From our user testing we learned photos are essential for travel inspiration. 73% of users wanted to see more photos in a more inspiring, visual experience. However, changing the UI to use larger photos had major implications for multiple teams.
- Issue 1: Volume. The platform already had 300 destinations, each with hundreds of POI images that had been originally curated for a UI that used smaller images in fewere places.
- Issue 2: Cropping. The new UI took one phote and re-purposed it in 11 different places, each with a slightly different aspect ratio. This occasionally led to unintentionally idiosyncratic, if not downright comical, cropping. We needed a solution that would scale.
The Product, Engineering and Content teams strategized how to address:
- How to handle photos for new partners/photos being uploaded from this point forward
- How to handle photos for existing partners/retrofitting existing photos.
Because a single photo would be used in a variety of contexts, we figured out what an optimal original photo size & aspect ratio should be. Then we created documentation for the Content and Account Management teams to share with the partners.
To retrofit the thousands of existing images, the Content Team combed through each image and either optimizing/re-cropping the images or finding new images as necessary. While most of this was a manual process, the Engineers helped to automate some of the work by building an image optimization & cropping tool in our internal Admin CMS.
It's one thing to say "let's including personalized recommendations of desintations too!" It's another to enact it. One of the biggest blockers to launching the Discovery page was the fact that our recommendation engine, UtripIA, was optimized for in-Destination inventory (e.g. POI, restaurants, tours...etc). It was not built for top-of-funnel destination recommendations.
We worked with our engineering and content teams to find a solution. While full incorporation into UtripIA was the ideal, a recommendation solution that utilized Elasticsearch would:
- Be faster to implement
- Be adaptable for other destination-adjacent content (e.g. cruises)
- Enable additional functionality, like special filters, that we had been previously unable to accommodate
- Our Data Scientist could focus on UtripIA optimization for other projects that were happening alongside the Destination Discovery project.
- New partnership possibilities with airlines and cruise lines
Because Utrip is a small startup, we were often iterating on the internal processes along size the external product builds, to ensure that every step forward meant we could do more, faster. One consistent spot of friction in the product development process was the turn-around time for creating design specs, a Google Doc, which could be up to 40 pages long and easily became out-of-date since all spec updates had to be manually managed.
A.B.I. Always be iterating. This mindset applied to everything from UI to internal processes. For this project we switched design software (from Adobe Creative Suite to Sketch), and the design documentation process (from generating 40+ page Google Documents to streamlined Zeplin files). Within a week, we were fully on-boarded. Within a month our output had doubled.
- 95% reduced turnaround time
- 291% increased document output
- 167% increased engineering satisfaction
