Gemini, Google Travel and the threat to OTAs

This has to be the best mission statement in history. Not only has it survived to be just as relevant 25 years later, but by sticking to the plan, Google has built a $2 trillion company.

When ChatGPT launched on November 30, 2022, it immediately kicked off the debate about the threat to Google’s dominance in search. The only thing that has really threatened this near-monopoly in search was the rise of social networks. But this is bigger - could generative artificial intelligence disrupt the very core of Google’s business?

Google Shopping

It’s a simple concept. For certain search keywords, there is a clear intent - the user is searching for products, not websites. And where that intent is understood, the original 10 blue links of search results are not the optimal format. You can skip a step by displaying products directly in search results. Users don’t care about the home page of Macy’s or Best Buy. Those sites require a secondary search to find products. Bypassing these brand home pages is a faster way to get the user what they are looking for – so it’s better. This started with Google Shopping, versions of which have been around since 2002.

Google Travel

Google Travel largely sits on a similar platform. It probably seems obvious in hindsight, but it took a long time to apply to travel products. It started with Flights around 12 years ago, and since then has expanded to Hotels, Things To Do, Transportation and Vacation Rentals.

These searches for flights and hotels are often very distinct. The familiar fields of destination, dates, guests are on the home page of every online travel agency for a reason. Once you recognize that, it’s a better experience for the customer to bypass these OTA sites and display flights or hotels directly and allow the user to modify and filter searches within Google.

Generative AI

So how does this all fit together? There has undoubtedly been a high level of panic at Google in the past 15 months. I haven’t heard any tech leader say that they saw this coming. Before ChatGPT, previous versions of generative AI were clunky and didn’t appear threatening. That launch changed everything. The space is moving so quickly, led by OpenAI and Microsoft, that it’s easy to see how it could disrupt Google’s entire advertising model in a very short space of time.

Innovator's dilemma

It's the ultimate example of innovator's dilemma. For many user searches or customer objectives, these LLM (large language model) chatbots clearly give more useful responses than a Google search. The LLMs rightly or wrongly have trained on, and learned everything on the internet, so in theory they already have all the information and answers.

So by following their mission (which they will), Google needs to switch to generative AI responses for an increasing percentage of search results (Google OneBox is just an early version of this). Generative AI responses attempt to answer the question or search in full, sometimes with a chat back and forth and without the need to go off and explore other websites. But Google’s entire model is built around displaying organic results as well as ads, and the chatbot format doesn’t immediately lend itself very well to that. So how do you go head-to-head with ChatGPT or AI search websites like Perplexity?

Gemini

With Gemini, Google’s latest generative AI release, they are finally competing in quality with OpenAI’s ChatGPT. Combined with Google Travel, there might be a nice opening for Google to test new search result formats without forgoing revenue, overcoming the dilemma.

For top-of-funnel discovery searches, changes might be arriving soon. Most travel writers think their sites are special and unique. Most are just re-hashing the same old lists, recommendations and top tips. Google can probably ditch 99% of it. That content can all be replaced by the Gemini LLM response. The best 1%, which provide actual unique and valuable content, will be just fine, and Google will find a way to mix it in. That would be a better experience for users. Today.

When it gets closer to the transaction, the traffic is far more valuable for Google. Here’s where Google Travel comes in. Gemini is already integrated with Google Maps, Flights and Hotels. Searches with Gemini feed directly through Google Travel surfaces to end up on booking pages on OTAs, hotel and airline sites. Still missing on Gemini are the Things To Do products, but those could be rolled into the Gemini format in the same way Flights and Hotels are.

Gemini makes Travel and Maps more useful because it acts as a deep search, which wasn’t possible with just algorithms. And Google Travel makes Gemini better because users are often looking to purchase flights, hotels and activities. Each makes the other part stronger.

Other integrations in Gemini, like Gmail, Docs, Calendar and more, will no doubt add value as well. Many will think this is the “Customers search 35 sites. Travel is broken. Here’s my app” trip planner argument. It’s not. It’s just better discovery. What if users go from searching 35 websites to book a trip, to just five websites plus Gemini. We know many people enjoy the planning phase, and Gemini could replace those websites, not just because it has all the answers, but also because it also recreates the fun part of the trip discovery.

These changes might happen much faster than people expect. ChatGPT, especially with the integration of GPTs, is challenging Google daily. Google can no longer afford to roll out changes with a timeline of their choosing. Those changes could be drastic for many publishers, much more so than the previous algorithm updates. Optimizing content for LLM training is not well understood – it’s new to everyone. People in the travel industry should be following this closely.