ihi G Glossary it's going to be the same speakers, the same type of content. And please make sure to even stay to the very last minute because there are some gifts for all of you at the very end. So without any further ado, we have a packed agenda. We're going to go straight into the presentation. So I'm going to start doing some introductory words about AI and marketing, bridging those two together. I'm going to hand over my word over to Kasper, who's going to talk more about his perspective at Acutics. We're going to bring in perspective Christina, who's going to talk more about her experience working. Sorry, working with AI, personalization and marketing from a consultancy point of view. And last but not least, we have a case study provided by Maths at Newellfisk. And at the very end, we're going to have some closing thoughts and Q&A for those of you in the audience who would like to ask any questions. We are going to be open to address that at the very end. Perfect. Now, just some introductory words between about who we are at IAH, but also at Acutics, because we do come in with a strong partnership. And for those of you who haven't really heard about IAH before, or Acutics, or both of us, then you have a chance to try and understand a little bit what we're doing and moving on. So IAH Nordic is a data and AI consultancy established for the last 20 years. And we have a very strong anchor in data and AI strategy. We're also doing a lot of cloud architecture, web analytics, and also are renowned for our stellar digital marketing team as well. And we are the ones who can provide with the strategies moving forward, whereas we have with Acutics, who are more of the machine, basically the engine that powers our strategic work as well. And Acutics is essentially a two-product company. It has a standardized, so it standardizes your data language. It also has a validated product as well, which also focuses on making sure that all your marketing sources are healthy. And it all powers this accelerated insights, getting speed to insights quickly, and making sure that the output of the data is actually something we can act on. And what we have seen generally across various scenarios from our experience is that building a solid data foundation is going to give you heaps of benefits out there. And here are just a few examples. So if you want to dive into more advanced analytics, basically doing a smarter personalization, making sure that the reporting is faster and more accurate, that in turn will allow you to do a lot more of your, basically orchestrate your customer journeys in a better way. And at the very end, if you would like to go even more advanced, start doing some predictive modeling, predictive insights. And as we enter post -summer season, there are many companies out there who seize this urgency. We've taken a bit of a check on what the market and industry says about this. And there are some interesting facts that we are seeing when it comes to AI adoption out there. 63% of organizations, they do not have, or they're even unsure if they have the right data foundation for doing AI. 87% are going to fail due to poor data quality. And the ones who do work with strong data governance, they are more than three times as likely to succeed in this field. So with that backdrop, I'm going to hand over to a person who knows the ins and outs of a shared data language. And this is the CEO of Acute, Hasbro, who's going to bring us on board. Thanks, John. I'll do my best to talk about a shared data language as we have done it for quite a lot of companies already. And it's something that we are not only praising, but talking a lot about in terms of the current state of AI as well. It's like it has been, have a greater focus on the foundation part. Because if you see here, I mean, in many situations, I mean, in the world, if we don't have a common language, we're speaking English now because we have peers that are listening and that don't understand Danish necessarily, could speak German, French, et cetera. And that's the human language part here. We need to have this, this common language to understand each other. If we talk about AI models, we could take MCPs, I mean, model contact protocols, the common communication language between AI agents, et cetera. So it's about forming this language that makes our business departments, channels, activities, anything that we do, speak the same so we can actually start comparing apples to apples and pairs to pairs and so on. It's like the enablement layer, so to say. And we see it, I mean, a very examine, I mean, to the point example here because, I mean, data languages comes in many ways. In this context here, we are speaking marketing languages, for sure. And one case, we often see, that's the naming conventions. Naming conventions within media platforms could be journeys in the CIM activities or creative naming of our assets, et cetera. And the problem that we're always stumbling upon, that's inconsistencies. I mean, we're creating naming conventions, I mean, on our own. If you don't have a structured way to do it, we'll name it differently. I mean, you might even have, as a good example for segments here, marketing might call it SMBs, sales calls it mid-market, operations calls it something completely else, could be like tier two accounts, and then we have someone else that calls it something with revenue, could be like a financial operations project. And if we go one slide back, then on the solution part here, that's where we want to structure, right? We want to have these consistent patterns and structures where data is named consistently, so we don't need to clean it afterwards. I mean, we can fit it into our models and we understand that, well, regions are named this way here, channels are defined by this, and platforms, this, objectives, can be awareness, leads, conversions, etc., but no more than that because that's the structure we're creating. And it's not about control, it's more about enablement, I mean, enabling the power of our marketeers to trust and actually use the data so we can start driving value instead of arguing about what numbers to use. We had actually a session a few weeks ago where we talked about the optimization truth and the business truth because, well, I mean, I'm not going to start an attribution talk here because it's going to take, that's a whole other session we're going to do, but thinking about attribution, everything falls back to that one and a half in the market saying accurate data so we can attribute it correctly and see what has an impact and what can we optimize against and so on, but we often hear that companies are struggling to find, I mean, what attribution models should we use? Well, I would always say it depends because what are we optimizing against? Is it the business outcome? Is it budget allocation? Is it meeting our key numbers and our revenue targets, et cetera? Or is it the optimization truth that we're looking into the media platforms? Because you can, even though the meta Facebook data is siloed in that platform, I mean, there are so much optimization possibilities we can do with the signals in that platform, but that's like one way of the truth. It's not wrong, it's just one way of the truth and then we have the business truth, but this shared data language connects the dots also. And then another question we always get is what is it? Casper, you always talk about taxonomies. What is this? And I would say our definition, again, it's subjective here, but our definition is like it's shared data language is a common established taxonomy of standardized structures and governance rules and something that spans across the organization. So not something that alive in Germany and we don't use it in Denmark and we use half of it in France for the US part. We need to align that we're all in the same picture and we're doing it for the same part. And the easy way to see it and have a little cartoon drawing here that's in the kitchen. I always tend to use this for people that are very new to taxonomy building and data languages here because if you think about us cooking something here and not with my kids because they like to put everything in one pot and then it's a mess. But then the ingredients I mean, that's going to be the inputs and we're going to have the recipe what we're following here, that's the standards, right? That's how do we structure and how do we use the ingredients in a right way to actually make the dish that we want to do. And then we have some kitchen rules, the governance aspect, I mean, washing our hands, we're doing, I mean, all the rules that I mean, working in the kitchen, that's the rules that we have applied to make sure that the dish is created in the right way in a sort of way. And then, of course, the chefs, I mean, that's how it seems, right? And all our marketing peers, everything that are working in this, that's the one who cooks the dish. And data standards, I mean, it can be so much. We recently went to a data governance conference. I think Cindy and Kasper also went there and they can say that data governance is broad. It's a lot. And in this case here, we want to focus on marketing because data governance could just as easily be the supply chain management, manufacturing, and HR. I mean, so many other things here. But governance, essentially, it's something where we embed the rules into the data. And in this case here, and for the nine examples I have, that's for marketing specific. And it ranges from naming conventions and media platforms. So we have marketers that are optimizing in media or Google Ads or a display platform or Salesforce for CMS or a Bloomreach or whatever. They often rely on conventions for naming parts to make sure that they make the right choices and do reporting and optimization to see what actually performs in here. But it could also be UL parameters. We all know in terms, Google Analytics or Adobe CID tracking codes or PIVIC or Metomo or Piano. I mean, it doesn't really matter of the system, right? Also, I can't mention any more systems or probably, Snow cloud, maybe. But it's a way of standardizing that the tracking signals that we get linking from the media platforms or the activations to the website where we then see the conversions. And then making sure that we're tracking correctly so we don't suddenly have a lot of direct traffic that actually came from a social platform or something that came from Facebook but we were missing the paid part of the tracking so everything was attributed to organic. And these things here are small things that actually have a big effect in reporting and budget allocations and maximization patterns etc. and then if we move on here to the next one we always talk about why is it critical here. I think nothing actually many of our conversations around data standards and governance people are saying well why is it important how do we measure it and to be honest that's always where we have the hardest part because what's the quality of good data but if we turn it around what's the consequence of bad data wrong decisions and everything and for ages there have been the saying and I have said a lot as well garbage in garbage out we do have a new saying golden gold out right I mean let's make it positive here I mean the better we feed it the better the outcome and we tend to say this with AI as well that's our approach to it and we see it from the market as well there's a greater focus on establishing the foundation because if we spin it back to the slide you had and the numbers you mentioned John here there are a lot of AI projects that are failing I think the MIT study a few weeks ago that came out was at 95% not all bound to data I'm aware of that but bound also to the outcome wasn't as expected which may have been because of the data foundation like an indirect implication and it's like the market is realizing that cool I mean AI we can do all this magic stuff here not all magic but I mean automated stuff and fast moving optimizations and automation and these things here but if we don't have the right foundation I mean we'll start seeing hallucinations in our reporting patterns not just in our AI prompts as we see already but in our reporting where we're actually basing our business future so that's why we want to have that so if we go back to the slides here so we have in and out of course what we put in can also be what we put out right because we don't want to have this amplification of issues if we go back one more in here the second part here that's trust I mean if we don't trust the inputs how would we ever trust whatever outputs that are coming here because if it's made by humans or made by AIs I mean again if everything if I don't trust the data that we're capturing let's just say Google Analytics then I may embed some you know Google has some rules that is for capturing missing data and all these things here but if I know that most of it that comes in are wrong then when Google tries to actually create the lost data will that be correct probably not so it's kind of like establishing that truth in the data here and trust and then there's the scale part because AI allows us to do much I mean we see it ourselves when we do marketing activities I mean if we can do it a little bit it works I'm always like let's do more how can we do it faster but if it doesn't work and we're not aware of it doesn't work when we then put power to it and scale it with AI or whatever we do it will just amplify the mistrust and the incorrect decision levels here so it's so important that this data foundation there is correct before we start doing and what can we do from here we have three steps I mean I don't think it's an option anymore to look into this it's kind of becoming like a necessity and of course there's the definition part but actually after we did the physical event here I want to turn it around because establishing the ownership of data I would like to put that at the step zero that's the initial party who's going to have the data ownership we're working with a few different clients who are actually talking about this who will be the owners and control of our taxonomies and data structure because only when we have that established we know who should be included in step one that's the taxonomy alignment and it's more I mean and that's also a whole talk that we can do another time because there's a lot to it of course the skills with the company complexity is stakeholder management and aligning teams and just asking the questions what are we using our data for dashboards cool then probably we shouldn't use it are we using optimizations oh we can now take action on the data then there's definitely some value and so on but I think that the last point here is the fact that AI is moving fast and the ones at least the ones we're seeing here and also I think I have some shared views here the ones that wins are not necessarily the ones that embeds AI first and fastest but it's the one that ensures the foundation part they're doing it right that's that will probably be my last saying here to get something that will be get the foundation over and then start working from there otherwise you're not it can turn into a fast fail it's probably a hard word to say but difficulty 100% it's super insightful thank you so much Kerstin for sharing your thoughts and speaking of someone who has seen what works what doesn't work we have a person in the room who has been building data strategies for many large brands out there in the world and who can speak a lot into many experiences where companies are aiming to nail the foundation in order to enable personalization and AI and so forth and that is Kristina who is the director of IIH Nordic so I'm gonna hand it right over to you Kristina thank you and I think I will actually just continue where you were stuck and then take it a bit further into the personalization layer so the activation part and why it's so important in that sense to also have the data foundation in place and I actually think that it's McKinsey has said it very easily they said that if your data isn't ready for AI then your business isn't ready for AI so that is kind of a lot of studies summed up and I think that is why we also see a lot of the failures right now and then again there's so many buzzwords out there so what does it actually mean when you say what does it mean to be AI ready and one thing is the data language having the same taxonomy as you just mentioned is something about the trust so actually a lot of the same as you just said seeing the data real time and ensure that we actually can trust it that we can have this connected view so not just for reporting but actually also so we can follow our customers and we can act on the right time when it comes to the personalization and then I think where a lot of companies are struggling that is the scale part it's so easy to do the pilots it's so easy to have some of the organization to do stuff I know you also know that conversation that you have your clients but how do we actually ensure that we can scale it so we need to have this in mind and yeah figure out how to do this and there there's a lot of cost when we look into the data and you already mentioned some of it so I won't go into details but one thing we haven't mentioned today yet at least that is the time spent when data is not the right then we simply just use so many hours in looking into numbers and what are the right numbers can you find this number for me yeah you'll have it in two days so there's really also something about efficiency here that's less technology but more the part of what we spend our time on and what can do easily and just some examples from that we pulled that out so maybe some of you recognize these Monday morning why is not LinkedIn conversion fitting our Google analytics that was because the UTMs as you said it's not fitting Tuesday of the budget increase were not as we thought when we forecasted because there were some issues with the data in itself and then the content was not as we thought why don't we see the full journey there's some breaks and that can also be a thing and the Friday plan as I said before hey can you give the numbers for last week and yes you can have it on Wednesday and today it's simply just not fast enough if you would like to be competitive and all of these when we have these conversations in the organization then it's typically because the data foundation is simply not there so what we often do and what we recommend is that we kind of go in and then assess where are before it was data now it's AI readiness so new buzzwords same story I would say so where are we are you on the first level doing the manual fragment operations looking into spreadsheets figuring out how was things doing can we do a people table or something moving into the more automated where that's also often where we see acutics coming into play where you can actually support having this common view and make it easy to do the taxonomy and really make it easier to have a common data language moving into the unified personalization so that is where we really start stitching the customer journey we have different channels that we can cross over and we can start looking across how is the behavior and then there's the AI readiness where not that many to be honest on today right but that is really where data can work for itself and we can have it to optimize them and I think for that part technology is one thing but there's also the whole organization being ready so yeah really share the part of ownership and stakeholder if you don't have that in place it doesn't matter then it's just a cost in the technology and so on so we kind of look it into these four layers so it's really about in the bottom where we have a lot of what you just talked about so ensure that we have the connectivity between the data that is where MCP can help us that we have the UGMs the naming that we collect the right tracking so what is our KBIs what is that we would like to see from a business point of view ensure that that is actually available in the data and then of course the governance layer that we still especially with AI right it's always been important but now it's become even more important that we know what we are collecting and why and then move up the ladder which is more about the unified profiles to 60 view we can do the campaign orchestration the quality insurance automated that's also validated and we can do that with some acutics magic and then we are moving more into pattern recognition using the data for the activation and for the insights and then the real fun cool stuff but again it's really about starting the button it is boring but we need to do it but it's also about ensure looking at the business and then ensure we build it from that so don't build everything it's not just nice to have it's really need to have when it comes to the business yep technical so here is a bit more again so it becomes actionable hopefully for all of you sitting up there listening it's really bad if you are on the first layer ensure that you know as I said before what is the KBIs what is the commercial output the business objective would like to achieve by going into this journey of personalization audit what have today where do have the gaps and so on and establish the ownership of the governance and so on as said before and then can move in to the exam parameters standardize them and so on it can go pretty fast those steps can go actually it doesn't have to take long and do it by connecting not everything but when makes no sense and then scale it into maybe a CDP there's also tools that can use for that today with AI and so on but really start looking at how can we look at the customer journey and see that from start to end and then go nuts in data and do the predicated personalization right and there's also tools for that today so it doesn't have I think that's what has actually changed five years ago it was hard work to get to that point today with the AI it is easier not easy but it easier from a technology perspective so when have this architecture in place what we see is it's easier to do smarter personalization you can scale insight much faster and can also scale it with the AI and ensure that everything goes into the channels without having the manual work what we also sometimes see is that everyone talks about hybrid personalization or personalization in itself also here need to take it into steps I know a lot about that so ensure that we start just by being able to see the inside start doing segmentation get some learning and then can move more into micro segmentation identity and hyper personalization but again always have in mind what is the maturity in our organization and our data foundation instead of just going crazy with something that the management would like to have in both words yeah so it is now that is the good thing we still see that there's opportunity for getting up to speed maybe not be the first mover but then at least a fast follower so as I said before go and assess it look at the commercial part analyze and do a plan we do see that people that skip the first steps here that they are very more likely to fail so that is one of the main reasons when these projects are not successful that is because chose to go in just to the quick ones and build something fast it simply can't scale so just to sum up from as said we have done this for many years so what is the common pitfalls that need to avoid don't start with AI before have fixed the data ensure that know what are tracking ensure that have some C level sponsorship it can start somewhere below but at some point when would like to do it for real need to have some management to back it up and be the voice of this and then don't understand maybe change management because as I said before and as you also mentioned technology is one thing easy easy but working with people and having them to use it that's a whole other story that we can also do a session about so really it's about starting small learning fast and then scaled out strategically and smart that's my view on this that's right and that's a very fresh and insightful view and I really hope that a lot of you who are watching learns as much as possible from the frameworks we're using because that's a lot of what we approach our clients with to make sure that they get an understanding of what the effort and time is actually needed to drive success from here so a huge thanks for Christina we are going to head into the last bit of this webinar the time has been flying by I wonder if there are any questions coming from the audience if there's anything top of mind it can be anything regarding how to get know internal buy-in to create some kind of urgency for naming the foundation in order to get to the AI stage where want to be or if want to ask anything into the product itself of Acutics or anything of that sort because we have a lot of sharp brains here today a lot of years of experience coming in so this is your time to really get the answer to the question you've always wanted to ask say that a question we often get when we're sitting out there that is the buying part how because organization wise ownership how do we make sure that our colleagues stakeholders working in 40 markets how do we make sure that everyone buys into the idea of let's get some good data to the business we have often seen that it feels miserable if we just enforce a structure and a new process for people I guess it might work for someone that a very special type of company let's call it that but for most of what does work is enabling the marketer what value do they get out of it so each single person that are part of establishing this data foundation they need to know the value that they're getting from it it's about creating maybe creating reports or data flows where they see themselves where can access the data what are using the data for what does it mean when I follow this form in the codex or follow the data flow set by IH to actually create this personalization part here what does it mean in the end here what values I'm getting and this needs to be communicated to each single level of the hierarchy and in organization that's by far where we see the greatest outcome more then it's also about choosing some of the more mature markets of course also important strategically markets but often they go hand in hand so why do we have some fund runners that can buy into this and use it and it's also ready to say okay it looks a bit weird I will ask a question around why and then we can have that conversation and learn and then take the other markets so I think the part of thinking you can do everything at the same time exactly and I see that's coming something here but if do it like this as would like it to do you're getting all this as well then we from a company invest in setting up this reporting framework and everything that you can actually start utilizing because you're following the patterns that flows into this framework here yeah yeah yeah we have a first question from the audience how long will it take to implement the first steps for good data what is your experience experience in this I can say it depends on how I want to go right I mean we're working with businesses that have let's say for a utent tracking perspective I mean that goes in some metadata there the media source maybe and maybe a campaign launch year maybe descriptive text maybe are targeted objective and so on we have some companies that are working with five elements and some that are working with 50 obviously the 50 takes longer because then need to align but actually the five can take long as well if align in 40 markets through those five elements that you're working on so I think it's about starting small process I mean regardless of how much work we spend place it's not going to be followed so I always say focus small on the data part but really go deep into the process establish the data process who owns it I mean what does it take to request a new field how are we using the data I mean all that because once that's there well then we also have to flow for taxonomies then we know who's in control we can update that and we have the process already so it's just a matter of well if they're using acutics have a new field or still storing spreadsheets we'll add a new column or whatever that's right yeah when we do project together it's often around the quarter right 90 days from when we start until we actually see the first outcome but then there's the whole change management and that I think meal fizk is a example of that takes years before you have a lot running to face where it's ultimate and everyone is using it but I would say first step to get started if can do that quickly and get the learning and as much as this is a technological consideration there's also a process as said Caspar and behind every process there are people so I'm wondering Kristina what would say are the most important skills to have as a team as an individual to move forward with what we were discussing I can think it's about being adaptive so it's less about competences that is also good but that can learn also with good people around but really having the mindset that we don't need to have everything in place and if this correct and what to do over here but be adaptive and ready also in general when work with AI right that data I don't know if you haven't done it no no but I will continue cool and speaking about starting small starting somewhere become curious about what the current state of affairs is we are reaching the last bit the last cookie from today and that is coming back to the gift that I wanted to present to today so have the opportunity to request an AI ready data report that is going to allow to get a status check a health indicator of where your campaigns or how healthy your campaigns are and here can choose anything from your marketing stack whether that is LinkedIn meta Google ads or anything else and this is an extensive report and it addresses what goes well what doesn't really go that well and luckily this process is very simple can start by reaching out to me or reaching out to any of our dear friends at Acutics to basically set up a call to understand what channels want to investigate and that's essentially it after a couple of days you will get this report and that will provide you with a lot of insights to where your first steps really can take you so yeah I hope the audience enjoyed this as much as we liked hosting it so if there are any questions or comments we are very happy to answer them feel free to latch on to any of us after this because we're very passionate about this subject as you probably could hear so if there's not anything else I would like to thank the speakers here Casper Castilla and Mass tuning in from remotely and to the audience who also listened throughout this session perfect thank you very much and hope to see you around soon