Brands, marketers, and product designers need to understand their customers. Traditionally, market research was driven by surveys and focus groups of limited scale. Today, digital signals like web browsing behavior can provide a stream of observed behavioral data that is rich with information about a user’s interests, needs, and preferences. This type of data, coupled with machine learning techniques, holds the promise of freeing market research from the constraints of self-reported customer data.
In its raw state, web browsing data is both too detailed and too sparse to be comprehensible, let alone actionable. Melinda Han Williams explores semantic embeddings as a novel approach for understanding observed digital consumer behavior and details how to use a semantic embedding of web browsing behavior to drive unsupervised clustering for customer segmentation. You’ll learn how Dstillery has trained a neural network on 15 billion behavioral interactions. The resulting model can be seen as a much lower dimensional embedding of the internet and, if projected into two or three dimensions, as an interactive map. This taxonomy of internet behavior can be used as the foundation for a number of applications, providing unparalleled insights into consumer behavior and needs.
Melinda Han Williams is the chief data scientist at Dstillery. Before joining the ad tech industry, Melinda worked as a physicist developing third generation photovoltaics and studying electronic transport in nanostructured graphene devices. Her peer-reviewed journal publications have been cited over 8,000 times. Melinda holds bachelor’s degrees in applied math and engineering physics from the University of California at Berkeley and a PhD in applied physics with distinction from Columbia University, where she held a National Science Foundation Graduate Research Fellowship.
Comments on this page are now closed.
©2019, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • email@example.com