Editors’ note: Recently, Selligent Head of Partnerships – Benelux, Nicolas Van Vijnckt, sat down with two experts in marketing automation: Alexei Kounine, VP of Product Management & Innovation (Selligent) and Michael Straathof, Commercial Director & Co-founder at Selligent agency partner 100%EMAIL, for a roundtable conversation about the importance of data and how artificial intelligence can assist brands and marketers today. These topics are addressed in a synergistic way, from both a technological point of view (Alexei) and the perspective of the marketeer and marketing consultant (Michael). The conversation will be presented in a multi-part series on our blog, and we hope you find the insights inspiring and practical.
NVV: Marketeers have many questions around data and AI. For many, these topics still sound quite undefined, like buzzwords. Are these new exciting developments relevant for all digital marketers? Does using AI ask for a well-defined context? Can I afford this and is my organization ready for it? These questions and many more will be addressed by our two experts, with their specific views, creating a unique synergy between technology and strategy.
NVV: AI-driven and personalized experiences are at the centre of reaching out to your target market. Multichannel campaign management is obvious today. Michael, how do you see that brands today think about using data? Are brands aware about the value of data and what data is interesting to manage?
MS: Thank you, Nicolas. 2021 is a data-driven era, whether you’re talking about marketing, sales, product development, or service. Data is in everything we do these days. It reminds me of my father, who had a little shop in The Hague. He knew every customer by heart; their name, where they live, what they bought last time, what their interests were. Also, he asked them for feedback. The basics of what marketeers do today and what he did are still the same. The difference is that the volume of available data has exploded.
It’s not human intelligence only anymore today. It’s becoming artificial intelligence. You have to be able to store all that data in the right place. You also want to make use of that data; the profile data, interest data, data of what people bought last time. You want to have that stored; and secondly, more important, you want to use that data, so that the customer feels he is being recognized. That’s the major part of why we’re all so busy with data in marketing, and why brands are thinking so much about data.
What is important to manage? We’re talking about demographic data, but also third-party data may be added. Do you know what kind of car he drives, what newsletter he reads, or what other interests he or she has?
NVV: What you are talking about is consumer behavior. The story of your father shows that a customer has a very unique personal profile. Today, in a much larger, digitized economy, we see this reflected in all the types of data we collect. But of course, we do not want to only capture the past data, but also want to influence future behavior – facilitate upselling, prevent churn, etc. Alexei, what Michael says, does this have an impact on how you structure your data and make it flexible to be future-proof?
AK: First, we’re talking about consumer behavior and the type of data that a consumer leaves as a trail behind. The Selligent platform is about understanding all this data and making predictions on top of it.
If you look back twenty years, the trend was that consumers have profiles you could segment and it’s quite easy to send one email to reach a lot of different people. As a marketer, what you track is engagement, based on consumer interactions. Doing that, at a mass level, was something that worked at the time, because email was rather new. As consumers start using more channels, they leave more digital trails, behavioral data: essentially views, clicks, etc.
The second trend was omnichannel data and how you not only represent one consumer and his interaction versus one campaign, but how do you actually represent one consumer from a data perspective, across all the different channels? You’re not looking at a mass of persons in one channel, but one person across many channels. This is where concepts like the Universal Consumer Profile, or “golden profile,” started emerging. And with that, we started having more and more complex data. So how do you actually store all these millions of interactions and how do you make forecasts on them? At Selligent, we store all the data around each consumer across all channels, like a star model, which allows you to segment things easily and predict.
Now we’re at the stage where we have so much data, that if you want to offer personalized experiences, you need to use this data and make spot-on predictions on what the best content is, what the best send time is for a specific campaign, etc. Once you manage making prognoses, you already offer each consumer a relevant experience. This is where most brands are today. The next step is about real time, doing everything we do now, but closer to the moment in which a consumer would be engaged. So we have this evolution from mass marketing and one campaign for many millions of contacts, towards very specific and in real time using predictions and, therefore, AI.
NVV: You’re talking about engagement based on consumer interaction, requiring an omnichannel approach. Omnichannel these days is not really a choice anymore. Your customers ask for alignment across every touchpoint, and this can only be achieved if you take care of your data. Data is an undeniable asset for your company. As the market shows an ever-growing awareness in the deployment of AI in digital marketing, will it continue in the coming years? Are we just at the beginning or not – and do you see impact on marketing performance?
MS: I see some real mature examples and also companies that don’t have a clue yet; they are busy storing data, but don’t know exactly how to use it. Making predictions, we’ve been doing this for over forty years. The thing is that we have so much more data points included in this propensity model, that it’s becoming harder and harder, for the tools, but also for the datasets. So it’s really a good time now to start thinking about how we can do this more efficiently, so that our infrastructure is able to manage all this data; but also, maybe even more important, that marketeers know what to do with all this “gold” they have in their hands.
So yes, there are some really good examples. But how do we make this more approachable for the mass-market marketeers, who want to be relevant at every contact moment, as Alexei mentioned, in a real-time manner? At the moment of contact, you want to have the right content in place. If they open your website, if they call your contact center, or whether you want to send an ad on Facebook or LinkedIn, whatever channel you’re about to use to be in contact, you want to be as relevant as possible based on the propensity you have in place.
NVV: After this interview, we will have a webinar where we elaborate on these examples. Alexei, do you have something to add on what Michael just said about the maturity of different brands regarding AI?
AK: Yes, I agree with Michael, that many brands are at different levels. But if you’re a brand that is not considering AI as being an important piece of how you run your marketing, you’re making a huge mistake, because machine learning is the next wave in which people will automate things. We are talking about omnichannel experience delivery to a single client. What you actually want is predicting in a much more granular way per individual.
Adding to my previous example, it was possible to do that on millions of contacts at once, sending a message to that many people. If you want to give a personalized experience, you need to collect data about these consumers across channels; you need a feedback loop collecting data from all channels. Even more than before, you need tools to automate the decisions in your marketing journey. We’re used to using rules: if a person is aged between something and something, you send this message; else, you send another message. But this is a very low level of personalization. What you actually want is systems that predict, in a much more granular way, per individual.
I expect most brands to use AI-driven tools, even without them knowing anything about AI. It’s about using alternative automation techniques that exist today and are that available in platforms like Selligent, out of the box.
NVV: Today, we see already a lot of marketeers start thinking about AI, but hesitate. “My brand is not big enough, I have too few data. The type of data doesn’t give me useful insights.” They often use arguments for not doing it. Alexei, you talked about granularity. Can you start small and grow in the process?
AK: Definitely yes! And are they right not to consider it? Definitely no! Because it’s very easy to use out-of-the-box tools that allow you to do very simple things. I’ll give you an illustration. If you send campaigns only to a few, ten-thousand people, you may think, “Oh, my audience is not big enough to use machine learning tools.” This is what you would think if you had a team of data scientists that are waiting for data to flow in from your email marketing tool. But this is not the way you should think about this.
There are two different angles of looking at this. First of all, email is only one channel in a whole range of channels that you can use to track individual behavior. The second thing that works really well with email is a website, because it has data that works intrinsically well with email. You click in an email, you land on the website, and so on. Getting email data, coupled with web data, actually helps you improve your email marketing by using machine learning tools on top of all the data you collected. And it works, even for very small clients with a low number of products in their catalog, with low web traffic, and so on.
The second angle in which you can see that is, “Oh, I’m a small brand, I don’t have the resources to look into machine learning tools, because it’s way too technical and complicated”. My answer is you should not look at building a data scientist team, but rather use off-the-shelf tools from your marketing automation solution, that do machine learning for you, without going through the tedious process of hiring data and machine learning experts. Most of the marketing automation solutions today, including obviously Selligent, offer tools that use data for making predictions.
NVV: Michael, based on your experience as a marketeer, how can you introduce AI at a brand?
MS: It’s a matter of time before the majority of marketeers get used to these new kinds of techniques. In your personal life, you’re already used to it in a lot of different ways. For example, if I watch TV, if I go on Netflix, what kind of movie is being shown to me which is in my top 10? That’s AI. If I listen to Spotify, what kind of music is in my playlist today? That’s AI. If I go to Albert Heijn and it predicts what kind of groceries I need this week, that’s AI. There are so many techniques already based on data in your daily life, that it’s only just a matter of time before marketeers get the real thing. And it’s not something new, it makes life easier and more effective for marketeers.
When I talk to these brands that work with 100%EMAIL, I ask, “What kind of contact moments do you have? And what kind of data do you collect and how do you collect it and how do you combine that in your next best offer?” And most of them, they do have an idea once you explain how to use it. And cost-wise, you don’t need hundreds of thousands of euros to start working with it.
NVV: What you’re saying is you should not only be concerned with the data itself; it’s also about the technology. You gave a few examples of use cases. Realistically, what can you expect on the return on investment? Does it differ from sector to sector?
MS: We just talked about email and if you do it in a more relevant way than “batch and blast” method, it’s a return on investment of one to 40. So, how much do you spend at Google or Facebook on your advertising when you can also use your first-party data in a more effective way? That’s always my question if they say, “I don’t have the budget.”
NVV: Let’s look at how you can drive a project. What is the context you need to successfully deploy an AI project? Consider the data and the technology you already have? Do you have to revise the data processes? What could be the objective or goal?
MS: It starts by defining or making the customer journey visible to your marketing team. Start drawing all these contact moments and start drawing what the information need is in which stage, what channels do we use in which stage of the customer journey. And then, when you have a vision on your ambition of how you want to contact your customers, then you start looking at how the data is organized and the technology you have in place, in order to facilitate the marketing automation process.
Start by drawing your ideal customer journey, see what’s happening now, whether the customers are happy or not, and then write down what’s your top priority to optimize. Which stage of your customer journey, in the short run, is delivering the most value? And start there, just start in month one, start by optimizing that moment of the customer journey.
© 2023 Marigold Engage
Marigold: where relationships take root.
