everything feels like broken. And then there are, you know, then there are a lot of solutions started happening. It's like, no, no, no, it's an attribution problem. Let's solve the attribution problem. Your ad is working as fine. You just cannot see the results. Well, then that was the, you know, then a lot of investor VC money and a lot of company attribution-focused companies. Okay, let's see. Maybe your ad is working equally well. You just cannot see it now. But in all this experiment, that's actually not the case. Yeah, attribution, you can build a better mousetrap outside and you can get some of the efficiency, but it's still not adding up. If your loss is still like 50% plus efficiency loss, you're reporting. They're like, okay, what is the problem? Platforms have been very tight-lipped about this, right? They couldn't tell how severe this is. Like almost like if you look at any world garden, like of course I'll say meta, if you look at there, there are almost like two universes exist within engineering teams. There is a 700 plus engineers working on attribution solutions. There's a very few set of engineers working on auction side of things. Auction basically means is And it's a very efficient auction platform. So, who do I need to show the next ad to from this advertiser? Then, so that, you know, it win-win-win for everyone. You know, user has a better experience. Advertisers get the desired outcome and the platform also makes more money. So for that very complex decision, that auction system suddenly was starving for data. They were flush with data. Now they are getting, only they could use 30% or less than 30% of data they could use for auction training. And then the very poorly understood concept is not that they didn't have the data. Brands are still sending data. They are anticipating these changes. They are collecting more data, like asking survey, quizzes, before you become a customer, all of those data collection tools, CDP, customer data platform, all those investments went in. So, not that they're getting more data, they're sending all the data back to ad platform, but the auction engineers, their hands are tied by, not by that data, by the legal team. Legal team says, no, no, no, you cannot touch that data. If user has opted out from my digital tracking and iOS, you cannot touch that data. So, it became like, well, I have the data, I'm dropping the 70% of the signal on the floor, and my line of sight suddenly is broken. I can see some users all the way through the cradle to grave, but majority of the users have no idea when they're leaving my platform what they're doing afterwards. And it creates a major disruption. Speaker 2: I think this was the best explanation of the situation that I have ever heard. So, congrats to that. And I think our listeners should listen to that twice because that really explains the situation what has happened in the last year absolutely perfectly to the point. Now, also, everyone opted out and then people were surprised that they get completely random ads in their news feeds and get annoyed by that as well. Now, you came up with a solution to help with that, and we want to talk about predictive AI solutions on that side. Talk me through. How does it help? How does it work? Speaker 1: Right. So at the core, when you think of it, is that before these changes happened, iOS 14.5 and all the changes happened, the platform was doing certain work on advertisers' data. So you send the data, they were doing some, think of it as data pipes going in, but they will enrich that pipe before they use it for auction. And they could do it because they have unrestricted permission to it. But you as a brand, when you think of it, users opted out from platform tracking. If you're a brand, this is your visitor, your data. Of course, you have a consent form and everything. Most of the users are okay to share that data with their brand that they're buying from. So, if you own the brand, you still have all the data at your disposal. So what we do is we sit in between that data flow. So think of the data supply chain or data flow as a simple pipe, like where you transmit all the data as is. We intercept that flow. We do some extra work. In the middle, then we send the data, you know, then before we transmit the data. So in a core, what it means is, and the communication mechanism is essentially is a conversion API or API system that each of the platform has that's basically saying, this is how, this is the standard protocol for you to send me your transaction data or your browser data or your offline transaction data. That's open for everyone. You don't need any permission. So what we do is we get all the data, The enhancement that happens is we, at the core, is a predictive AI. So the predictive AI that does, you know, we use a deep neural network. We train, we, in the early days, we had about, you know, Today, I'm going to talk to you about predictive AI. So, we've come two years and a few months into the journey. The first eight months, we've spent a lot of time and energy. Being a former data scientist myself, we make sure that we build the system, this prediction, as robust as possible, as general purpose as possible, so that we can cater to a lot of different customers' prediction problems. So, the prediction that we do is based on what outcome you as a marketer want to do. Visitors who are likely to have high propensity for conversion, and we send that data back to our platform. So, then telling Meta, okay, your line of sight is broken. I get that. But can you only focus on this good stuff? Don't focus on the less of the fluff stuff. So, we basically bring in efficiency back to that mechanism. If someone says, well, I'm going to acquire customers, my wills, the customers will repurchase in the next 90 days. Or if you have a subscription brand, the customers are never churned, okay? These are the good stuff. Auction, find more of these people. So that translation service that we are doing on behalf of brands, and if the brand has their internal data science, machine learning, and all the technology skillsets, they could do it themselves. What we find is most of the brands, even if the brand has a billion revenue plus, The skill set is not as simple as, well, I hire a data scientist and I do it. It's a different skill set. It's a very super specialized skill set, a lot of know-how on ad tech, marketing, plus data science, machine learning, plus maintaining the system, plus changing, constantly evolving as platforms change their requirements. So it becomes a lot of undertaking, and we are platformizing all of that. Do it once, and you can distribute the cost amongst more users on the platform. Speaker 2: I think that's a very smart solution there. And as you said, you need a lot of skills, knowledge, and expertise to build something like that. Now, you said you're pulling the data from different platforms. Now, we have listeners who are on Shopify, who are on other platforms, who are using other apps. All of this data can be used. Is that right? And how can you use it? Speaker 1: That's right. So when you think of it, we have customers. We're really lucky, actually, in the early days, For our journey, we actually got some tough problems. We have some customers who are using completely homegrown solution. They don't use any of the Salesforce, BigCommerce, Shopify, any of the WooCommerce, none of it. They're like homegrown. Or some of them using Shopify, but I only use Shopify backend, Shopify headless. So the technology standpoint, when you think of it, we see it as a data in, data out. Data in standpoint, You know, we make it easiest way for our customers or prospects to get us data. So, if they're using Shopify, of course, it's very straightforward. Under 10 minutes, literally takes four clicks and creating an account and that app installs and we start, we're all set to get data from. Same thing happens for if there is a commerce platform like Salesforce Commerce Cloud, BigCommerce, WooCommerce, all these we have apps or plugins. Now, let's say there are some customers maybe not just having homegrown, but they have invested in customer data platform. Great. Customer data platform like Segment, Telium Segment or Amperity or other platforms, we actually have built-in connections connected for it. So, we have a destination. Angular is a destination from the CDPs. If none of this works, like if you're something some other ways or you do not have CDPs, then you still have a mechanism we can connect directly to their cloud data warehouse. If they're using Snowflake, Redshift or other places, or they can deploy Angular tag directly on their JavaScript or tag managers. So we make it data ingestion as easy as possible and as customizable as possible based on the need from the client. And then on the destination side, of course, we work with, you know, all those world gardens we mentioned, and we are building more destination for programmatic web and outside of world garden as well. But, yeah, that's how we tackle the data in and out problem. Speaker 2: Let's talk about the distribution side, about the ad platforms themselves. How do you measure the attribution, the results? What kind of reporting do you supply? Speaker 1: Right. So, the biggest thing is when we started on this journey, my co-founder and I had two things, which is, can we build a technology where it is very easy to experiment? While experiment, when you do the experimentation, because the performance marketing is the end, you need to run a lot of experiments, a lot of tests, and you need to have at scale, running those, getting inferences, what's working today may not work, you know, two weeks later, and then you constantly evaluate and testing. And then taking that same theme, when someone, not a customer, just want to pilot it, we didn't want to, you know, we'll be as less intrusive as possible within their marketing operations function. So, the way to imagine if customer working with us or prospects working with us, install their, our app, then we do not want to mess up any of that existing setup. They're already having a way to send data, they're optimizing, it's all fine. The way we say it is, set up a new pixel, new datasets, you know, so give us a new container, completely fresh. We don't have all the history. Your legacy setup has years of data, all the data we have. We barely started getting data from you last minute, you know, and now let's start pumping the data with this new technology into this dataset. Then, after a week, we are ready for a test. So, when you think it, like, one hand, we are already starting at a disadvantageous position because your legacy pixel has years of data. We barely have a week, in some cases even less than a week. And we're now saying, well, okay, you run an A-B test. A-B test means, you know, in platform A-B test. So, Facebook has an A-B split test. So, they are making everything being equal, same creative, same copy, same audience even, you know, but only thing is those campaigns are optimizing with this new way versus your legacy way. Everything being equal, only thing is that difference. And then sure enough, if you run it, we have run over 200 A-B tests, 96% of the cases we won. We won at the statistically significant best level. That means if you run this test again, very likely, you know, 70, 80, 90% chance that you're going to find the same winner. And then when you think of it now, it's like, well, you're already, you're winning already on a, you know, disadvantageous position because you have, your hands are tied. Now what happens, natural question is, okay, what happens if you start sending all your new technology data into my legacy pixel, then I can use all my history plus going forward better data. Then you see another state function improvement. But by that time, you know, the technology is somewhat proven. Because they already tested in their data and then clients start using it. Then they see the state function improvement in their overall KPI. Marketing efficiency ratio or blended CAC, all of these things started improving. It's because you get out of this attribution noise. Because attribution, last-touch attribution in this day and age doesn't work as well. So you might win. Something may look very good on the surface, on the click, but you may lose the performance marketing. And I'm going to talk about the battle overall. So in this way, when you start, you experiment, you prove it out. When we start being the system of record, then you see a state function improvement and that performance stabilizes. And it also gives an optimization capability or experimentation capability of test out new things. You have a hypothesis, very likely platform can support with some labor event orchestration. So we orchestrate new event, you want to start optimizing on it. Great. If you work it, then you found an unlock. If it doesn't work, there's iteration on those things. We pretty much create a smart data loop that complements all of the creative, all of the strategy, all of the things that are very much still fundamental of performance marketing. We see three pillars of performance marketing is strategy, Creative and copy and data loop. We take care of the last pillar, very essential pillar. We make the data loop, smart data loop. As long as you have the first two pillars, You know, you should be getting – it's a significant unlock. On average, we have seen 32% improvement. Speaker 2: So your $100 spend certainly works like $132. I think the performance marketers listening to this episode will have a big smile on their face right now. I want to dive a little bit on privacy issues and maybe data limits. Is there anything from the platforms where they cut you off or where you see a risk? Speaker 1: Actually, it's the opposite of it. So, when you think of it, the platforms like Meta and others, we are actually working pretty close with them as they are rolling out new offering and product because they just realized that someone needs to do this translation job and they cannot do it. And after Cambridge Analytica and all of the data privacy changes, they are not We need to take that chance. You need someone who is a marketing service provider to the first party, to the brands, that data. We take privacy and security very seriously. In my past lives, I implemented HIPAA systems for health tech companies. I worked in other regulated industries and entertainment where there's a VPP, a Video Privacy Protection Act. We build Angular as privacy by default. The way you think of this is it's not your data contributed to other data. It's strictly non-contributory. We are a marketing service provider. If a brand stops working with us within 30 days, all their data is gone. All the trace is gone. And then within the data itself, the way we handle it, there is very sensitive data we handle with, like customer information, customer data, and this. So the way is, again, we take very good best practices from this thing, which is decouple sensitive data as early as possible. Keep it in a vault. And only use it when you need it for transmission side of things. When you look at prediction, I cannot tell any, you know, if you go and look into Angular system, it's completely ID-based. I cannot tell you whose data it is. Because I don't need to. Machine learning systems or deep neural network doesn't need to know it's Claus or it's Reeto. It needs to know this is the user ID and this is the content ID, this is the product. So that's how we abstracted the concept. And then, of course, when we communicate back to that platform, then we need all the data again. We go back to the vault. We get the only permission-based access and very few human beings actually can access it. I mean, some of the data vault, I even cannot access it. And we did it by default because That's how important it is, the security and privacy at the forefront. And so we are building it for the future. We expect more regulation to come. Only requirement is as long as brands have permission use of their data, their consented data, that all goes into the platform. Speaker 2: Yeah, and I think that's the way it should be. Once you have the content from your client, then it's in the best interest, in the best service level for the client, and then you can use it. Can you share some success stories or case studies of brands that you have worked with and what kind of results they saw? Speaker 1: We have many use cases we unpack. So when you think of it, try before you buy a brand. It's a one-table, one of our early design customers. They are like stitch-fixed competitors when you think of it. So they were optimizing. They were the early customer. Again, they are homegrown, so they help us build all of the technology flexibility. So, they have been seeing, you know, over 30% improvement in, you know, their new customer acquisition cost. And then what they started seeing is about an 11% improvement in their retention of the customer. So, it's not only you're acquiring customer cheaply, but you're also, you know, those customers are ordering more retention instead of behavior. So, almost like that becomes a free cash flow on top of your acquisition cost efficiency gain. Then we have a customer, you know, called Borotation. So, they are Shopify. Actually, they recently became, you know, the top 10 innovative brand. They are, you know, think of their Nespresso for cocktails. So, they make cocktail machines, Nespresso for cocktails. Very innovative brand, you know, grew really fast, you know, in the last few years. And they had an existing setup, what they started using. And we are seeing, you know, Close to 90% improvement in marketing ROAS on the North Star they care about, which is the new customer acquisition selling them first product. We have a women apparel brand where they're seeing, you know, they're interesting is they're a holding company, PE-backed holding company, and before they started with Angler, they had this constant battle between the CMO and the board about new customer acquisition because the numbers they saw in platform didn't match with, you know, what they were seeing in their overall system, financial reports. And even though they are falling to the T's, all the best practices out there, their data sanity, data practices was the best in class I've seen. But it was still not working. And it became like really a perfect use study is why this is not working or what is broken and how we come into fixing that. So that's Christy Dawn. They saw 33% over improvement, over I think 35% improvement in new customer cap. We have some brands, and Facebook actually came up with a case study. There's a digital landscaping service, and they recently had an exit. But when we started working with them, there's a digital landscaping service, $1,500 class digital landscaping service. So people don't buy on impulse. People come back over and over again. So the way they are using paid marketing is they buy leads, and then those leads convert. You know, then they have different mechanisms, CRM, tele-sales, you know, all of the sales force to convert. With Angular, they got same cost per lead, but those leads converting 55% better. And it's not us telling us, it's the platform, Meta, published that case study with the agency and the brand and Angular in the mix. So things like that. We have some outdoor furniture, long consideration cycle, and a very innovative outdoor furniture brand. They were using, you know, when they started using Angular, they saw immediate impact on, you know, bottom of the funnel. Then they started expanding to the middle funnel, and it become a full funnel optimization strategy. Beauty brand's Wonderskin has great use cases of that, identifying the movable middle for their unlocking growth. Like, you know, think of it, beauty is such an impulse driven buy. And, but this is also the category where there's a lot of competitions out there. You know, there are so many brands out there. Even Sephora has a lot of, a lot of people are conquesting Sephora audience for their purposes. And in that category, the growth actually comes from not the in-market. Interestingly, it will be like what we call the movable middle, which is you talk to the right people. They may not be ready for an eyeliner yet or whatever product or skincare and a cream that you're selling. They may not even know you exist as a brand. But once you engage with them, then they are likely to convert in the next seven days or whatever time they will take, whatever the usual window is. And when you tap into that audience, interestingly, your competition is much less at that point. Your media cost decreases, but that actually, or media, you know, impression cost decreases, but those users also convert, not in the same session, within the next seven days. They will come back as a different session. They may come back from another channel, but your first engagement you're driving from the channel, that will drive to the ultimate conversion. And we're seeing a stiff function improvement in their overall funnel efficiency. On the one end, there's a very long consideration cycle, niche market to mass market impulse buy. This technology works because all we're doing is segmentation, and segmentation with the data that brands have and the business model they have. Speaker 2: Sounds a little bit like the good old times when that was working from the platform itself, but with your solution, I think we're coming back to that. Now, walk me through the typical onboarding process. What steps are involved? Is there any kind of homework to do before you can get started? How does it work? Speaker 1: Right. So onboarding process for most of the platforms. So if someone is a Shopify merchant or they're on Shopify or BigCommerce or any other commerce platform, the onboarding process literally takes under 10 minutes. So I'll use the Shopify as an example. There is an Angular app. You can find it. You install the app, then in the process, you connect to the destination. So, you'll see all the channel destinations like Meta, Google Ads, TikTok, Snap, Interest on the platform. Let's say you say, well, I'm going to start with Meta because that's where most of the spend is going in. So, you connect there and that's it. That starts the process from the back end. So, we start listening to events. We start exporting the data. We, of course, in your workspace, then you start getting about three or four emails from us, like as each step is complete, well, we are data ingestion complete, just ignore. You don't need to do anything. Just when progressing is that, you know, the pizza opened, right? The pizza is made, and we just move to the next stage, last stage. What happens after the last stage is done, we are already starting events into your new setup, so not disrupting anything. And then at that point, our customer success person or team will reach out for a follow-up. So in that follow-up, what we discuss is essentially, this is what we learned. It's not a black – we don't want to make it just magic. It's like, this is what we learned. This is what's going to happen. You are now ready for a test. So, we shared a test plan. Literally, like any hands-on keyboard marketers, it's like we never spend more than two minutes, maybe five minutes on a test plan because they're like, oh, this is straightforward. I know it. You just clone your existing and you optimize in a new way. Just drop down, select it. You run your own experimentation now. Usually, two weeks later, we come back. In many cases, actually, before two weeks, the platform will say, well, I actually find an early winner, and 96% of the time, it's us. But again, it's like, well, after two weeks, we reconvene. What did we learn? Here is the thing. By that time, at that point, we already have a business case for it, how we are going to be helping, and what does it mean for this top-line and bottom-line impact. At that point, and then we talk about the services, our service fee becomes one of the easiest sell from the internal selling standpoint because this is the technology that will give you at least 10x ROI from the services and everything you factor in. Speaker 2: Okay, that sounds very, very straightforward. Let's talk about your pricing. How do you charge for your service? Speaker 1: Our pricing is based on the size of the brands and how much media they're going to optimize through us, through the channel. Typically, there's a minimum fee, about $1,000 minimum platform fee. Then usually what we try to do is we try to get about 3% of media that optimize by us as our fee. That's our target. Of course, we have some really big spenders where we'll be making a lot more money if they're paying 3%, but there's a ceiling. So if there's a floor, there's a ceiling, and we try to strive for about 3% of media spend optimized. And 3% may look like it becomes a very trivial decision from this standpoint, because if we are delivering 32% improvement, getting you 3%, you can take your 3% budget cut, so you can keep your flat budget, and you're going to get 29% of the benefit. It becomes a very easy sell from that standpoint. Speaker 2: Okay, that's very easy to calculate for a marketer and to put into their budget. Reeto, before our coffee break comes to an end today, is there anything you want to share with our listeners that we haven't covered yet? Speaker 1: Oh, so I would say the landscape is changing very fast, you know, and it's a very exciting time to be in industry, you know, and with all the JNI, LLM, the innovation happening on the creative and copy. And what we are focusing on at Angular is the predictive AI. So when you think of the JNI innovation and predictive AI on that thing, when you join these two things, almost like a left brain, right brain coordination, and that's where the magic happens. So that's why it's a very exciting phase. We are creating a smart data loop. Performance marketing is just a wage for us, we think. We focus on performance marketing because we are passionate about it. We had a founder market because we've been Working with marketers, we saw firsthand how problem it is and we think it's really a painkiller, not a vitamin. You need that solution. But then possibilities, once you create the smart data loop, you can use the smart data loop for site personalization. You can do better retention. There's endless possibilities on this and even collaborative commerce. So that's why we're excited about what we have built and what we are learning with our customers and design partners together. And it also opening up what is possible once we. What you know the extensibility of the platform so and we made it very easy to test out. You know, I know there's a lot of claims out there, you know on the what is technology possible and we are making also very bold claims and I think we believe is the best way to prove it out prove it out for every account in a very in a simple way so that it does not require a lot of investment on anyone's time to see the results and that's why we built so we're willing to Today, we're going to tackle a more hard problem. It's the use cases that we can solve. The use cases we solve, we think it helps everyone. It helps us. It helps other brands. We're trying to solve the same problem. So, we're definitely looking for people who are questioning the status quo. If you're not happy with your performance marketing and if you think your data loop Okay or not okay, even if you think it's top-notch, still I think there's an improvement opportunity, so give us a shot. Give us a chance and we'll have a lot of fun along the way. Speaker 2: Yeah, I would totally agree to you. I'm in digital marketing for 25 years. I'm a dinosaur and exciting times are actually on yet. Since AI came around the corner, things became much more exciting, much more productive than they were before for a couple of years. So, I would totally agree and I think there's so many opportunities right there and also with your solution that marketing will become better, will create more results, better results than it was in the last couple of years. And I hope a lot of people reach out to you. Where can people go to find out more about you? Speaker 1: Sure. Our website is getangler.ai. Getangler.ai. And of course, you can find me. We have LinkedIn presence. And we also, you can start a trial without talking to any one of us. You can start the process, complete self-service, or you can also book a demo. And we will be happy to, one of us will be happy to talk to you and walk you through the steps. Speaker 2: Excellent. I will put the links in the show notes, as always, and you are just one click away, and I hope a lot of people reach out to you. Reeto, thanks so much for giving us an overview, and I hope to talk to you soon and see what's in the box for you and our business. Thanks so much. Speaker 1: Thank you. Speaker 2: Hey, Claus here. Thank you for joining me on another episode of the Ecommerce Coffee Break podcast. Before you go, I'd like to ask two things. First, please help me with the algorithm so I can bring more impactful guests on the show. It will make it also easier for others to discover the podcast. Simply like, comment and subscribe in the app you're using to listen to the podcast and even better if you could leave a rating. And finally, sign up for our free newsletter and become a smarter Shopify merchant in just 7 minutes per week. We create content from more than 50 sources, saving you hours of research and helping you stay on top of your ecommerce game with the latest news, insights and trends. Every Thursday in your inbox, 100% free. Join now at newsletter.ecommercecoffeebreak.com. That is newsletter.ecommercecoffeebreak.com. Thanks again and I'll catch you in the next episode. Have a good one.