Position: Product Manager, Search and Ranking
Led the Search, Ranking and Relevance Platform team, rebuilding the core discovery personalization architecture from the ground up.
Targeted Search and Autocomplete
Some of the major challenges Groupon faced when I joined the team was the difficulty to accurately find the items you were looking for. Because of the company's age and number of teams that work on the search feature, over time the search feature has become bloated, making it challenging to find the right items in the catalog.
I worked with the engineering team to improve the autocomplete and search experience by reordering and targeting the right items that should display when search, replacing the old approach with a new NLP query matching process. I also worked with the design and marketing team that updated the filters and search display options to better show users why certain search results were displayed.
This proved to be a success at increasing CVR by 18% and OGP by 6%.
Personalized Ranking
As more and more users came onto to the platform, Groupon was lacking in the identifying and choosing the right features to rank on in order to display the best products for that customer. After an investigation into the types of available data we could use, I prioritized which models we would proceed with testing to increase personalization for users.
The final model used in the personalized carousels on Groupon before I left helped increase OGP by 9%.