In the time it takes to read this sentence, Google fielded half a million queries. At over 100K searches per second, Google has become the extended brain that former CEO Eric Schmidt envisioned. Meanwhile, The Information reports that OpenAI wants a piece of that action and is planning on entering web search. That action, as of Q4 '23 raked in a staggering $11,000 USD per second or $85.5B dollars in revenue for Google alone. 

There’s a lot to unpack. I'll discuss two big questions. First, where is Google vulnerable and what can we learn from prior market failures? Second, how will OpenAI’s success change the web ecosystem as we know it today?

‍Can Google Adapt

Google has been threatened many times since it took over the top spot in web search 24 years ago. 

With every major shift in technology, Google needs to adapt to be the quickest way to ask our question. And we’ve seen how resilient the platform is to attacks. 

Thus far, they’ve adapted to three major web search changes. 

First two were the rise of the mobile web which was quickly followed coupled with the rise of mobile apps as the App Store released in summer of 2008. Google’s simple and fast search-engine seemed to be ready for the mobile web and was well-positioned for that transition. Former Google VP, Marisa Mayer’s insight that “slow and steady doesn’t win on the web” was already the standard operating procedure. 

Second, mobile platforms and operating systems threatened to fragment the market. Google was threatened by the iOS ecosystem. A business strategy kept them on top through an expensive partnership with Apple and Safari to keep them on top. Google reportedly pays $18B a year to Apple in revenue shares from searches originating on their devices. Again, Google promotes Chrome heavily to reduce this “tax” on iPhones. Moreover, Google’s acquisition of Android was the second strategic win which has wildly paid off. The open OS has strategically secured its dominance on mobile devices and most Android devices. Many believe that the Android acquisition is the best acquisition Google’s made and arguably in all of modern business history.

The third attack has been the rise of voice search on mobile devices and new screen-less audio-first experience. Today, according to a Google Mobile Voice survey, 41% of American adults and over half of teens in the country use voice search every day. Again, Google deploys two approaches to maintain their default place fielding searches and releasing its own competitive devices like Google Home and Google Assistant (sunset now for Gemini).

‍AI: The Next Front

The new attack is from LLMs, Generative AI, or let’s just say “AI” to simplify the discussion. To understand how this attack impacts Google we need to step back and segment the Intent of the searches. 

There are two ways that “AI” can attack Google. First, they can remove the need to search when asking specific questions. “AI” today are quite good at providing answers around complex topics, evergreen questions, or getting “how things work” explanations. Moreover, as “AI” is used more for planning, product procurement/selection, or finding local destinations, these use cases could reduce those lucrative local and product searches. It’s important to realize that according to BrightEdge, over 68% of online experiences begin with a search engine so if people “start with AI” instead of search and, even worse, if AI can search the web -- it’s possible AI may be the default approach.

Back of the envelope, Google owns over 82% of the search market so let’s assume that nearly 56% of all online experiences start with Google. That would mean that each point drop in online experience that starts with search is taking 1/2% or $500M off the table from Google per quarter. This is an incredibly naive analysis because Google could turn dials on search advertising to increase revenue and, effectively, inflate prices as they set the target -- nevertheless, this would eventually fail to hide stagnation.

Given the reality that Google’s value requires aggressive revenue growth, shrinking their accessible market could be catastrophic for them in the public markets. 

Understandably, Google isn’t sitting back. They have launched Gemini which is a page out of their Android playbook to protect their assets by fragmenting the market and limiting the scale that OpenAI, et al., can reach. Moreover, Gemini will protect their productivity tools from attacks from AI-first productivity tooling.

Is OpenAI an existential threat? Due to the optics we discussed earlier, OpenAI and “AI” just needs to slow Google’s growth to have catastrophic knock-on effects on the company and market. 

‍OpenAI: The Death of the Open Web

Aside: If you’ve been in the industry long enough, you’ve seen the headline will X kill the web and I had to go there again. Every time we examine this question with a new “X” we come to a realization that things change, are different, but the web has stayed resilient.

Given how “AI” experiences we’ve seen work we can assume that it’ll take more eyes away from web pages as the user’s curiosity is satisfied by the generated summary or synthesis. 

‍Who does this harm? All of us. Especially people want to introduce something new or try to change the narrative.

At the moment, the open-web and, to some extent, App stores allow some assemblage of a ‘near-democratic’ distribution that allows for the rise of new ideas. As people trust “AI” to provide them answers and go there first, they’ll also sacrifice discovering things through search that are new. “AI” recommendations will, likely, bias to what results have more data and social proof. This bias harms everything new unless a user specifically asks for “new” or “with no reviews”: new restaurants, new stores, new products, and new voices. All things “new” require time to build a reputation, and that time takes money.

The reality is that Google already has some of the symptoms of creating homogeneity in our diets because it’s a single algorithm; however, they do a lot of work to surface new and recent along-side other results.

The Impact of “AI” on the world can be looked at through this lens. Imagine that we all spoke to “One human” to understand the world. What would happen to us? When we want to know what happened in the “War in Pangea” we’d all see the same perspective that isn’t optimized based on our location or history of searches. I can’t see how “AI” will or should personalize responses to fact-based questions.

‍Cost of Marketing and Advertising May Rise

With drops in search-engine inventory and clicks from search to publishers, we’d see a supply of ad inventory drop which may lead to a rise in ad costs. This would limit new voices as well as they may be priced out. It’s very possible that OpenAI or new “AI” search may integrate paid results in some way in the interface; however, I can’t imagine they’d perform anywhere as well as they do within Google and Bing Search, today.

We’ve seen this impact already as users spend more time in apps vs surfing the web. We will continue in the trend where we squeeze our publishers' top line. That said, it’s possible that publishers will find a new revenue stream by publishing real-time feeds to “AI” platforms at a cost per article or, potentially, “cost-per-reference” which would incentivise good content.

‍Conclusions

The potential impact of OpenAI's entrance into search, alongside the rise of “AI” tools, should not be underestimated. Google's incredible adaptability has solidified its position, but this technological wave is undeniably different. “AI” offers unprecedented convenience but simultaneously threatens the fundamental openness and democratic discovery that underpins the web as we know it.

Google's counter-maneuvers will undoubtedly influence the evolution of AI-powered search and the entire web ecosystem. Whether they can truly blunt the growth of OpenAI and similar technologies remains to be seen. Their track record speaks to their resilience, but their core revenue model – reliant on search supremacy – is increasingly exposed.

A critical issue is the erosion of discoverability for new ideas, products, and voices. This may stifle innovation and further concentrate power in the hands of already dominant players. While “AI” can improve our access to information, its centralized nature risks homogenizing experiences and reinforcing existing biases.

As individuals and as a society, we must carefully consider this trade-off between immediate convenience and the long-term vitality of the web. The future of knowledge creation and distribution hangs in the balance. It's crucial to promote diversity in AI development and advocate for ethical guidelines that ensure these technologies work in the interest of the user, not just the provider. This new era of search demands mindful consumption and conscious choices as we navigate towards an uncertain, but undeniably transformative, future.

‍