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.
About the author…