Creating a Better Search Experience with Dynamic and Personalized Sponsored Ads

Millions of customers love the convenience of online grocery shopping, and its popularity continues to grow. But despite the convenience it allows, digital device screens have limitations that can pose challenges to the user experience.

Grocery stores such as Kroger are committed to seamless online shopping – along with the promise of fresh products and incredible value. The role of data science teams in that promise is to deploy technology that enhances the customer experience and helps them fill their baskets faster. To achieve those goals, 84.51° developed an innovative machine learning solution known as Dynamic Positioner that helps increase relevancy and engagement for shoppers.

The Problem: Screen Constraints and Inefficient Browsing

Every e-commerce experience is constricted by digital device screens, which impact the shopper experience. Only so many products can appear on a screen at the same time – and scrolling through a long digital aisle is a waste of shoppers’ time.

When a customer searches for products using the Kroger app or website, the most relevant items appear alongside featured ads. However, the relevancy of the sponsored ads to the search terms varies. This dynamic presented an opportunity to better engage customers with sponsored ads that were relevant to the context of their search – personalized and timely. The challenge was to develop a model to predict customer propensity to click on a sponsored ad and dynamically serve up the most applicable featured product in real time.

The Solution: Dynamic Positioner

The Kroger Onsite Ad Technology team engaged 84.51° Data Science and Research Labs teams to create Dynamic Positioner, a solution with a proprietary two-phased approach.

In the offline first phase, a model uses a combination of historical insights from website interactions to estimate how likely an ad is to be clicked. To ensure the model can handle uncommon search terms, historical searches are clustered based on similar clickthrough insights for the most frequently clicked positions in the search results. This approach helps to make the model more robust and improve its performance in handling a wider range of search queries.

The online phase incorporates real-time optimization and searches the set of positions of the sponsored ads returned from the auction. It uses the modeled relevancy and position effects, and the clustering from the offline phase, to search for the optimal position of the sponsored ads, all in under 20 milliseconds so that customers experience no latency effect.

The Results: Improved Customer Experience and Engagement

These two phases worked together to enable Dynamic Positioner to improve the search experience on the Kroger app and website by presenting ads to customers that were more likely to be relevant and useful. The results broadly indicated that the search experience was not only improved, but customers were more willing to explore and engage with the sponsored ads.

Meaningful Customer Experiences

Dynamic Positioner is an example of the power and importance of leveraging innovative technology to enhance the customer experience. It’s important to remember that it’s not always necessary to use expansive and complicated artificial intelligence to engage customers; oftentimes, the solution can be elegant and lightweight. In the world of data-driven machine learning, sometimes the best approaches are derived from sound probabilistic theory and use heuristics designed from domain expertise.