Findify - Dataset enrichment tool for e-commerce stores

Purpose and goal

We built a live-traffic prototype of the technological infrastructure that enables automatic enrichment of ecommerce datasets. We use external datasets, clean them, and then use machine learning to automatically cross this data against the customer´s internal dataset, and the data we mine from online shopper behaviour in the store. The entire process is autonomous and scalable, eliminating what otherwise would require a significant number of man hours and with less chance of human error.

Expected results and effects

The project achieved all KPIs: Lower bounce rate on search queries associated with the output of the tool, higher engagement of online shoppers, and an overall increase in conversion. More than 30 customers adopted the solution using it at the beta stage. Customers have indicated that the enrichment tool provides them with insights they´ve never seen and reduced operational overhead in attempting to achieve the same results manually.

Planned approach and implementation

We´ve designed the tool to be very efficient in collecting the data from the external source. Once a single source is collected once, it can be scaled to all customers. Our machine learning algorithm is efficient, in that it requires minimal training since it uses crowdsourcing techniques; customers accept/reject enrichment suggestions which in turn trains the model. The tool integrates seamlessly into our search solution so the customer simply clicks a button to reap the benefits it delivers.