Search has become a fiercely contested space in the US at present and likely to be followed in other parts of the world, owing to the spread of language AI engines such as chatGPT among the masses. Search, as a business, has a special significance for Google as it funds R&D, so critical to any tech business. Quite possibly, its developing battle with Microsoft will have consequences that can have greater impact than we can viusualize now, irrespective of how a language AI works with search engines.
The numbers are striking.
Google’s 2022 (year ending December 31) was $ 283 billion with a net income of $60 billion. Search & other services generated revenues of $162.4 billion (AdWords), Google Network Members $32.8 billion (AdSense and AdMob), Google Other $29 billion Google Cloud $26.9 billion. Its R&D spend was $39.5 billion. What is at stake then, for Google, is its bets on the future, subsidized by the cash generated from search. Any reduction in search revenues will cause an unwelcome dent in R&D and compromise its future.
Microsoft, on the other hand, does not depend on Bing to produce the cash to fund its R&D. Its revenues for the fiscal year 2022 was $198 billion, experiencing growth in all its revenue segments – Intelligent Cloud, Productivity and Business Process revenue driven by Office 365 Commercial and LinkedIn and More Personal Computing revenue driven by search and news advertising and Windows – with a net income of $72.7 billion (https://www.microsoft.com/investor/reports/ar22/index.html). Its broad-based revenue structure is inherently risk mitigating.
Chat and Search
Now we can understand what is at stake for Google if its search engine is under threat but let us analyse how serious and immediate is such a risk. OpenAI’s CEO Sam Altman has announced they will charge $50 for a month, which works out to $600 for a year, assuming users avail of the services for a year. If one million users sign up, it will earn $600 million in revenues. Let us assume it reaches 10 million users which will take the revenues to $6 billion. I am not estimating net income as I have no clue to its costs. But just comparing gross revenue is revealing. Google’s quarterly revenue from just search and you tube ads is close to $ 50 billion. Hence, even assuming chatGPT does pose a threat to Google’s search, which has a market share of a little over 84% in the US (https://www.statista.com/statistics/216573/worldwide-market-share-of-search-engines/), it is not immediate.
But, someone can counter by saying that Microsoft plans to use chatGPT’s AI engine to dramatically improve results from Bing, which has a nearly 9% market share. Microsoft is obviously keen to increase its market share and eat into Google’s revenues, unless the overall market sees a surge that can accommodate both. Satya Nadella, CEO Microsoft, has already announced their intentions when speaking about the future of search and Bing. And now that we read about Google’s Bard having ‘failed’ and Alphabet’s share price dropping, Google is in for a battle.
Crunching the numbers
It is not surprising that chatGPT has offered users a price of $50 a month in the US because Google dominates the US search market but encounters different alternatives in different countries. To quote from Statista quoted above: “As of the third quarter of 2022, more than 62 percent of internet users in Russia used Yandex, whereas Google users were nearly 36 percent. In China, Baidu was the most used search engine as of December 2021, with nearly 86 percent of internet users in the country accessing it. In Japan and Mexico, people tend to use Yahoo along with Google. In the first quarter of 2022 nearly 56 percent of the respondents in Japan said that they had used Yahoo in the past four weeks. In the same year, over 27 percent of users in Mexico said they used Yahoo. Bing operated by Microsoft, was the second most popular search engine in the UK after Google”.
Baidu’s market share in search (in China) is 58% in desktop followed by Bing (23%), Sogou (9%), Google (7%); and a 94% market share in mobile, aggregating to 84% of all platforms (https://seekingalpha.com/article/4560177-baidu-my-number-one-china-stock-to-buy-now). And it has been developing other revenue streams such as cloud (where it is the fourth in China) and driverless AI systems, just similar to Google’s own forays. Media reports say that Baidu too is venturing in a similar direction. Clearly, the battle is going to be intense and each of the three giants will put all their might into this battle.
Lead kindly search – targeted advertising
It is too early to say how a language AI will work with a search engine but we must understand how search generates revenue for Google, which makes money not through search but through advertisements to which users are ‘directed’ by the search engine – although different people have different ways of expressing. Let me refer to a concrete example which has figured just a few days ago in a story by The Guardian titled ‘Google targets low-income US women with ads for anti-abortion pregnancy centers, study shows’ (https://www.theguardian.com/world/2023/feb/07/google-targets-low-income-women-anti-abortion-pregnancy-center-study – mandatory reading). The story has an arresting start: “Low-income women in some cities are more likely than their wealthier counterparts to be targeted by Google ads promoting anti-abortion crisis pregnancy centers when they search for abortion care, researchers at the Tech Transparency Project have found. The research builds on previous findings detailing how Google directs users searching for abortion services to so-called crisis centers – organizations that have been known to pose as abortion clinics in an attempt to steer women away from accessing abortion care”.
Different states have banned abortion after different number of weeks of pregnancy. Katie Paul, the director of the Tech Transparency Project says that “By pointing low-income women to [crisis pregnancy centers] more frequently than higher-income women in states with restrictive laws, Google may delay these women from finding an actual abortion clinic to get a legal and safe abortion”, and adds that while Google users can selectively target groups, the groups themselves aren’t aware that they have been targeted. Incidentally, it is ironic that we speak of ethical AI when there is complete absence of ethics here in the case we are discussing. I am reminded of a famous book of yesteryears – Hidden Persuaders by Vance Packard.
Summing up
This example is illustrative of how Google’s search engine works. It is unlikely if it will dramatically change except perhaps in such sensitive issues. Meanwhile, even after two decades of extensive use by millions of people, Google does often return inappropriate and even absurd results. Hence, I will suspend any view on how language AI will work with a search engine, since the architecture underlying each is probably different. The only thing we can be certain of is that intense battles are ahead.
Takeaways
Search has a special importance for Google for its R&D
Microsoft’s broad-based revenue structure is risk-mitigating
Baidu is foraying into similar areas
The US is the initial place of contest
Unclear how Language AI and search will work together
