In this podcast episode, we interview Carolin Bink, VP of Customer Success at 1plusX, an AI-driven data management platform.

Carolin shares what makes 1plusX different to other data management platforms, how their AI-first approach helps both publishers and marketers to utilize data, and what the impact is for marketers with third party cookies going away.

Listen to the podcast now via the links below:

Transcript: Interview with Carolin Bink – 1plusX

Speakers: Mike Maynard, Carolin Bink

Mike: Thanks for listening to marketing b2b Tech, the podcast from Napier, where you can find out what really works in b2b marketing today. Welcome to the latest episode of marketing b2b technology that podcast from Napier. Today I’m joined by Carolin Bink, who’s the VP of Customer Success at one plus x. Welcome to the podcast, Carolin.

Carolin: Hi, nice to meet you. And happy to be here.

Mike: It’s great to have you on now. I mean, I’ve been really keen about this interview, because I looked at your LinkedIn profile. And it seems to me that you were a customer of one plus x and then joined the company. I mean, was it really a case of joining a company that you were using and love so much, you felt you had to work for that?

Carolin: Yeah, basically, it was pretty much like that, to be honest. So I was searching in 2016, DMP back when for the publisher in the sales house. But I worked at Axel Springer holes owns a lot of classified. So I also got into contact with a lot of markets classified data. And yeah, basically, I found this Swiss exotic looking startup that didn’t do any marketing and just had engineers as employees and started working with them. And then yeah, really enjoyed it. And yeah, then I got this opportunity to, to do what I what I love the most full time consulting publishers and marketers worldwide about data strategies, and that’s why I moved to work with the technic technology, I introduced this.

Mike: Well, and a big change in terms of company size from Axel Springer, which is huge. To a start up.

Carolin: Yeah, yeah, indeed. But I had the lack of Axel Springer to be kind of in start up environment like data strategy was always Yeah, it was an innovation hub. So yes, it’s a change. But it was mainly changed in terms of people engineers knowledge. And also, for me a really high learning curve to learn more about AI. And really what’s going on in the backend? Yes.

Mike: Awesome. So, I mean, one plus x is a data management platform. I mean, there are no end of data management platforms to help publishers and marketers. So can you just explain a little bit about what you do? And what makes you different in this market?

Carolin: Yeah, so I think what makes us different is that we always had this AI first approach. And this was something that I felt also extremely attractive when I chose one plus excessive DMP back then. Because, yeah, frankly, we had another technology before. And there were so many high expectations on using this technology. And I just remember, like a really concrete case of classifieds page that was offering a price comparison. And then people were so excited, yeah, we can finally use, people are searching for washing machines and directly sell them to washing machine providers. And yeah, obviously, the washing machine providers were quite excited about those news. And then in the end turned out like, there may be five kg of people interested in Germany in buying a washing machine, visiting a price comparison site. And still, the media currency was CPM. So basically, there was a lot of back and forth a lot of emails, a lot of high expectations.

And in the end, the marketer couldn’t spend budget because there was simply no reach and the cost and the publisher was obviously also frustrated, because there was also no money coming in. And so, this is the this was the setup that I was used to when I was searching for companies that have this AI first approach like how can we utilise the data that is coming in and predict on top of the seat said users that make sense to Target and I think that made us a little bit different because this was the approach for us from day one on and for the customers that are marketers particularly, we prepare to cleanroom product that is mixing both best of both worlds where you can upload as a marketer your data set set and then you can use the publishers embedded spaced like a publisher database, obviously, privacy compliant to do expansion based on your own seat set. So we have this best of two worlds approach, which I would say is also different towards other cleanroom solution, which I personally fear or going back towards the situation that I faced in 2016. When I tried to sell Yeah, a washing machine intend to campaign to washing machine marketers to be honest.

Mike: That’s true. The interesting I mean, you’ve used a couple of technical terms that might be be worth just explaining to people listening. So you talk about a cleanroom. What What do you mean by that? When, when you’re referring to advertising?

Carolin: Sure. So yeah, obviously, right now, because of GDPR, because of CCPA. It’s not easy to match data, right. So you can’t just, you know, use the third party cookie that you used to use like couple of years ago, and match your users in your publishers database. So if you would like to use your CRM data, or any kind of data source that you own your first party data, and you would like to find those users into an in the publisher audience, you need a cleanroom, which is making sure that you’re doing this matching privacy compliant.

Mike: Okay. So the cleanroom is, is making sure that you’re meeting those requirements from GDPR. And the other regulations around the world. Yep. And obviously, that’s hugely important. Because if you’re not doing that, then then clearly any products are not going to be able to be sold. So it’s an interesting term there. So what you’re trying to do is use AI to understand how people are thinking, so whether they’re considering it, in your example, buying washing machines or not, what is the benefit of AI? For the marketers who want to use that data as compared to using something simple? Is it just reach as you talked about? Or are there other benefits?

Carolin: I think, obviously, reach is part of it. So if you aren’t just referring or if you’re just using this technology that you always use, you will be always fishing in the same pond, more or less, especially as the internet goes blind, your opponent is drying out a little bit. So there will be less and less fish to target. And obviously, yeah, kind of it’s kind of getting harder and harder for you to find your right audience or retarget your audience. So AI is helping marketers significantly to keep reach up, and also reach, let’s say qualified leads, right instead of just random audiences. And yeah, I think that’s also a concept that is highly known, right? If you are uploading your email addresses to go or Facebook, or you’re uploading your audiences based on device IDs, and then you click on Generate look alike, audience. So I think this approach now is just getting more wider, wide. So we’re also trying to have similar strategies for the open web. So I think one clear benefit for the marketer is to use AI, to increase a, let’s say, qualified audience group. It’s, yeah, hopefully, also, or should be also still precise, right. So it’s not about just reach. But obviously, it’s about this holy grail of meeting that say, the sweet spot between reach and quality. And if the internet goes blind, if the third party cookie goes away, it’s also the only way I think, to at least try to have this more or less transparent customer journey tracking, maybe need to call it this way. Because, yeah, obviously, it’s getting harder and harder for you to know, with whom you were already in contact on which platform platforms are fragmented. So I think these are really big benefits also, for marketers to step into the game.

Mike: I love your use of the term, the internet going blind. And maybe it’s worth you just explaining that a little bit about what the the impact of the third party cookie going away means for marketers, and why that that means the internet then appears a bit blind to marketers.

Carolin: Okay, so yeah, basically, data is, I would say, also still really dominated by the demand side, if it’s not by the marketers directly than it is by the agency world. I just remember a world where, you know, group had a pixel and each and every publisher page worldwide to retarget, to user they were in contact with. So they knew like I users, they have seen, I think programmatic media buying is only based on audiences that you know, already, right, you don’t buy unknown traffic. But this is all based on third party cookies. And if the third party cookie is going away, because Google decided to restrict it as well and Firefox, it’s already blocked on safari, it’s already blogged, then it’s getting harder and harder for you to identify the user in front of the page. And this is what I mean when I say the internet goes blind. So the measurability and the matching of users is what is missing, but it’s still a basic concept on how programmatic works

Mike: That’s, that’s such a good, good explanation of what’s happening. And so what you’re doing is you’re bringing more intelligence, presumably so that publishers can understand, you know, or predict who’s likely to buy without having to track people all the way across the internet. So, so how are these publishers using this AI to generate products that they can then sell to marketers?

Carolin: Yeah, I really like this question, because that’s really where I put a lot of, you know, brain power in in the last couple of years. So I think that publishers have an amazing database. So I think, if you would like to understand this, this concept of marketing with AI, you need to have both, you need to have a really good database that you can use for predictions. And you could, you also need a seed set that this really high quality that you can use as a seed for your prediction. And I think the publishers are really good and having we call this the embedded space. But what is an abandoned space embedded space is basically your database. And on the publisher page, publishers see, or it’s the publishers I work with, they see the users nearly daily on two pages. So they know exactly what the user is interested in. And the users have a lot of different interest, right. So you can also check Google knows where a user is what he’s intending to buy, Facebook knows exactly the relationship starters. And where what the people do in their private life, by publishers really know what people are interested in, right, and what they’re frequently reading, and which sports club they support.

And if it’s more about celebrity news, if it’s about local politicians, if it’s about global economics, so they really know what users are interested in. And they have this tonnes of data points that they can use to build up really, let’s say, differentiated models, and that they can then use to, yeah, predict how these users who would I don’t know, squat towards a specific seat audiences if they have likelihood to be interested in this specific product or not. And that’s, this is the superpower that publishers to have that they can now use for marketers. And that’s obviously what they do already now. So with their classified data, with their looking data from the subscribers, that they use the seed set, and then they’re expanding those users towards a specific likelihood. The second thing that the publishers can also do is that they can use this data and enrich their assets. So the publishers basically have two superpowers, they have, on the one hand, the users, but they also have the assets. So what does this mean? It basically means that you can say I have a likelihood of 85%, this user is male. But I think I can say at the same time, I have a likelihood that this article is read by 85% male audience. And then even without knowing exactly that the person in front of the camera, or in front of the page is is male, you have this likelihood per asset that you can use to identify, which could be the right audience, even on the first impression, and that’s the second superpower that publishers have.

Mike: Fantastic. I love the idea of having two superpowers. And publishers cleaner clearly have a lot to add. I mean, if we look at the first superpower, I mean, what you’re doing is you’re actually saying that publishers have the ability to look at a group of people your seed set, as you say, understand what they’re interested in. And then what you’re doing is pulling in people with similar interests. Is that right?

Carolin: Yeah, definitely based on this, this behavioural data and all the data that you can collect from the publisher. Exactly. That’s what we do. We predict the likelihood for somebody to be in a specific segment based on all of this data interactions that we can collect.

Mike: And that’s great. But that is a little bit as you said, like the Google audience tools where you can have look alike audiences. The other superpower you said was was really interesting, which was around knowing who reads each page or who looks at the assets. See, can you talk to me a little bit about how you help the publishers understand, you know, who’s being targeted by each particular story on their website, so that they can then deliver ads that are even more relevant.

Carolin: Yes, sure. So, basically, again, you can do two things like you can crawl the the content, you can use the semantic understanding, you can identify interest out of the article itself. But you can also use the audience that you are that you are allowed to use. For example, you can use your subscribers and check them out and see how their interact with a specific article or video. And then you can make, again, a prediction and then you can use both information sources just to make a really complete prediction on the user itself, and then use this again, for users that you see for the first time that you’re not able to track anything about just to personalise your, the feeling of the user, this can be an ad, but this could also be, for example, personalise the page itself.

Mike: And presumably, also, there’s a benefit for publishers in terms of personalising the page because they can recommend content that the visitors more likely to read next. That’s, that’s great. So I’m really interested, you know, how do marketers approach publishers about this? Because that, you know, one of the things I see is a lot of the time it feels like publishers want to work with only their very biggest customers on the exciting stuff. And some of the smaller advertisers maybe don’t seem to get as much attention is, do you think there’s a way that more marketers can work more effectively with publishers and help them sell better services?

Carolin: Clearly, so I think, first, there is, I think, right now, there’s a huge demand for trying out this new partnerships, as Google postpone the decline of the third party cookies. So everybody’s still working in the old environment. But now, they really talk about alliances, Id alliances, for example, they talk about cleanroom setups. So I think, for smaller advertisers, there’s also always this possibility, I mean, obviously, you need to have an automated version of the solution. So I think, especially for smaller advertisers, if publishers need to do a lot of things manually, then it’s getting unattractive for them. But this needs to be the goal for the cleanroom providers to have like a 100% optimised data onboarding setup, that is allowing the publishers just to do this with a lot of advertisers, and not waste time on this 101 data transfers that are indeed unattractive for smaller advertisers. So I think, right now, the whole tech ecosystem is heavily investing in automating all of these setups. And obviously, we do as well. But I think in the future, this is exactly how it’s going to work. So publishers, will this just have plugged into their system, and then advertisers can use it for matching, and then for expansion purposes without the need to go to the publisher, and, you know, do something by hand on their side.

Mike: And certainly, I think that’s, that’s a really good point, reflecting a lot of marketing, if you can’t automate it, it’s very difficult and very expensive. When you’ve got a tool that can automate the process, it becomes much more accessible to you know, pretty much everybody. So yeah, I love that idea. I mean, one of the challenges I see, particularly with with your solution at one plus x is it’s obviously very heavily reliant on artificial intelligence and machine learning. And realistically, none of the marketers or the publishers are experts in this technology. So how do you have a conversation with your publisher or a marketer about this technology, when you don’t have a deep understanding of how it works?

Carolin: I think you don’t necessarily need to have a good understanding on how it works in the backend. I also honestly didn’t happen before I joined one plus six. So you just need to trust us that we are really experienced and exactly that. And besides that, the rest is only onboarding, which is quite lean, and not so complicated to do. I think in I think, in general, it’s just important to be open and open minded, because I also see in the industry, some people are just afraid whenever there’s AI on something, they’re like, Okay, I’m not going to understand it anyway. So I’m not going to try it out. And obviously there are also the people that are still really focused on their gut feeling. Because that’s something I need to admit right if you would like to. If you would like to use AI, you need to trust the AI. You need to say this is my seat set. And now I find the right users for me, if you keep saying but they need to be male, they need to be between 35 and 45. And they need to have high interest in in buying cars, but only convertibles then it’s not really needed us, then you can still use the standard segmentation that is offered somewhere, right? So I think this openness is something even if you don’t need to understand how it works, but you need to trust this algorithm and you need to be open to tested. And that’s something that I face quite often that there’s still this personas going around that the market research institutes created, that the media agencies can only try to rebuild based on data. And it can be first party data, second party data. But yeah, if you really would like to try out this cleanroom approach, you need to be open, that you don’t know exactly before you run the campaign, how other people look like that you’re targeting, because whom you were targeted, is not defined by your gut feeling, or your research is defined by AI.

Mike: I love that. And I think that’s, that’s such a good indication of how marketing is changing. You know, previously, people used to create an ad, for example, an ad everybody decided they loved it, they ran it in printed publications, nobody really knew whether it was effective. But if everybody liked it, that was great. And we’re moving to the situation where, you know, Google on Google ads will tell you which headlines work and which headlines don’t. And it doesn’t matter what you think you’ll know which ones work and which don’t. And, you know, it’s very humbling to be wrong. And I’ve certainly been wrong on that. But, you know, with with products, like one plus x, you’re actually helping the market or defining for the market or a lot of the audience. And that is another, you know, step for a marketer to trust technology to deliver the audience rather than to define it themselves. That’s, that’s fascinating. I mean, I think, you know, there isn’t obvious questions. Well, there’s been quite a lot of products that have hyped their use of AI, both in marketing and other areas, that actually have been very disappointing when people have tried them. So I mean, why do you think some of these AI products have failed, particularly in marketing? And what are you doing at one plus x to make sure that it’s not just applying technology, but it’s generating a real benefit for marketers and publishers?

Carolin: I need to say the real differentiator is the consulting, I can give you a really concrete example where I failed, to be honest. And maybe that’s also something that not companies talk so open about, but I will just do because I think it’s important. So I had a customer, it was a really nice customer, and they had a portal where you could buy tickets, and they had a lot of amazing ground truth data, they had like 3 million Lockton users. But the problem was that the specific tickets you could buy, obviously, it was a transportation provider. So there are differences if you’re travelling with kids, if you’re travelling business related. But in the end, everybody is booking a ticket the same way. So even if you have a lot of data, all data sets look the same. And we started to try out the algorithm, like our algorithms and their database, and then the machine learning expert came to my place. And he said, you know, what the seed set is shit. And I was like, No, this can’t be the case. I’m 100% sure the seed set is amazing. It’s locked in users, they are verified, so no way. And then I started digging deeper.

And then what we found out is it was simply not working for this particular marketer, because he had a lot of data. But the data was so similar that you could not predict any patterns that made sense. And this was the moment it’s already like quite some time ago, when we thought, okay, we need to pivot. You know, we need to, we need to bring those two worlds together. Because publishers are really suffering from, let’s say, seats, it’s a lot of people are anomalous, a lot of people will never buy a subscription, especially in specific age groups, right? It’s a, it’s really a bummer. But it’s the status quo. And the marketers, sometimes they have a lot of data. But if it’s a platform that is not a, I would say, an online store, or if it’s not a classifieds side, where you can really see differentiation, and it’s going to be hard to use AI. And even the best trend algorithm is not able to do anything that makes sense with this data set. So I think a lot of products failed also because of this missing consultant and dismissing reality check. And that’s also why we came up with this connect idea to connect the the strength of two sides to build something new on top of that. And I think that’s one of the reasons why consulting is really the differentiator like not just accepting what the AI is telling you, like the data set is shipped but really go there and understand why is this the outcome and what can we do to change it and then be open to pivot and yeah, just go completely change your system towards a new architecture in case it’s needed.

Mike: That that’s amazing. Because I think, you know, sometimes people think, well, there’s some technology, we just apply it, it’s gonna work. And, you know, it’s great to see that, actually, you do need good data, that’s gonna work and you’ve made the point with, with audiences, the audiences need to look different. If they look the same, then there’s nothing to say. So I think that’s great. So it’s about understanding the data, and that needs experts that needs people to come and provide that consultancy. So I love that as an explanation. That’s fantastic. So, I guess if people are excited, they believe that AI absolutely can help them. I mean, how do they get started? You know, is it best to rush in? Should they be talking to a provider? Like you? I mean, how should people start adding AI to their marketing? Do you have any advice for them?

Carolin: Yeah, I think screen what’s in the market? Like maybe do some basic checks, like I just told you, like, do you have enough data? Do you how many data sources do you have? How big is your data? Silo right now? Do you think that you alone with your own data will be able to have prediction that makes sense? And if yes, then try out some some some tools. If not, then search for solutions that will help you, for example, that are allowing you to run your own train your algorithms in I don’t know better environments for more precise outcomes. Obviously, a cleanroom, I think has a, like the solution like that one that I just explained where publishers and advertisers meet, there’s a relatively low entry, because you just need to find one publisher who is open to do it with you, you need to try it out, you will have like one test campaign, you can a B test, do it now, now that you still have a third party cookie just to check if it works, right. Because now you can really do a B testing in terms of performance. Don’t shy away the first moment the first campaign might not have the results you were desiring because they’re always you need to add optimization, you need to add some more brainpower. But I think it’s not so complicated to start, if you are searching for tools that might help you to overcome your personal challenges. And this doesn’t, I think that the biggest problem a lot of companies in enterprise have that they think, Oh, I’m building this all on my own. I’m I have such a great tech team, I have such a great Data Silo, I would just build everything in house and then it simply takes too long. So I think here we have again, this, tried to find like an MVP, tried to pivot your ideas and fail but fail fast. I would say this.

Mike: Yeah, I love that. Just give it a go and see what happens. And don’t be worried if it doesn’t work first time. That’s, that’s great advice. I mean, obviously, you know, with your product, it’s particularly around serving ads and marketing content that way, but how do people really understand that the system is working? So do you integrate with other parts of the marketing technology stack to help people measure performance? Or is it very much an independent product?

Carolin: No, like obviously, you use your data audiences in your in your activation channels, like whatever activation channels you have from obviously from from media buying to email marketing everywhere where you you can utilise that. Furthermore, a lot of our publishers particularly are challenging our machine learning algorithm again, market research panels, and as the third party cookies to their so we’re getting challenged a lot against nears and for example. So we are really used to getting this external feedback, and are really proud of us there because, yeah, this is really our bread and butter business. But in general, yes, we are completely integrated in this in this edtech system. We are also completely I think one of the biggest or I say one of the biggest advisors or most important things for tech providers is not to be a standalone solution, right. But to fit in perfectly as a puzzle piece and most of the mahr tech stacks that companies use. And we hopefully we are with our API’s and raw data access and all this touristry provide hope that we are fitting seamlessly in most of the martec ecosystems and stacks that the customers are choosing.

Mike: Awesome as that sounds great. That sounds like you’ve really thought that one through. And this has been fascinating. It’s been really interesting. I’ve actually feel I’ve learned a lot about AI as well which is great or a lot about what you need to think about when you’re using is AI? Is there anything else you think we should have covered in this discussion?

Carolin: Yeah, no, I think we covered it quite well, I think trying it out now, while the third party cookie is still there, to have this ability to see, the potential new world in the still existing world is like, I would say, this is a luxury, we will not have in a couple of months. So I think for the marketers, the, you should now urgently move towards this kind of directions, because you will learn so much, right now. As soon as the third party cookie is gone, we will rely on all those cleanroom solutions matching partners. And that’s just maybe too late, right, because you would like to see the before and after effect yourself and your own data. This is one thing. And the second thing is obviously it cleanroom is also dependent on identifiers. And you should invest in ID alliances or check where you can join alliances in general, which IDs you can provide. Because if you have the best data set in the world, if it’s not matchable, with anything else, then you will not be able to find your audience even in the most sophisticated publisher embedded space you will find. So that’s maybe my second advice, I would say for marketers. Yeah, and then obviously stay open and let the AI do the magic without trying to influence the algorithm with overfitting.

Mike: Trust the technology. This has been brilliant. I mean, I’m sure people would be interested to know more about one plus x and what would be the best way for people to find out more about the product and also contact you if they’ve got any questions they’d like to to ask you.

Carolin: Yeah, so I think the website is a good starting point, you can always click on book a demo. And depending on what you would like to see, we can do a bit of consulting, or we will show you the platform. So just get I don’t know, your our neutral view and things may be in in a little session together with us. If you have specific question to me, you can also find me obviously on LinkedIn. You could also reach us on LinkedIn as a company.

Mike: Awesome. This has been great. I really appreciate it. Carolyn, thank you so much for your time and for being on the podcast. Thanks so much for listening to marketing b2b Tech. We hope you enjoyed the episode. And if you did, please make sure you subscribe on iTunes, or on your favourite podcast application. If you’d like to know more, please visit our website at Napier b2b dot com or contact me directly on LinkedIn.