Marketing campaigns are more successful the more personalized and accurate they reach their target group. This increases the conversion rate and prevents frustration and newsletter cancellations by spammed customers. But addressing fine-grained target groups individually is complex, time-consuming and expensive. AI-based solutions can reduce the effort, save costs and improve the efficiency of marketing activities. In an interview with Lünendonk, Dr. Lucas Calmbach from KPS AG provides insights into the application of AI in marketing.
Where do you see a need for the use of AI in marketing?
Lucas Calmbach: Many marketing departments invest enormously in solutions to create content that meets the needs of their customers. The goal is to personalize the assets to fit their needs and to tailor them precisely to individual characteristics and requirements. This usually takes the form of several marketing campaigns that differ from one another in terms of content, which leads to a high degree of complexity in the prioritization of campaigns. Finally, there must be no overlap and thus no flooding of the customer with marketing content. Nobody likes to be spammed. With AI, the effort and complexity are reduced considerably. In addition, the marketing content can be personalised more precisely and thus address the customers more precisely.
Why are classic automation potentials not sufficient for this?
Lucas Calmbach: Classic automation potential is based on fixed, mostly rigid rules. These would have to be adjusted manually again and again to improve the degree of personalization. AI, on the other hand, enables dynamic and permanent optimization of the logic and, as a result, the continuous optimization of marketing content in the individual customer approach.
Does this mean that AI takes over the selection of marketing content independently?
Dr. Lucas Calmbach: Correct. The idea is that the marketer no longer dictates which customer should receive which content, i.e. which images, texts or personalized offers. From now on, the AI takes care of that. As a result, each customer will receive different marketing content, according to their individual characteristics and needs.
What about customer targeting? Is AI going to do this as well?
Lucas Calmbach: Indirectly yes. In the past, we created many different and fine-grained target groups that were sent different marketing content. Today we address a large target group with different information in every single case. The AI determines the exact content at the time the campaign is being launched. The selection is therefore no longer made at the beginning through target groups, but at the end by means of content.
Can we speak of a new generation of segmentation and personalization in this context?
Lucas Calmbach: Absolutely. The approach delivers a significantly higher degree of personalization with less effort for the marketing department. Conversion rates are also significantly higher than with the traditional approach.
In what way is this more efficient?
Dr. Lucas Calmbach: The marketer no longer has to tediously define the different target groups and assign appropriate marketing content to them. This is now done by using AI. It also eliminates the complex administration of finely grained campaigns to avoid flooding customers with (non-relevant) marketing content.
Is AI limited to the selection of content?
Dr. Lucas Calmbach: Currently, yes. In addition, AI can automate translations for different content and adapt it to the language of the recipient.
Can marketing content also be created using AI logic?
Dr. Lucas Calmbach: When creating content, it still makes sense to specify the structure and thus define the scope of the marketing measure. AI can, however, fill the structure with content - i.e. marketing content that fits perfectly - at the customer level.
What does the approach you have developed require?
Dr. Lucas Calmbach: In essence, a tool for marketing automation is required. This tool works with a large amount of customer data from various sources. It's the only way to determine individual characteristics and needs at the customer level.
What kind of data is this?
Dr. Lucas Calmbach: On the one hand, it is customer master data such as age, gender and other personal information. On the other hand, it is behavioral data on previous purchases, clicks or interests.
What do you mean by interests and where do you get this information from?
Dr. Lucas Calmbach: Modern marketing automation tools make it relatively easy to identify the specific interests of customers. For example, if a customer clicks on a link to a whisky, it signals an interest in the spirit. If the customer repeats this more often in a certain time window, the marketing tool automatically assigns the interest "whisky" to him. AI in turn logically incorporates this insight into the determination of the optimal marketing content.
This means that AI learns with every marketing campaign?
Dr. Lucas Calmbach: Correct. As the number of marketing campaigns increases, so does the interaction with customers. This in turn generates more data that AI uses for personalization. So it's an iterative process that continuously improves customer knowledge and personalization.
To what extent must AI be trained at the beginning of the iterative process?
Dr. Lucas Calmbach: At the beginning, there is very little data available. Now the data scientists are required to train the AI. They identify customer master data and behavioral information from other systems and "feed" it to their AI-tool as part of a learning process. Only a trained AI can be used for interaction with customers.
What role does social media data play in this process?
Dr. Lucas Calmbach: Social media channels have been connected to marketing automation tools for a relatively long time. Technically, this information can be very easily linked to customer master data. However, restrictions result from data protection regulations.
Which system do you recommend for collecting relevant customer data?
Dr. Lucas Calmbach: We rely on standardized solutions for marketing automation. These aggregate data from other systems, such as customer master data from a CRM solution, sales from an ERP system or clicks from a web shop. The marketing solution then bundles this data into a so-called 360-degree customer profile on which AI is based.
Now that sounds as if integration is a central task. What other roles of employees do you need for your approach?
Dr. Lucas Calmbach: Right, employees with integration expertise are indispensable. But we also need data scientists to define AI and content developers for the initial creation of marketing content.
Would you describe your approach as particularly complex?
Dr. Lucas Calmbach: From a technological perspective, there are certainly more complex AI approaches, especially in medicine. But what is much more exciting is the paradigm shift and the change for marketing departments that our approach brings: Whereas in the past a large target group was provided with exactly the same marketing content, today many finely granulated target groups are provided with different information. In the future, a large target group will once again dominate, within which each customer - in contrast to the past - will be targeted with very individual and specific content. This paradigm shift has the potential to significantly streamline existing processes and reduce costs. In addition, marketing content can be tailored more precisely to the needs of customers, which promises higher sales.
This article was originally published in the Lünendonk® Magazine 2020, issue 05, "Künstliche Intelligenz":
The Lünendonk® study is available for download in German language:
The 80-page "Künstliche Intelligenz" magazine print edition can be requested from KPS or Lünendonk.