Anyone who visits online shops gives an indication of his or her buying intention by means of his or her clicking behaviour: Does the user have something specific in mind and does he or she definitely want to spend money? Or does he just browse and possibly leave the website without buying? An analysis of the click behaviour can increase sales figures.
Our project partner from the e-commerce industry wants to get to know the customers of his online shop better. For this purpose, the behaviour of website visitors is to be analysed in order to distinguish determined buyers from occasional surfers. By means of different parameters such as click frequency, users can be classified into categories - if a customer navigates very specifically through the product range, he is more likely to buy a specific product than a customer who seems to browse the offer in a relatively erratic and random way. Depending on this "customer journey", effective sales incentives can be set.
This is our contribution
With the AI method of a "neural network", we can visualize connections of varying strength between behavioral patterns and actual purchase decisions of users. In this way we enable our customers to make even better purchase forecasts. Since we also provide the customer with long-term database maintenance as part of a further project, the network learns The adaptation of the network to updated data is set up automatically - so the network learns with every order and every wish list entry in the web shop.
Thanks to our neural network, the mail-order company knows which undecided website visitors it is encouraging to make spontaneous purchases by means of targeted special offers - and which targeted users it is better not to distract by such insertions. In this way, sales can be improved and customer satisfaction increased at the same time.
The project at a glance
- Establishment of a neural network to analyse customer behaviour