Supermarket data sweep
“A retail company is a player in three ways; as a publisher, as an advertiser and as a data provider,” according to Ali Dasdan, engineering director for data platforms and business intelligence at Tesco. His view of the marketing and advertising ecosystem within a retail context splits tools and platforms into three categories.
The demand side partners includes ad agencies, agency trading desks, ad servers, as well as demand side platforms for real-time bidding (RTB), for search and for social.
Markets or auctions sit in the middle, encompassing public and private RTB exchanges, search engines and social networks. At the other end are the supply side partners, including supply side platforms (SSP), ad servers, as well as site analytics partners.
Right n the midst of that are the data management platforms (DMP) that are helping brands to extend the audience and the data that they hold, as well as doing selection and segmentation.
According to Dasdan, one of the benefits of having a data strategy is that it would result in “definitely” having a unified 360-degree view of the customer, especially if it were enhanced with second and third-party data.
“It’s an unbelievably high-speed, low latency operation."
Dasdan went on to outline some ways in which machine learning can be applied in marketing and advertising in a retail context, one being assisting with traffic management.
He said: “It’s an unbelievably high-speed, low latency operation. You cannot keep investing and it costs a lot to operate multiple data centres across the world.” Alternatively, using machine learning would enable retailers to recognise which sites to pay more attention to and which ones to ignore.
“The benefit we saw was 10% to 20%, which is actually a big deal."
When Dasdan and his team applied machine learning to budget optimisation, the benefit was tangible. “The benefit we saw was 10% to 20%, which is actually a big deal because you are spending sometimes hundreds of millions of pounds,” he said.
Dasdas also stated that using rich customer data can improve the service that customers are provided with. “Customer service can be improved if you know more about your customers, as well as the fact that you have unified customer data . You can use that to speed up the application developmentp.”
The engineering director added that retailers could take advantage of all the data they have to hand by unifying the online and offline presence and keeping one source of truth on an internal data platform. Prudently, he advised paying attention to security, privacy and compliance because “marketing is one area where you need to take extra care.”
Ali Dasdan was speaking at AI Congress London.