Machine Learning might seem incredibly complicated for non-technical types, although the latest application of the tech is very easy for even the most technophobic web users to get their heads around: TripAdvisor is using ML to launch a new photo feature.

Essentially TripAdvisor reviewers will be able to auto-select from an array of primary photos for whatever accommodation they happen to be reviewing or reading about.

Ideally, this will mean that, in the future, TripAdvisor pics will be much smarter, non-blurred and more appealing in general. As opposed to the rather awful low-quality tat that litters the site right now.

Particularly when an owner doesn’t manually choose a photo to accompany their listing, users will often see a completely random traveller’s pic of a toilet door or something else that’s equally unhelpful and irrelevant!

The latest machine learning initiative will hopefully mean that, in future, all pics of accommodations and restaurants listed will be far more representative photos of the business in question.

TripAdvisor’s new auto-select tool analyses a number of factors that are known to enhance user engagement, including image resolution, orientation and sharpness, having people in the photo and more.

“We know that good photos are a key part in attracting new guests,” said Martin Verdon-Roe, VP B2B Hotel Product & Marketing TripAdvisor.

“According to research, 79{a990e605127f06bac58d8f530ec8d3ddc1721ced564bd12be3752b381e1e9f7f} of TripAdvisor travellers said that photos were important when choosing to book an accommodation and in more recent live site testing with partners, we can see that optimised photo selection has been shown to increase user click-through rates. This exciting new feature capitalizes on our strength in machine learning with insights from millions of professional and traveller photos across TripAdvisor to help properties make the best first impression to our global travellers.”


Via Travel Daily News

Newsletter sign up

This field is for validation purposes and should be left unchanged.