EasyJet tool lets people use photos to search for flights
European airline easyJet has built a visual search tool that lets people book flights based on just a photo, without them even needing to know the destination.
The “Look&Book” feature launches on the airline’s mobile app today and uses image-recognition technology to match a photo of a location from anywhere in Europe with one of easyJet’s holiday destinations. It not only recognizes the destination, but tells the prospective traveler where the nearest airport is and pre-populates the booking form in the app with those details. As easyJet sells more flights to places outside of Europe, the visual search tool will recognize those destinations.
Matches are most likely to be successful when pictures from Instagram are used, which easyJet purposefully intended due to those images being easier to process than pictures from other social networks or personal collections. Instagram photos are more likely to be geo-tagged, said Daniel Young, easyJet’s head of digital experience, who led the project. Aside from the location tag, each airport has been weighted with a number of tags such as beach, building, architecture and culture. If the location is not present, the tool will identify the key tags in any pictures and match that to the most accurate airport matching those tag percentages.
The “Look&Book” feature is relevant because Instagram is being increasingly used as a search engine. Young wanted a way to sell those people using the social network to search for holiday destinations by looking at specific hashtags, locations or areas. According to him, there’s no real ad product in Instagram to be able to capture this change in behavior currently, so the airline built one with agency VCCP and app developer Travelport Digital.
“When we were looking at how people use Instagram, it quickly became apparent that as people look through Instagram at nice photos, there was no way of turning that into a transaction,” said Young. “We’ve asked Facebook and Snapchat over the years whether it’s possible for them to create a feature that allows us to drive bookings from their social networks, but buying a holiday is a lot more complicated than buying a T-shirt, for example. There’s a big disconnect between searching for a holiday on Instagram and then being able to book it.”
Pulling the user back into the easyJet ecosystem prior to purchase means the airline benefits not only from Instagram’s propensity to drive discovery, but also search, selection and purchasing data inside their own app, said Gracie Page, innovation lead at VMLY&R. “Although there’s been plenty buzz in the visual search space this year already, easyJet has gotten ahead of the curve by connecting the dots back into their own ecosystem,” said Page. “As consumer devices become increasingly more sophisticated and able to handle complex computational processes, we’re just getting started with the delight that will be the visual search paradigm.”
A campaign launches this week to encourage people to upload their Instagram photos to the app. As easyJet’s visual search tool learns to recognize more images, Young said the brand would encourage people to share images from other social networks or from their own collections.
“The holy grail for us is a customer waiting for a train and they spot a cool holiday location on a billboard. They don’t know where that place is and so they take a photo of it and upload it to our app, which tells them where it is,” said Young.
Image courtesy of easyJet.
More in Media
Media Briefing: Publishers say a cookie-less future remains murky, overheard at the Digiday Publishing Summit
Publishers who attended the Digiday Publishing Summit opened up about their top challenges, concerns and curiosities during closed-door, anonymous town hall meetings.
Why Warner Bros. Discovery is leaning harder into YouTube and Threads to monetize and engage social audiences
WBD is seeing ad revenue growth from its YouTube channels and engagement on Threads surpass performance on X.
Powering all of Amazon’s hardware is an updated large language model, that could help connected devices actually be smarter.