How you can control the user journey through real-time personalization in a targeted manner – eCommerce Magazin

AI algorithms in use: high quality Real-time personalization technologies, in which machine learning takes place using deep learning networks, are considered to be groundbreaking in optimizing the customer experience. They allow it for each individual Visitors to a company’s digital touchpoints Determine product affinities, purchase probabilities and risk of churn in real time and immediately deploy coordinated marketing measures.

You can usually for every objective work with industry-specific algorithms that are trained for the respective use case and whose analytical and action-related performance increases with every user action (click, interaction, view, etc.). The time and intensity with which each individual user deals with the content play an essential role.

New dimension of customer experience and new efficiency of 1: 1 personalization with AI algorithms

With previous personalization processes, it is possible to optimally guide interested parties or customers at the moment of their action. However, marketers are increasingly asking what specific action the user will take next and how long he will be involved in a session in order to be able to optimally manage the user journey in terms of the time available and the touchpoints that are likely to be used.

Processes that work with AI algorithms of the LSTM category (Long Short-Term Memory Networks) enable companies to predict the user journey in detail and to control the users – customers and prospects – in a conversion-optimized manner. LSTM algorithms, which belong to the group of recurrent neural networks and, due to their complexity, have so far hardly been used for real-time forecasts in marketing, have primarily been used in the areas of text and speech recognition. They enable precise prognoses on influencing factors that are directly relevant to the transaction, such as

  • how long the user will stay on the website,
  • how long it will take to convert,
  • which pages of a website the user will visit next.

With the knowledge of these factors, marketers can now manage the user journey individually, fully automatically and dynamically.

Use Case Next Best Action (NBA): The Next Big Thing

The hierarchy and complexity of a website make it difficult for users to easily find their desired destination. Personalization can help with this.

AI algorithms
Figure 1: Guide users efficiently through the website. Image: Smart Digital

The “next best action” is an offer to the user to navigate directly to one or more landing pages within a website hierarchy, which are probably the next best steps for him on the way to an efficient conversion.

Practical example: An automobile manufacturer wants to increase sales of its models. For the purpose of optimizing his on-site marketing ROI, he relies on an AI-based prediction of user interest during the first session, and not just target group identification. For this purpose, LSTM algorithms are used as part of real-time personalization. The aim is to use it to introduce potential buyers to the topic of e-mobility and thereby generate additional sales.

For a visitor to the manufacturer’s website, the LSTM-KI used as part of real-time personalization now predicts which content he is interested in with what probability based on his usage behavior after the first page view.

AI algorithms
Figure 2: Scoring interests using LSTM algorithms. Image: Smart Digital

With the anticipation of the next steps and touchpoints, individual user guidance for purchasing decisions can now be implemented with highly dynamic and fully automated communication tailored to individual needs. Content is only offered to the user if it is likely that it will be relevant to him and promote a conversion.

The user is no longer intrusive in the sales funnel, but receives the most relevant message and the right content at the right touchpoint at all times.

AI algorithms
Graphic 3: Conversion-optimized user experience.
Image: Smart Digital

The result: the lead rate was more than doubled during the pilot phase.

Conclusion on AI algorithms

AI applications with the LSTM method not only have a significant leverage effect on the efficiency of conversion optimization. With the possibility of providing every single visitor at every digital touchpoint with exactly the content for user guidance that actually interests them at that moment – and before they can formulate the need for it themselves – creates a completely new dimension of the customer experience and achieves one Unprecedented new efficiency of 1: 1 personalization.

Also read: Customer interaction: what will the future look like after the corona crisis?

AI algorithms
Image: Smart Digital

About the author: Gregor Wolf has more than 20 years of experience in building and scaling companies in the digital industry. Before joining Smart digital The data and marketing expert was managing director of Experian Deutschland GmbH, one of the leading international cross-channel marketing specialists. Other stations in his professional career include the Global Group, Schober Information Group GmbH and United Internet Dialog, in each of which Wolf held managerial positions.