Tripadvisor already seeing “compelling results” from generative AI investments

Tripadvisor, the self-proclaimed world’s largest travel platform, which helps 463 million visitors each month plan their trips and experiences, has said that its early generative AI investments are paying off. CEO Matt Goldberg said during the company’s Q3 2023 earnings call that the platform’s recently announced travel itinerary generator is leading to higher visitor engagement and more revenue. 

Given the hype around generative AI and the technology industry’s insistence that it is the future of all things work, digital and experience, it’s worth assessing to what extent companies are already making use of it and what impact it is already having. It’s also worth reiterating that twelve months ago hardly anyone was talking about generative AI at all, and although growth in testing and usage has been rapid, it’s still incredibly early days and lots of organizations are unsure about how best to make the most of the technology’s possibilities. 

Tripadvisor announced its AI-powered travel itinerary generator back in July, a feature which creates personalized travel itineraries using OpenAI’s generative AI technology. It takes insights from over a billion reviews and opinions of the more than eight million businesses listed on Tripadvisor and delivers users with guidance from ‘real travelers’. 

To create an itinerary using the tool, travelers are prompted to enter their destination and travel dates, who they’re traveling with, and the types of activities they’re interested in. In just a few seconds, they’re presented with a “personalized, day-by-day itinerary, which they can save, edit, and share with travel companions”.

As travelers progress through the trip planning process, they have opportunities to engage and interact with Tripadvisor’s travel guidance content, from reviews and opinions, to experiences and photos from other travelers. Users can save all content they interact with to a ‘Trip’ as they build their travel plans.

It’s exactly the kind of use case that proponents of generative AI talk about as being ‘revolutionary’- taking an existing, unique dataset and creating AI generated content that’s personalized to users within seconds. Instead of travelers having to trawl through different recommendations and other user reviews, they can request a guide that’s useful for them. 

But how is it performing so far? According to CEO Goldberg, the results seem promising. He said: 

We're still early in the journey and are iterating continuously, but we're seeing compelling results. During the first three months in beta, members who created itineraries returned at a much higher rate within the first seven days than members who hadn't created an itinerary and we've observed that higher return rate continuing for the full month thereafter as well. 

Importantly, members who build an itinerary also generate on average three times higher revenue than the average Tripadvisor member and the average member already monetizes at approximately 10 times the rate of non-members. 

We believe the initial data is a good indication of early product market fit and we're on a plan to launch this full trip planning and itinerary product on our mobile app, on iOS and Android this quarter.

Engagement and monetization

Goldberg said that Tripadvisor is focused on building engagement and delivering “world class guidance products” to create a diverse set of monetization paths. In Q3 2023, the company saw revenue grow year-over-year by 16% to reach $533 million and net income came in at $27 million. Goldberg said: 

I'm pleased with what our teams have accomplished in the quarter. We're leaning into the strength of our existing assets while ensuring that we build and innovate for the future. 

We're executing more efficiently and effectively across each of our segments, and we continue to solidify our position as the most trusted source for travel and experiences with Tripadvisor leading the way on the launch of the Coalition of Trusted Reviews, where we've brought together the leading platforms in the online review space to combat fake reviews and set global standards to best serve consumers.

Goldberg added that key to Tripadvisor’s future success is driving deeper engagement with travelers, making use of the platform’s “unique content and data at scale”. He shared two examples and said: 

First, we've been leveraging generative AI and machine learning tools to bring together guidance from our community and our editorial team on a set of our most popular destination pages, think New York City, Paris, and so on. Curating fresh, relevant content on these pages is driving over 15% more saves so travelers are engaging more frequently by adding a desired hotel, restaurant, or experience they discovered through our product to the trip they're planning. 

Second, we recently rolled out a test across tens of thousands of hotels that uses Gen AI to summarize reviews and deliver clear insights to travelers on key quality attributes as they consider their choices.

Let me give an example, for me the single most important factor in booking a hotel is that I can get a quiet night's sleep. We're now leveraging our data to provide a clear signal about the noise level of a particular hotel, along with other critical qualities we know are important to travelers like atmosphere, service, value, and more. 

And Tripadvisor is grounding its generated content in the opinions of other travelers, which it sees as key to trust. Goldberg added: 

While generative AI provides the perfect way to summarize content at scale, trust is still at the center of why people come to us, so in this feature we share the specific underlying reviews from our community of travelers that were used to generate each insight and provide a direct path to go deeper on what matters most to the traveler. 

But Goldberg recognizes that these features are only useful if they impact the top line. He said: 

These examples are just a few of the many encouraging proof points that our teams are driving, which gives us conviction in our strategic direction, prioritization, and sequencing. But we know this only matters if it translates the financial impact over time, and we're also pleased with the leading indicators we're getting on monetization. 

I mentioned before the significantly higher revenue we see among users who create a trip. As these users build an itinerary and engage more deeply they're exposed to monetizable hotel and experience recommendations to relevant ads from our partners, all of which were increasingly integrating directly into this user journey. 

And the refreshed content on our destination pages is a perfect point in the journey to surface these options to a traveler. We’ve already observed double-digit increases in experiences revenue on the pages with these content enhancements and we are just beginning to scale. 

All this adds up to more engaged travelers exposed to monetizable products. 

My take

A promising start for Tripadvisor, but the true test will be longevity and how customers engage with these tools over time. The travelers that trust the generative AI tools’ recommendations and have good experiences off the back of them - that will be the proof in the pudding for Tripadvisor. If travelers end up having poorer experiences relative to curating their own experiences through research (as is often the case at present), then that will mean bad news for the generative AI experiences. As noted above, it’s early days and we need to see how this all plays out.