Social Capital and the platforming of VC
Social Capital positions itself as a cross between the Gates Foundation (mission), Berkshire Hathaway (duration and outlook) and Renaissance Tech (tech and data) with the ambitious mission of “advancing humanity by solving the world’s hardest problems”. “We can’t fix democracy, but we can fix capitalism” says founder Chamath Palihapitiya.
Ahmed Husain at Lombard Odier invited Chamath to a private breakfast here in London which gave me a chance to hear the story firsthand as he was out touring Europe for additional funding.
With such lofty stated goals, I arrived a skeptic; I came away fairly impressed by the strategy, if he can pull it off.
The overarching goal behind Social Capital (SC) is to access what CP repeatedly refers to as ground truths. Data has no bias; you can choose to ignore it, but you can’t readily deny it. The challenge Social Capital sets itself is essentially the following: how do we move away from what appears to be a crapshoot to systematize the generation of alpha through better data. In other words, let’s stop trading on credentials and start trading on information.
Chamath is at pains to point to a narrative fallacy that surrounds tech as a quasi mythical engine of disruption : risks are taken, but they are taken incrementally. A company like facebook was built through years of micro-iterations. Success comes from a process of micro-experiments, micro-failures and continuous incremental improvement. This is a debate I have found myself having countless times with non-tech investors who think we play the lottery. We don’t. We de-risk constantly.
The core operational issue the SC organization is trying to solve, both for itself and for its portfolio companies, can be summarized in the following data-related problems:
- how do we collect data in a scalable way?
- what data matters ?
- how do we detect false positives ?
- how do we avoid rebuilding the wheel ?
- how do we use one company’s gain to benefit another ?
The solution designed by Social Capital is to build a platform that leverages first and third party data through machine learning and apply it to funding products. The paradigm is identical to that used by facebook, google or Amazon, except that here the products are investment funds and funding solutions.
To that end the company has built three pillars supporting its strategy:
- first party data collection (access to a company’s logs, AWS infrastructure etc) augmented with third party inputs, much the same way as a DSP would do
- a core machine learning platform
- a knowledge engine that captures best practice, relative performance, predictive analytics etc.
These three pillars aim to power a full suite of funding products, running from black box seed investing to venture, growth, credit and public investments, effectively covering the entire lifecycle of a company.
At rhe time of writing, the organisation has 65 people of which half are focused on tech, data science and engineering.
Chamath argues that this creates a sustainable competitive advantage grounded in:
- the aggregation of strong data science and engineering talent
- the compounding benefits of proprietary data acquired over 7 years
- the ability to continue to innovate around data-driven investing
The benefits to startups are well articulated — SC board members arrive at board meetings armed with objectively observed data-driven insights, not opinions, and use their operational experience to work with founders on achieving excellence. The narrative makes sense, though of course you still need chemistry to make great boards sing.
This paints the picture of a different type of venture firm: data scientists enhancing investment decisions, arming operators with data to sit on boards, and sharing best practice insights across the portfolio.
Social Capital is one of three companies that are attempting to move the venture model forward in recent years, the other two being of course Ycombinator (harnessing the power of networks to fund companies at scale) and Andreessen Horowitz (world-class infrastructure support). Exciting stuff if they pull it off.