Alphabet’s Google franchise sits at the centre of the internet. In effect it is a contributory database as it tracks billions of users ‘journeys’ and records precisely who clicked on what. This gives it an enormous advantage in serving relevant search content.
As such, it is a natural repository of information from third parties (e.g. weather reports and sports scores), generated by Google employees (Streetview, business opening times), and crowdsourced from individuals and companies (restaurant ratings, hotel inventory).
The value of connection
This data asset is so powerful that consumers continue to get the best results from Googling, despite the more engaging user experience of a large language model (LLM) chatbot. For example, clickstream data shows that ChatGPT usage has remarkably little effect on consumer Googling behaviour. Meanwhile, OpenAI pays a third party for grey market access to Google’s search index, despite having deeply subsidised access to Microsoft’s Bing index. The Department of Justice estimated that Microsoft have spent $100bn over 20 years to build this index, yet it still falls badly short of the Alphabet index.
In short, innovative technology alone cannot displace the value created by being at the nexus of a vast number of users. We prize these network effect businesses as classically defined in the 1908 AT&T annual report:
‘A telephone – without a connection at the other end of the line – is not even a toy or a scientific instrument. It is one of the most useless things in the world. Its value depends on the connection with the other telephone – and increases with the number of connections… the Bell system… has assimilated itself into and in fact become the nervous system of the business and social organization of the country.’
Indiscriminate selling
LSEG and S&P were hit hard by the fall-out of Factset’s fiscal 2025 results which came out September 18, despite being dissimilar. Factset derives 83% of its revenue from its desktop research business, at the core of which is the resale of third-party data from exchanges and elsewhere. This is a workflow solution used by investment and wealth managers and investment bankers. It is at a remove from revenue and can be switched or multi-sourced without penalty (we switched off Factset ourselves two years ago). The equivalent businesses at LSEG and S&P are 11% and 8% of their toplines respectively.
Instead, most of their revenue and even more of their profit comes from classic network effect businesses. Equity indices, credit ratings, and price reporting agencies (PRAs) are all quintessential benchmarks which large, fast-paced markets have agreed to use a single reference point – in effect, ‘the nervous system’ of their market.
Client revenues are tightly bound to the product and at risk if it is switched. Bond funds commit to buying only credits rated by S&P and Moody’s, ETFs are marketed based on their named equity index exposure, and energy derivative contracts cite specific PRAs as their basis. In total, benchmark revenues are 11% and 62% of LSEG and S&P revenues respectively.
Other businesses within the umbrella of LSEG and S&P also benefit from different network effects. Some of these are classic contributory databases, like Carfax Mobility for used car history at S&P. Others are based on indirect network effects, like LSEG’s real-time pricing franchise which rests on its aggregation of many low latency data feeds with built in fee liability calculations and a vibrant ecosystem of third-party developers.
Finally, there are the classic trading venue businesses within LSEG, such as Tradeweb, the leading electronic venue for cash rates and credit trading, and LCH Swapclear, the world’s dominant interest rate swap clearing venue. In total these other network businesses are 57% and 12% of LSEG and S&P revenues respectively. They share the typical network effect characteristics of high margins (as most of their data is free), high capital efficiency, and low competitive intensity, usually working in a monopoly or a duopoly.
Long-term lens
These businesses in total comprise 68% and 74% of LSEG and S&P respectively. Much of the residual businesses of both are high-quality (though not network effect driven) but are inessential to our investment theses for them.
Over an investment period of many years, their shareholder returns will be dominated by the internal cashflow growth of their core network effect franchises. The competitive differentiation of these franchises and their economic centrality make them positively geared plays on global GDP growth. As the world grows, we expect it to drive outsized gains for the people holding the ring on oil, bonds, interest rate swaps, equity indices, and even the humble used car.
Over time LLM businesses will come to co-exist with these networks which have in truth adapted themselves to many technological shifts over time – as Google’s Search franchise is already showing signs of doing. Markets used to be physical locations where traders congregated; for example, the original S&P ratings were distributed on massive leatherbound books. Any time when there is an indiscriminate and panicky selling of ‘buckets’ or ‘themes’ is bound to create opportunities.
