A global meal kit delivery company generated large quantities of textual customer feedback for analysis. These unstructured data contain a lot of information that could potentially provide valuable insights for product management, such as product and operational improvements.
However, the current techniques being used did not perform well in translating these data into actionable insights.
Amsterdam Data Collective developed a Natural Language Processing (NLP) based model to conduct open text customer feedback analysis at scale.
The model shows the performance of the company and its products in various categories, such as price-value, product, customer experience and operations.
Per category, the development of consumer sentiment and the number of reviews are shown, together with the most important subjects, to allow for the most important customer feedback topics to be analysed. For example, the product related reviews are automatically divided into topics such as portion size, recipe, flavour and meal choice.
Impact and Benefits
The model provides product management and other departments with actionable insights for product improvements. Additionally, the model signals potential operational issues in an early stage based on the data from the customer feedback analysis.
The first version of a NLP model for customer feedback analysis can be developed in just a matter of weeks. The model is further optimised based on the first outcomes and stakeholder validation.
Do you want to know more? Get in touch with Hans van Avendonk at firstname.lastname@example.org or check our contact page.