Evgeniy Kulevnich: Yeah, my choice would be personalization.

Exactly as it goes. Maybe I will switch AI-powered content marketing with optimizing marketing operations. But it again depends on what we understand behind the AI-powered content marketing. But personalization is definitely the number one

Anthony Lamot: For sure. Well, meanwhile, we get the results. And what I can tell is that our audience apparently has a very similar preference. So the majority of the votes, 63 of the votes, actually went to personalization. So that’s good to know because. That’s gonna be the first subject we’re going to tackle today.

I will stop sharing my slides here just for a second, and instead, just head over to some questions started where you have gaining. I know you’re more the technical side. What are some of the things? If you look back at the past year that you implemented to further enhance and automate your campaigns

Evgeniy Kulevnich: well from from my technical side. And we we tried to be as much relevant to in our communication with our customers and as much optimistic as we could be. So I would say personalization. It was quite a big deal for 2023, and it will be definitely a goal for 2024.

So we try to be more relevant. So everyone says for a long, long time that third party causes dying, and you can’t really rely on it. So of course, first party data in 0 party data issue kind of steroids is especially this high like this 0 party steroids. So we try to use as much the euro party data from our users as we could to to make our communication more relevant, and, of course, to increase our automated sense as much as we can, because you set up at once, and you just track it. How relevant it is. And you can slightly just rules. So it it’s slightly operating with the optimizing marketing operations in this, in in some sense.

So I would say, yes, it was. And it will be 2 main focus areas for from our site, for for the past and for the future.

Anthony Lamot: Very interesting. And since we’re going to be talking more about personalization a second. But if you look back at 2023, seeing the interest of the audience in personalization, was there some something specific in your setup or something specific. You implemented that you thought was really interesting.

Evgeniy Kulevnich: We have a couple of new campaigns, email campaigns where we try to to use is set as 0 party data. And we see it works. Really good in terms of looking at KPI’s of our emails and emails that are based on 0 party data are most profitable, are most engaged across all hours, and we added a couple of more journeys from from our site where we use this data into, try to promote and remind people about their favorite items, favorite brands and purchases they might have any future. So this was recent our recent enhancement in our campaigns.

Anthony Lamot: That’s awesome. Thank you for sharing that, Helen. If we go over to you, you’re more on the operations side. What was the biggest operations change you undertook in 2023. And what was the impact of it on marketing performance?

Helen Clements: But I think I think for us, really implementing both the segmentation tool from deselect. But also engage has really been a kind of a game changer for us. And I think, yeah, I’ve been working across emails, journeys, digital marketing for 20 plus years and highly skeptical that anyone would actually be able to solve the problem around kind of consumer saturation, and how you allow journeys to be but for the sending dates at the very simplistic way to be customized and personalized to that person when. So when they took an action, it was 4 days later versus the date you wanted. So always, really skeptical to be able to yeah, allow so control while solving. You know that allowing to keep that areas of of personalization. So I think what’s been really nice for us is that we’ve been able to take Engage. Take something that yeah, that kind of AI, like a black box, AI that can seem quite scary and unknown. Put some controls around it and say, right well, here are our boundaries. This is what’s most important to us. So these are the emails that are most important, these the ones that we want, our supporters and those people that we support to get every time, and these are the ones that can be delayed, etc. But we’ve also really kind of integrated it across absolutely everything that we do, every single journey we send, every email that we send out. We use the Engage tool both to see in terms of helping, you know, with with how we actually schedule things across the organization. But also, you know, every single journey at every single point before an email goes. Yeah, has this person, you know, been saturate? Have they sent too many emails? And we’re constantly testing this. So you know, to begin with, we didn’t really know what our saturation points were. So yeah, through putting it into everything that we do. We’ve been actually being able to test that we’ve actually been able to go back to the organization and say, well, actually, we’ve got really strong evidence that highest engagement rates come from people receiving either 1 or 2 emails within a 3 day period, and that really shocked people because they were very much holding onto no more than one email a week that that was your limit. And actually, we’ve been able to test and show that that’s not the case, and we’ll keep testing or keep trying new ways of of putting it in. But knowing that that is just happening in the background has been brilliant, and I think it’s also it’s one of those things that’s as you say. AI is a buzz word, and I think sometimes it can be quite scary an organization to be pulling these things in. And this is being something that we’ve been able to bring in and integrate into all of our operations in a really non scary way we’ve been able to bring people through. And now we’ve got a little bit more freedom to do a little bit more exciting stuff as well.

Anthony Lamot: That is super interesting. And first of all, thanks for the call out, but really interesting that you actually found. So it sounds like you found out that you could be sending more than people thought. But you now have data to support that, and you can do it in a way that you’re not over saturating people within that 3 day framework.

Helen Clements: Yeah, exactly. Exactly. And we’ll and we’ll keep testing. And it might be that, you know, the more we test, we’re able to segment that data we’re able to truly understand. You know, other different saturation points for different personas, for different audience segments. Yeah. As I said, in a way, we can both understand it and then implement or find things in a way that we just couldn’t before

Anthony Lamot: Alright, thanks for sharing that. I think it’s very insightful for for those who are new in this subject, or just getting familiarized for that. Let’s go back to the slides for a second. So with that we have quickly touched upon 2023 in retrospective I think one sentence from our report. I really like this action and adaptability create opportunity, because so much has changed. And it’s still changing. I think it’s truly what makes our domain so interesting and so fascinating. And for a somewhat switch of topic that we’re going to talk a little bit about personalization.  We’ve already touched upon the subject. But it’s a key part of the study. And what’s interesting is that you know, we have historical data to compare what our respondents, who are people working in Martech, people working with Marketing Cloud and people in marketing operations, how their responses are evolving year over year. And what we see is that 52% of marketers believe personalization tactics actually impact their company’s ability to generate revenue. But that’s an increase, right? So whereas in the past it was probably sort of the top of the class that actually saw the impact of personalization. It seems that more and more people are actually seeing the impact of that. And I think, personally. And this was discussed at a user group that we held recently as well.

I think that might have to do a little bit with the fact that so many departments are under budget constraints, resource constraints. It’s a very recurring theme. And this drives marketers towards better tactics. Just like before. I would like to do a quick poll at this point and check with our audience what personalization strategies you mostly use when working, working with the Salesforce Marketing Cloud

Number one. “We personalize few campaigns”. Number 2: “using strings or dynamic content”. Number 3: “AmpScript” and number 4: “leveraging segmentation to prepare the data”, so structuring the data in such a way that you can just personalize right off the bat that could be through SQL deselect, or something else, and that

Poll should now pop up. and while we wait for people to reply to that Helen, what is your most used tactic? And what’s maybe your favorite tactic? Maybe those are different things.

Helen Clements:  I love a good content block. And I and I think being able to personalize who gets a content block allows you to, I think, elevate. Your journey is from something relatively linear to something, you know, beautifully personalized. And I think what we’ve also started to investigate and look at is, you know, this kind of kind of evergreen dynamic blocks that we can change. But we, you know, program right? Using AMPscript a a block into a journey. And we’re not having to go in and find it. We just update that block that sits somewhere.

So yeah, that’s what we’re really, you know, we’ve found works beautifully within, you know, Marketing Cloud. And then, you know, being able to leverage the segmentation that we’re doing through this the deselect segment tool. And then using those those dynamic blocks around script I think, has really kind of taken our, you know, our emails and our journey is from something quite basic of, you know. “Oh, look! We’ve got a first name in the subject line that we feel good to something that really shows that we we really know our audience”.

Anthony Lamot: That’s great. And I think that’s maybe something that our audience can learn a little bit from as well, when I share results, I think it’s interesting how just spread out of this. Not too many people, apparently, are using the dynamic content block option. Some people aren’t even personalizing their campaigns, and the others are kind of spread out between segmenting to drive personalization and AMPscript itself.

So. Interesting result! Thanks. Everyone for answering those questions. Let’s go back to to the panelists with some additional questions on on the subject of personalization. Starting with you, Evgeniyi. How do you influence or convince the 2024 consumer to purchase? How do you know you have the right offer at the right time for them?

Evgeniy Kulevnich: It’s a good question, a good question. But again, personalization at scale, it’s it’s quite 20 sentence. But personalization scale is I would believe that would be the 2025 topic for a quite hot topic for everyone, especially in retail, because we have tons of items to to promote to you, and the people consume in general quite a lot. It depends on area. It depends on the industry, but it’s a quite intensive in the industry, I would say so. We will try to personalize our communication, but we’ll try to be as much personal, as relevant to the customers as we could be based on data that we have based on finding some trends trends across our data that we which we we use on our daily basis. So try to be as much irrelevant is, it could be personalization. It’s what what we believe in in 2024, together with automations. Because automation since only saves your time for all the campaigns and for other stuff, but it’s also produces cost for all your ongoing campaigns.

So it’s also another another topic to to suggest, AI is another hot topic for recent years, and for an upcoming years. AI is a good to have a look, but for me personal feeling is, people more talking about AI rather than using it at the moment it, it again depends on the industry. It depends on where it could be used or not. But I think, AI, it’s really good tool to find hidden patterns that you can find with the people in some specific. AI generated content. I would say it’s a bit more questionable for now. So we are. We are looking at it. We are looking at this direction, but we if we start using it, we’ll start it like slightly, because auto-generated copy could be super nice. But we operate in 15 different markets and English. For example, if we’re talking about auto-translation translations with AI, English is our less populated market  across out because we’re number one in the Nordic department store. So Nordic languages are not at the top of least. For example, for tools we’re using that if if they want to include it as a language for translations for AI and a loyalty, loyalty is another big topic for us for 2020 phone, because loyalty drives people to go to your website. Loyalty brings your revenue. If if you look at this, at the, at the business, but with working on a loyalty. You can make your customers more happier and more engaged with you.”

Anthony Lamot: For sure. Thanks for sharing that so comprehensively. I think we’re going to dive even more deeply into AI, and the caveats surround it in a second. But I want to take a little bit longer with your 24 strategy. You’ve already highlighted things like automation personalization. What about the content? And how does that tie in for you?

Evgeniy Kulevnich: Well, in this case, I think if you want to work more in a personalization strategy and content on here, you need to own a lot of data because personalization personalization will be based on data. So we’re already on a quite a journey of controlling and gathering data on our own.

Anthony Lamot: Great thanks for sharing. If I would synthesize that, it’s that data drives your personalization which in turn will drive your content. And I think there we would all agree that marketing cloud is, and uniquely data-driven platform uniquely suited for that kind of activity held on somewhat similar question for you, how do you factor in personalization when strategizing 2024 content and campaigns?

Helen Clements: Yeah, I mean, I would say that it’s, you know, one of the very first things that we think about. Both in terms of our marketing cloud. Journey, but actually across all channels and many of our campaigns are multi-channel.

Anthony Lamot: we’re constantly thinking about it, and how about them reflects into our various websites. Social media we’re still doing, you know, print for and telemarketing.

Helen Clements: we are constantly kind of building this spaghetti architecture of what all these different campaigns are, how all these different journeys interact, all these is, is this a journey that is a one time journey for someone, enter it, go through various things and then leave. Or is it a longer term journey? So we are constantly kind of thinking kind of with 4D chess in terms of how that personalization fits together. But I think for us as a not for profit. It. It’s particularly important. Because yeah, we need to be talking to the people that are supporting us and who we’re supporting. But how we talk to them. What we tell them about is just so important that we get right.

Anthony Lamot: That’s a really interesting consideration that you have to take into account. One thing that, by the way, I love the 4D chess analogy really appreciate that one which leads me to ask many people using marketing cloud or other marketing automation platforms. They will. They sometimes wonder how to maintain the overview in the face of so many automations, and then some ad hoc campaigns going on. What would you answer to them?

Helen Clements: There’s various different ways. I mean, for us engage has been brilliant at being able to do that in a really visual and kind of practical, useful way. You know, just through schedule and calendar, and being able to go and look at it and see how it works. But I think, you know, fundamentally for us, it is constantly looking at the data constantly looking at what the evidence is showing us. Constantly looking at, you know. How can we evolve things, you know, in really small pieces? So that we don’t get kind of overall and daunted by what it is that we can do. But no, it yeah, it is. It is a challenge. But, as I said, having engaged is that kind of small thing in the background? That means that we can. You know, we don’t have to sit up at night worrying about it cause that’s taking care of it for us.

Anthony Lamot: Great! Thank you for sharing that. I’m going to jump back into the slides. At this point. Evgeniy already touched upon the subject of AI very briefly. And that’s exactly the next thing we’ll discuss. Obviously, since the launch of ChatGPT 3.5, but especially ChatGPT 4.0 last year. There’s been a lot of buzz about this, and a number of new providers on the market, and so on.

And so we asked, our respondents, have you personally used AI tools or technologies in your marketing efforts? And really, interestingly, already the majority of people have used it, although, you know, this question itself doesn’t answer how thoroughly, which is something we explore more in the study itself.

But basically, our takeaway was that as knowledge of AI becomes more integral to marketing efforts, across industries, geographies, and so on. Marketers will need to learn this technology to not only create impressive results stay competitive.

And I think that kind of ties into with what we’ve seen. Management consulting firms, venture capitalist firms publish about this.

So let’s going to skip this survey, for now let’s go back to our panelists here. Question for both of you. How do you see AI enhancing the impact of campaigns, or maybe the ease of creating them.

Helen Clements: I think as it’s already been touched on, you know, Chat GPT. As a tool to go and go. “Oh, wonder what it does when you do this”, or what you type it. You know it’s fascinating. You can get lost in it for for days, frankly, you know, but I think from in terms of kind of AI-driven content created content.

Yeah, I mean, there’s always been templated emails that you’re able to look at as a as a marketer. That have existed on the Internet. So it’s it’s it’s interesting. But I don’t think here marketers need to fear for their jobs quite yet. And I think what would be, you know, amazing is if you could have your own kind of AI content, generating model. That was very personalized. Your own kind of voice, your own brand. You could put a whole load of content into it.

But I think until we get to that point, III think for us it’s not something that we’re kind of actively using other than it can be a great way of generating ideas when we’re looking at brainstorming, a new journey and AI in the broader sense, though, as I said, we’re very much using it for our saturation control. And you know, we’re we’re using it for making sure that you know people are getting things at the right time and the way in the way that works for them. Yeah. But then, yeah, I’m sure that ChatGPT, just from my perspective, just suddenly popped up. I’m sure there’ll be new tools that will just suddenly pop up.

And you know, maybe in a year’s time it will be possible to create your own model. That would be brilliant. If you need a new thing to work on, that would be my request.

Evgeniy Kulevnich: Yes, I think I completely agree that on, on, on individual level, people use it a lot. And if we can train this model. Now I want it would be super beneficial for for us, and if I’m not mistaken, a couple of vendors on the market already trying to suggest your tools that way you can train your own model based on your own data. If I’m not mistaken. So we are quite close to it, I hope. And on the other hand, how can how can we benefit from AI in this case? It’s of course, it’s a cost operation. And it might replace some manual work when you create your campaigns. If you’re talking about marketing area, a reducing cost of copy to create a copy, to to generate a new content and etc.

But, as I mentioned, it’s maybe it’s not as trustable yet, is, it could be used in on a commercial on commercial side, because it’s interesting to look at. But when it’s when, when the question is, if I go into man implemented? Maybe. No, not yet. We’re we’re we’re closely looking at it. But what we see is the results are not really ready to be implemented in a, you know, commercial communication, promotional communication site and etc.

But from other perspective AI could really enhance customer experience. For example, I know that there are a new AI tools on the market where you can use computer vision. So you don’t need to have an attribution model for your product catalog.

And just AI can look at your pictures of any pictures whatever you have on your in your on your business. And based on computer vision. It get give you results. So computer vision based on picture and search in general that I think it’s quite interesting area to to have a look at.

And to find in different patterns in the data that you have in the user behavior. This, I think AI is really good already, really good at that. You can use and try. And your own business, of course, through AB testing, test out everything, test always.

But this, how you can measure and see their results. and yes, speed of maybe speed of different operations, as also could be quite beneficial. But at the same time I read and read really interesting research, saying that when people rely on ChatGPT or AI tools a lot, they can reduce the creativity of people. So because AI models give you the wider range of results that are more average in any topic that works. For most of the majority of your customers or majority of your people who try to use their models. But if you are on a different size, different range of of the this average model, you can struggle a bit, because creativity reduces.

Really custom, related solutions or custom, any custom solutions that will be more valuable, and only people can provide it. So if you want to go faster. And maybe in a wide range of solutions, go with AI. If you need unique custom solution, you still need people to an expert in the area to to be able to provide like the best for that. What you can get from it.

Anthony Lamot: I love how you summarize it at the end. And, by the way, on the visual recognition part I really liked. It’s bit of a personal note. But I really liked the the Demo by OpenAI recently, because now in the mobile application, they also have visual recognition, and they have this demo where they show someone struggling with a bike, taking a picture of the bike lock. And then ChetGPT figured out the solution based on online resources. So I think we’re only really scratching the surface of what’s possible.

But I would agree that we have to be really careful when it comes to content creation. I’ve tested this for myself. I post a lot on Linkedin, for instance, and there was a short period when I had CPT write my posts, and what I noticed was when I didn’t edit them a lot. The engagement immediately plummeted. I think there’s just something  unnatural generic about it. If you don’t train it if you don’t tweak it. However, meanwhile I’ve I’ve worked a little bit with that, and what I do find is that, provided I have a long session going where I’ve for instance, shown the AI examples of my writing. How? What? What? My style is, what people like. It starts to adopt, that it’s still not there, and maybe it will never be 100% what I want. But it does do like 80% of the heavy lifting. And that’s been a time saver for me.

On top of that looking at our own marketing operations, we’re also starting to experiment with it. But that includes sometimes for internal things. So, for instance, just the other day I had to provide. I was preparing a memo for a marketing department to look at something in the CRM system, and I actually wrote out my ideas to ChatGPT, and then it just wrote it out in a whole memo. So I think sometimes even for internal processes, or making sure you cover all metrics. So hopefully, those suggestions help our audience. I wanna switch back to Helen for a second. You’re a nonprofit. Are there some unique ethical considerations that you keep in mind when dealing with AI?

Helen Clements: very, very much so, and I and I think kind of touched on it. Earlier. In terms of you know, we are dealing with people in what can be frankly, the very worst times in their life. When they’re reaching out to us for support. And sometimes yeah, the the data the evidence can throw up some areas and things where, if you were to do it, it would be really, really damaging. So really, good example of this is, you know, one of us, like many other kind of charities that deal with you know, serious health conditions and concerns. Is that very often people can do fundraising and donate after they’ve lost a loved one.

Now, if you were to do a model to say, when is the best time to contact someone, though you’re most likely to get a donation from them. The data might tell them well, straight, after somebody has reported that they’ve lost a loved one.

And if you were to just let AI run with that that could lead to some really, really damaging things going out. So we have to be very careful with that. And this really interesting example from a big mental health charity in the US where they had, you know, a similar thing to ChatGPT, that they put on their website to run the to run that kind of their chat bot on there. And it went from seeming helpful to giving out devastating advice to people in under a month. And they had to quickly remove it. 

Yeah, we’ve just here in the UK with a big career company, you know, had a had a thing in the press where the AI on this website. Somebody got it to write a poem about how awful this Courier company was, and it sounds quite amusing and and quite funny that yeah, oh, you got the yeah, the AI on a website to swear and to do these funny things. But actually, when you’re dealing with really, really serious areas. Yeah, you just have to be really sensitive to you. Just can’t you know you, you can’t take a chance with it. And again, I kind of apologize almost for that kind of trivializing. You know, people who aren’t in, not for profit. But again, I kind of use a kind of sock thing if you tell somebody about the wrong size sock, it’s not devastating. If you tell someone give someone the wrong advice at the wrong time within the not-for-profit it it can be in very, very real ways. So yes, we have to be incredibly careful, incredibly cautious. Constantly testing, you know, constantly speaking to our community, we have an amazing group of people, who from within our community, who are directly affected, who we can go to and say, How does this sound does this work? We can test things out as well as testing on our kind of broader AI integrations. And the way that we’re using it.

So sometimes it’s not that we don’t, you know, as marketers, we’re not incredibly excited about it. We want to run really quickly and try new things. But there’s real real-world consequences if we if we do it wrong.

Evgeniy Kulevnich: I would want to quickly remark here that I completely agree that AI in communication should be used really, really carefully because it’s because of the communication, as you just mentioned, it could be not on the road, not to the right person, not at the wrong time in the wrong communication in general. So this is maybe not fully trustable, but I know that some big players on the market, are trying to put like a trust layer in between AI model in your systems. And it is just so you can really trust the AI answers in generated content, etc. But at the same time, AI, I think, could be used in social communication, but to listen to our customers. I think it’s really good with the huge amount of data, big mess of data.

And if you can teach or create a model that can listen and combine it with your CRM plan, communication plan, and etcetera. Then you can find really good moments and results.

Helen Clements: You know the drives. I’m not drives your communication in, etc. So now I’d agree. You know we use it in terms of our market research that we’re doing. We’ll use it across kind of text response on that. And that’s lovely. And within social media. So yeah, lots of other, you know, exciting things. And I think it’s really good for you to bring them up. I think sometimes those things get forgotten about with the kind of really all your chat Gbt and chatbox, and all the kind of things that buzz, wordy kind of area of AI.

Evgeniy Kulevnich: Yeah. Yeah. So, and you also mentioned that. Brand, you know tone of voice from your brand and branded content from AI, it’s also like, if AI can provide you just an average picture, an average copy, or that works for most of the people, that’s it’s not like what your brand communicates. But I know this company is already working in this direction, so they make they can make a model that will be trained on your brand book, on your photos, on your copies, etc., and then it can generate the voice of your company, and if you’re talking about generated images, it can generate images according to your brand book. So it’s where, at the beginning of this, but it’s, I think, companies already working towards this direction.

Anthony Lamot: I also love the idea for market research. We do this too. Obviously, we’re B2B, and it’s really important for us to know our customers are seeing in calls about the market, about the product, and so on. And we start to use call listening and then AI on top of that to sort of surface the newest insights for those topics that were really interesting to us. And then the marketing team can take that and feedback to sales for learning and education, but also the product team to figure out, okay, what do we need to build next?

We’re going to switch to the last part of this webinar which is about optimizing marketing campaigns in 2024, and like before, I’m going to start off with a little poll here for the audience this time about ROI.

Which is the most important factor when calculating the ROI of new tech? Is it cost reduction? Is it revenue increase, time value, adoption? If there’s any other that we haven’t listed there, we’re always welcoming you to post it in the chat. I’m just going to grab that poll real quick.

And it should be showing on the screen right now. And as we gathered results like before, I will ask maybe starting with you, Evgeniy. How do you measure the ROI? Maybe let’s stick with AI for a second, because we were still talking about it before. How do you measure the ROI of AI.

Evgeniy Kulevnich: Oh, it’s a good question. I can’t really say because I’m not sure that we actually adopted AI. Oh, on a scale for the whole business because the business is quite big, and we have millions of customers so good to know. But I think it’s from our technical perspective and in within marketing area, I think good. If we see on our main KPIs how engaged your customers are when you start using AI. If they are more engaged with your communication, if you have more clicks, if you have more items in a basket. If you’re talking about retail and retail as a business.

If you see average order of value for every basket that you have, every purchase that you have. I think that people would add more items to the basket to buy, and you see that everything leads to more revenue that you can get from one single email, from one single communication, or from one single campaign. So increasing in everything cost reduction as well. I like all the other answers. I think all of them are correct, and all of them are really beautiful to me. And if you’re talking about cost reductions, yes, of course, if, I think you already mentioned it, average campaign takes 12-14 weeks based on your reports the average time for one campaign with the across market years. So of course, it’s also a quite important moment. But it’s really depends on what type of campaign, what type of business, industry, and etc.,

But what what I would really want to highlight. It’s time to time to value in time to market. Now we live in a quite a complex world, and we have a lot of ideas, and everyone is super busy with the road maps and projects and etc. So time to market, I would say. It’s one of the main questions that you that you need to ask when you implemented or thinking about something because you can always do it on yourself much better than everyone. But if you can provide it and deliver it within the 3-5 years, maybe the world will be changed so much that it won’t be relevant anymore. It will be wasting of resources and time.

Anthony Lamot: Absolutely. I can really endure the time to evaluate as well as CEO when we purchase new software for our company. It’s one of the key questions. Anything that doesn’t, you know, if it doesn’t help the business within 3 to 6 months. It’s probably not something that we’re going to be looking at. And how long have you been using AI?

Evgeniy Kulevnich: In marketing? I don’t know. It’s a good question. It’s a long journey, so it started. If you’re talking about simple things, yes, as you mentioned, personalization. It’s also like an area, and the area of the AI, and we started implementing it in our company since, I don’t know, 2012 maybe 2013, something like this.

Anthony Lamot: And you’ve seen a good improvement over the years. And you’re glad that you started using it when you did, or do you wish you would have started earlier?

Evgeniy Kulevnich: I think that we started just in time for the business for the business at the moment when we implemented it. It was a really good moment because we already had quite a good quality data, a lot of things. That we had time to do before we started with this AI, and it was like for example, for average e-commerce store. If you’re talking about our size. It would be a waste of money because you don’t have a constant. You don’t have like good constant data and etc. So I think that we started really in a good moment.

Anthony Lamot: Alright, good to hear. I see we’re getting some results on the poll. We got about half the audience participating. And we’ll see if that changes here in a little bit, but most of them are leaning towards revenue increase.

And some are saying cost reduction. And those are the top two, of course. Which makes sense. That’s what we’re usually looking for in any kind of investment is, you know, how much is it going to cost me, and how much am I going to make from it?

Anything you’d like to add to that, Helen?

Helen Clements: I mean, I think probably our view will be a little bit similar. It’s kind of a mix of all of them, because I think time to value is, you know, really key, but that’s because the quicker you get a return on your investment. And, you know, we’ve probably got less in terms of a large amount of data, but you know, still quite a bit. And I think, you know, to show that we can do this quickly and kind of test and learn is really key.

And I think so obviously, revenue increase is always one that you know. That’s why we’re all here to make more money, I suppose. But then that obviously naturally leads to cost reduction, because if you’re working smarter, and if you’re not having to produce as much and you’re not having to do things as quickly, and you can be a bit more considered.

So I think, for me, it’s kind of like a journey, but it’s kind of, it’s it is about balance because you need to have a kind of a steady flow and I don’t think you can really say one is more important than the other. But yeah, for me, I’m probably saying revenue increase. I think that’s the ultimate goal for us, obviously.

Anthony Lamot: Great, thank you. We did get one question here in the chat.  It says, if, and I’m not sure if this is relevant to our conversation, but I’m gonna ask it anyway. If possible, could you please share examples of how you measure the ROI of AI in your work?

Or if that’s something you can’t share, maybe some suggestions on how others can measure the ROI of AI in their work.

Evgeniy Kulevnich: We are we. The main KPI for us is how our customers are engaged and how many items they can add to the basket. How much money they spend on the platform on the on our side. And I think these KPIs are really basic and would be relevant to most of the businesses. If we are talking about a small retail store.

If we are talking about the huge e-commerce marketplace, like we have at the moment. And it’s more for CRM for communication and for stuff like that. And I think in every every step, every small thing that you do you need to think about time, time to value, and how quickly can you deliver results. And you need to plan it carefully. I think that in today’s world, it’s one of the most important things, how you can plan and carefully and quickly deliver results.

Anthony Lamot: Yeah, absolutely. I’m gonna I’m gonna check on the poll real quick. Let’s see. It looks like revenue increase is in the lead at 56%. Followed by cost reduction at 22%. And then we got a few for adoption and time value. So it seems like most of the audience is leaning towards the revenue increase. So that’s good to know.

Anything else that you guys would like to add on this topic? Before we move on to our last part.

Helen Clements: Yeah, I think, for me, it’s kind of just how do you prove the ROI for it? And, you know, we’re trying to do that by all the things that we’re talking about is, you know, speed to market, we’ve kind of pushed, you know, with some of our tech suppliers where we’ve had kind of these pilot phases for a very long time. And you kind of go to these conferences, and you hear all these amazing things, and you’re like, wow, we want to do that, and then you try and do it, and it takes like a year and a half.

And actually, that’s kind of gone now because we’ve got to do all these other things. So we’ve really kind of just pushed to kind of do things quickly, which has been really successful. So, yeah, for me, it’s kind of, you know, be really clear on what you want to achieve and kind of challenge suppliers, I suppose.

And I think sometimes you have to be a little bit brave and kind of push people a little bit out of their comfort zone. But actually, you know, the results are really great. And actually, it’s just being really clear on what you want to do.

Anthony Lamot: Right, yeah. No, that’s great advice. Ok, let’s move on to our last part here.

So what’s the future of AI in marketing? Where do you see this going in the next, you know, three to five years? Is there anything that you’re excited about? Maybe any challenges that you foresee? Who wants to start?

Evgeniy Kulevnich: I think we are still at the very beginning of AI implementation in marketing. I think we have a lot of. Still, we have a lot of possibilities.

And it’s going to be something that we will see in the future in a couple of years. For example, for us in our industry, the fashion industry, I think one of the important things is going to be how we can use it for production.

For example, now we are still seeing clothes that are just being sold out after we produce it. We still produce a lot of things that are being thrown out or are not sold out. And I think that the next big thing for our business is going to be how we can use AI to predict demand. And how we can produce things based on the information that we have.

For example, one of the. Examples that I can give you is that we have such information because we have a lot of bloggers and influencers. And we have influencers and bloggers in all the countries. And when they post something in one country, we see that it’s being sold out in the next couple of hours or days, sometimes.

But we still don’t have this perfect mechanism of sharing this information between our departments, and we are still working on it.

Anthony Lamot: Yeah, yeah. Predictive analytics for the fashion industry, I think that’s gonna be really big. Helen, do you have any thoughts on the future of AI in marketing?

Helen Clements: Yeah, I mean, I suppose it’s slightly different. But it’s still data and kind of, I suppose. Kind of customer centric. And I think for us, it’s kind of the Holy Grail is kind of being able to do predictive modeling on the whole business.

So, you know, you can take your supply chain, and you can take your kind of you know, your sales and all those things. And we kind of, we’re really trying to kind of work towards that, but it’s really difficult because there’s so many different factors that kind of, you know, make your business go round.

But I think that’s probably the kind of the Holy Grail, I think, for kind of retailers and kind of the product that we sell, because obviously, our business is seasonal. So, you know, the better that you can predict kind of what is going to happen, the less stock you have to kind of, you know, carry over, which is obviously costs. And then you can kind of be much more kind of considered and how you go to market.

Anthony Lamot: That’s a great point. Yeah, using AI for predictive modeling across the entire business. Alright. Well, I think we’re we’re coming towards the end here. And I really appreciate both of you joining me for this conversation. And I hope the audience found it valuable. I know I did. And I look forward to maybe doing more of these in the future.

So thanks everybody for joining, And have a great day!