It was early 2016 when Marc was running a webshop and grew frustrated about the bidding process on Google Shopping. How could it be that this crucial element that was so decisive for the profitability of a webshop was handled so poorly? He shared his frustration with Jan, an ex-banker that had turned full-time developer of real-time trading algorithms on the S&P500.
While discussing, they got convinced that there are many similarities between buying promising assets on financial markets and buying promising clicks on ‘click markets’. However, there was one piece of the puzzle that needed to be resolved: “How strong is the purchasing-intent of the prospective customer?”. How likely is a visitor to buy a certain product? That’s where Alex came in as a hard-core Data Scientist and Statistician, specialised in sequencing and analysing DNA structures. By using advanced scientific techniques he started analysing search queries to distill the strength of the purchasing-intent of the person searching for a product.
The first version of the algorithm they built was tested on Marc’s webshop and it worked like a charm! And there was still so much that could be improved! The second version of the algorithm was offered as a service to befriended webshop owners that were equally enthusiastic. Typically sales levels skyrocketed by 50% or more and costs were reduced by 20% or more. What was there not to like?
After 3 years of continuously improving the system the team realised in the course of 2020 that the needs of webshops started to change. Amazon, bol.com, kaufland.de and other marketplaces developed into specialised product search engines unfolding interesting new advertising alternatives. Subsequently the challenge for webshops started shifting from ‘optimising the performance of Google campaigns’ towards ‘optimising the performance across a portfolio of search advertising opportunities’.
They also realised that webshops have access to a vast amount of information that is typically underused. Information about their own products, about competition, about customer behaviour, sales channels and so on. Making use of this information it should fundamentally be possible to build a system that could answer to the emerging challenge of ‘cross-marketplace advertising’. The task would be hugely complex though and the algorithm would need to be re-developed from scratch to incorporate the latest machine learning techniques.
After another year of hard work, Neural Search Labs GmbH was founded in April 2021 in Düsseldorf, Germany to mark the launch of NEURALsearch. NEURALsearch is a powerful Data-Science-as-a-Service that helps webshops unleashing the huge potential of Cross-Marketplace Advertising.