ChatGPT, Google and the war for the search bar
Some are viewing ChatGPT and other recent advances as a threat to Google Search. The argument draws energy from the fact that ChatGPT is (or shows promise of being) a superior mode of information retrieval for some set of queries.
ChatGPT does amazing things. I’ve seen examples of it prototyping code, playing chess, crafting jokes etc… But the above points to a kind of talent that is particularly relevant to the future of search–namely, its ability to effectively answer questions.
The reality though is that unseating Google as the dominant player in global search requires more than just a superior product experience. Google has spent the last few decades brilliantly building up a strategic position that would make it extraordinarily difficult for anyone to supplant them.
Google’s strategic strength, which works to assure their >90% share of global search, centers around their control over the search bar. That control is secured at multiple layers of the stack. On mobile, Android (70% of mobile OS) assures and protects Chrome’s dominance of mobile browsers. On desktop, Google lacks protection at the OS level, but its ownership of the search bar is well guarded by Chrome’s vast penetration (66%). Even when Google doesn’t own the search bar, they pay for the right to lease it, securing annual contracts with Mozilla and Apple respectively.
For Google to lose their search dominance, Chrome’s market share would have to decline significantly *and* the other browsers (perhaps even new browsers) would have to opt for a different search provider. Google is well defended against the first scenario because of Android’s dominance in mobile OS. And Google is defended against the second by virtue of large pockets. It’s hard to imagine anyone else other than Microsoft paying Apple $15B or Mozilla, the riper target, $450 Million.
The bigger problem for Google is that, despite their distribution advantages, a tension may emerge between user demands for a better search experience and economic interests that desperately want to keep things as they are.
What is that *better* search experience? Paul Buchheit suggests that LLMs can out-compete traditional search experiences by bringing answers to users without the extra baggage of a search results page. He thinks that the search bar might just show autocompleted answers as opposed to directing users to a list of search results and that the problem for Google is that this would mean eliminating real estate for search ads–an economically disastrous possibility.
If Buchheit is right, then Google will likely have to rearchitect their search experience. That would potentially cause economic injury but it likely wouldn’t affect their market share unless Chrome were to decline significantly or–and this is where things get loopy (in every sense of the word)–if a loss in economic power were to lead to stingier pockets when it comes to owning the search bar in Firefox or Safari.
On the surface, wargaming the future of search seems like a question about search engines. But the underlying conflict is really a matter of browser politics and ownership of the search bar.