Since December 2018, I have been working on transforming GetYourGuide’s user-facing search and discovery (S&D) experience. This initiative was launched in response to persistent customer feedback that it was difficult to understand how one activity was different from another in the GetYourGuide catalog. After analyzing GetYourGuide’s inventory, it became clear that not only were there many duplicate activities on the platform (e.g. dozens of 2-hour tours of the Louvre offered by different suppliers), but suppliers were also describing the attributes of their tours in different ways (e.g. “skip-the-line” vs “priority access” vs “V.I.P. entry”). This duplication of popular activities and lack of terminological consistency made it impossible for users to differentiate between activities resulting in a very poor user experience.

As a result of these customer problems, GetYourGuide elected to undertake a process very similar to Amazon’s decision to move from seller-centric shops to today's multi-vendor product detail pages. We elected to eliminate the old “supplier-centric” inventory model and create a new “product-centric” marketplace structure by (1) consolidating duplicate inventory and aggregating their availability into a single “virtual product” (VP), and (2) meaningfully differentiating these VPs for users through manual curation. 

My role was first to design a user research scenario that would provide us with user insights about how to organize our existing options into these aggregated VAs. This research about our customers’ mental models of activity structure allowed me to move forward with confidence and start manually curating our existing inventory. Working with product managers and data analysts, I then selected 8 high value points of interest (POIs) that were (1) complex enough to “stress test” our curation model, (2) strategically relevant to the company’s business interests with enough duplicate inventory and fragmented availability for curation to provide value, and (3) received enough traffic to reach statistical significance in an A/B experiment over a 2-week testing time frame. I then proceeded to differentiate activities according to 16 different attribute types and began creating a curation model and set of operational guidelines for the newly formed curation team to iterate on and optimize in the future. Once this curation sprint was complete, I started merchandising these newly created VAs by creating new standardized titles in accordance with the existing GetYourGuide style guide, UX best practices, and the curation attribute model. These new titles were written using clear, consistent terminology and defined what the VA is, while also clearly differentiating that VA from all of the other activities available at the POI in question. 

Starting with a set of 989 supplier-defined activities, this curation process resulted in 404 unique VAs being created. This represents a 59% reduction/simplification of GetYourGuide inventory across the 8 POIs that were curated. We then ran an A/B experiment of these new curated VAs against the status quo supplier-defined activities. The success metrics of this experiment were twofold: (1) to increase click-through-rate (CTR), and (2) to reduce instances where users encountered unavailability. We saw an ~10% increase in CTR for VAs with standardized activity titles, while also seeing a ~30% decrease in unavailability encounters as a result of the availability aggregation provided by the VA model.
 

This transformation of our inventory is still underway. It is not finished in any conventional sense of the word. It remains an ongoing operational task for multiple teams at GetYourGuide. Rather this project provided a viable proof of concept for the company’s executive leadership. As a result of this test, we have made S&D our number one company priority, the catalog engineering team has received increased resourcing and now has a dedicated Senior PM to preside over attribute standardization and data quality. We have reassigned our supply excellence operations team to preside entirely over existing inventory curation with a team of 8 currently undertaking the task and growing in 2021. We have also hired a merchandising team who have taken over writing new activity titles for these curated/aggregated virtual activities according to the best practice title playbook I put together. This project has become the largest product reorganization GYG has ever undertaken and our proof of concept provided the basis for the company’s product strategy moving forwards.