We created an Azure based platform scaling up and down automatically based on real-time traffic, active campaigns and size of Publisher network, keeping running costs at a minimum.
More importantly we created our own custom clustering and Natural Language Processing (NLP) algorithms that would categorise traffic based on reading preferences, going down to analysing the topic of individual articles.
Our client has already shifted its entire business model to revolve around this newly built platform and we are already considering other wacky ideas to increase conversion rates, such as applying sentiment analysis algorithms and targeting customers based on the sentiment of the articles they’re reading.