Towards Business Success with Smart Data
Behavioural economics and the digital transformation have something essential in common: customer orientation.
Insight in brief
- Machine Learning Algorithms and Artificial Intelligence offer the possibility to approach Behavioral Economics.
- Use cases are pattern recognition in shopping lists, heuristic prompts, personalisation and salience.
In behavioural economics, we examine why people do what they do. Thanks to digitalisation, the customer is given a stronger role in the market. I see shared possibilities in relation to the interaction in the virtual and augmented world with AR/VR technology, for example. One idea is that the insights gained from behavioural economics will interactively help the customer when making decisions. I mention online shopping here as an example.
I see the amalgamation of data analytics and behavioural insights in even more concrete terms. With Smart Data as an aspect of digital transformation, our customers provide us with new incentives. This enables us to address them more individually. Customers can be engaged at a more selective trigger point within their customer journey. Using data from various sources, we can analyse their behaviour and make data-based predictions. Data analytics leads us to rational, measurable insights.
The findings from data analytics and behavioural economics can be usefully combined. (Zühlke)
Behavioural economics uses experiments and investigations to show us how people generally react in certain situations. This is because the same customer A may react differently in situation ZZ than in situation XY, although the data predicts that in fact our customer A should always act in accordance with behaviour pattern XY. We must not forget an important point: the context, the big picture. What experiences have customers had in the past, what mistakes in thinking do they make, and where is their point of reference? This is precisely what plays an important role in behavioural economics.
Patterns in the shopping lists
With modern machine-learning algorithms and artificial intelligence, we now have the opportunity of moving from absolute, rational data to aspects of behavioural economics. There are also interesting interfaces and long-term combination possibilities in connection with behavioural analytics. I would like to take this opportunity to bring in my colleague Nadja Ulrich, Data Scientist at Zühlke. Using an example in the B2C sector, she shows which amalgamations would be possible in the future; how the success of behavioural analytics can be even greater with the insights from behavioural economics.
"In partnership with a customer, we took an app that had already been developed to help end customers with their daily shopping, and we expanded it with the clear goal of offering a customised user experience. It was about the approach: how can Bring! generate value from its data by using it as the basis for creating new, innovative customer experiences?
The specific initial starting points were the purchasing cycles and product combinations. With shopping basket analyses, for example, the purchasing themes are determined and association rules are automatically extracted from the data. This allows frequent patterns in the shopping lists to be identified, visualised and analysed. Determining purchasing themes, or rule-extraction, are classic tasks for machine-learning algorithms. But if we supplement the context with empirical insights from behavioural economics, there are many more possibilities: on the one hand, directly in terms of content in the design of the algorithm – when we draw on the findings of Descriptive Norm and Social Comparison, for example. In other words, people orientate themselves by the behaviour of their fellow human beings. They compare each other in order to behave in an acceptable and desirable manner. An exciting further step would therefore be not only to form the purchasing clusters thematically, but also to cluster those themes that appear frequently and those that never appear in a list."
Tailored to the customer
On the other hand, thanks to the theories of Behavioural Economics we can also directly derive further exciting tasks: for example, heuristic prompts – i.e. hints to perform particular behaviours at the right times – could be used. This can be extended from individual foods to seasonal offers. The automatic selection of the display is then the machine-learning job.
Personalisation and Salience is certainly also a classic – that is, we are more likely to pay attention to information and products or services if they are tailored to us personally. This is very useful and can be pushed even further – for example with recipe recommendations based on my purchases. It is therefore a personal shopping assistant who also understands my behaviour when I'm planning my shopping.
I see great potential in combining the two sciences. Both are ultimately about decisions. As in any relationship, there is a certain dependency as well as a certain interaction. If one uses the generated and analysed data as a basis and combines the insights from behavioural economics, new possibilities emerge, especially in B2C. Customers can literally be "nudged" towards their good fortune. In a similar manner, all components of the digital transformation also enable new interaction possibilities between customers, and with customers themselves.